/[BackupPC]/upstream/2.1.0/doc/BackupPC.pod
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Annotation of /upstream/2.1.0/doc/BackupPC.pod

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1 - (hide annotations)
Wed Jun 22 19:12:04 2005 UTC (17 years, 3 months ago) by dpavlin
File size: 159588 byte(s)
import of version 2.1.0

1 dpavlin 1 =head1 BackupPC Introduction
2    
3     This documentation describes BackupPC version 2.1.0,
4     released on 20 Jun 2004.
5    
6     =head2 Overview
7    
8     BackupPC is a high-performance, enterprise-grade system for backing up
9     Unix, Linux and WinXX PCs, desktops and laptops to a server's disk.
10     BackupPC is highly configurable and easy to install and maintain.
11    
12     Given the ever decreasing cost of disks and raid systems, it is now
13     practical and cost effective to backup a large number of machines onto
14     a server's local disk or network storage. For some sites this might be
15     the complete backup solution. For other sites additional permanent
16     archives could be created by periodically backing up the server to tape.
17    
18     Features include:
19    
20     =over 4
21    
22     =item *
23    
24     A clever pooling scheme minimizes disk storage and disk I/O.
25     Identical files across multiple backups of the same or different PC
26     are stored only once (using hard links), resulting in substantial
27     savings in disk storage and disk writes.
28    
29     =item *
30    
31     Optional compression provides additional reductions in storage
32     (around 40%). The CPU impact of compression is low since only
33     new files (those not already in the pool) need to be compressed.
34    
35     =item *
36    
37     A powerful http/cgi user interface allows administrators to view log
38     files, configuration, current status and allows users to initiate and
39     cancel backups and browse and restore files from backups.
40    
41     =item *
42    
43     The http/cgi user interface has internationalization (i18n) support,
44     currently providing English, French, German, Spanish, Italian
45     and Dutch.
46    
47     =item *
48    
49     No client-side software is needed. On WinXX the standard smb
50     protocol is used to extract backup data. On linux or unix clients,
51     rsync or tar (over ssh/rsh/nfs) is used to extract backup data.
52     Alternatively, rsync can also be used on WinXX (using cygwin),
53     and Samba could be installed on the linux or unix client to
54     provide smb shares).
55    
56     =item *
57    
58     Flexible restore options. Single files can be downloaded from
59     any backup directly from the CGI interface. Zip or Tar archives
60     for selected files or directories from any backup can also be
61     downloaded from the CGI interface. Finally, direct restore to
62     the client machine (using smb or tar) for selected files or
63     directories is also supported from the CGI interface.
64    
65     =item *
66    
67     BackupPC supports mobile environments where laptops are only
68     intermittently connected to the network and have dynamic IP addresses
69     (DHCP). Configuration settings allow machines connected via slower WAN
70     connections (eg: dial up, DSL, cable) to not be backed up, even if they
71     use the same fixed or dynamic IP address as when they are connected
72     directly to the LAN.
73    
74     =item *
75    
76     Flexible configuration parameters allow multiple backups to be performed
77     in parallel, specification of which shares to backup, which directories
78     to backup or not backup, various schedules for full and incremental
79     backups, schedules for email reminders to users and so on. Configuration
80     parameters can be set system-wide or also on a per-PC basis.
81    
82     =item *
83    
84     Users are sent periodic email reminders if their PC has not
85     recently been backed up. Email content, timing and policies
86     are configurable.
87    
88     =item *
89    
90     BackupPC is Open Source software hosted by SourceForge.
91    
92     =back
93    
94     =head2 Backup basics
95    
96     =over 4
97    
98     =item Full Backup
99    
100     A full backup is a complete backup of a share. BackupPC can be
101     configured to do a full backup at a regular interval (typically
102     weekly). BackupPC can be configured to keep a certain number
103     of full backups. Exponential expiry is also supported, allowing
104     full backups with various vintages to be kept (for example, a
105     settable number of most recent weekly fulls, plus a settable
106     number of older fulls that are 2, 4, 8, or 16 weeks apart).
107    
108     =item Incremental Backup
109    
110     An incremental backup is a backup of files that have changed (based on
111     their modification time) since the last successful full backup. For
112     SMB and tar, BackupPC backups all files that have changed since one
113     hour prior to the start of the last successful full backup. Rsync is
114     more clever: any files whose attributes have changed (ie: uid, gid,
115     mtime, modes, size) since the last full are backed up. Deleted, new
116     files and renamed files are detected by Rsync incrementals.
117     In constrast, SMB and tar incrementals are not able to detect deleted
118     files, renamed files or new files whose modification time is prior to
119     the last full dump.
120    
121     BackupPC can also be configured to keep a certain number of incremental
122     backups, and to keep a smaller number of very old incremental backups.
123     (BackupPC does not support multi-level incremental backups, although it
124     will in a future version.)
125    
126     BackupPC's CGI interface "fills-in" incremental backups based on the
127     last full backup, giving every backup a "full" appearance. This makes
128     browsing and restoring backups easier.
129    
130     =item Partial Backup
131    
132     When a full backup fails or is canceled, and some files have already
133     been backed up, BackupPC keeps a partial backup containing just the
134     files that were backed up successfully. The partial backup is removed
135     when the next successful backup completes, or if another full backup
136     fails resulting in a newer partial backup. A failed full backup
137     that has not backed up any files, or any failed incremental backup,
138     is removed; no partial backup is saved in these cases.
139    
140     The partial backup may be browsed or used to restore files just like
141     a successful full or incremental backup.
142    
143     With the rsync transfer method the partial backup is used to resume
144     the next full backup, avoiding the need to retransfer the file data
145     already in the partial backup.
146    
147     =item Identical Files
148    
149     BackupPC pools identical files using hardlinks. By "identical
150     files" we mean files with identical contents, not necessary the
151     same permissions, ownership or modification time. Two files might
152     have different permissions, ownership, or modification time but
153     will still be pooled whenever the contents are identical. This
154     is possible since BackupPC stores the file meta-data (permissions,
155     ownership, and modification time) separately from the file contents.
156    
157     =item Backup Policy
158    
159     Based on your site's requirements you need to decide what your backup
160     policy is. BackupPC is not designed to provide exact re-imaging of
161     failed disks. See L<Limitations|limitations> for more information.
162     However, the addition of tar transport for linux/unix clients, plus
163     full support for special file types and unix attributes in v1.4.0
164     likely means an exact image of a linux/unix file system can be made.
165    
166     BackupPC saves backups onto disk. Because of pooling you can relatively
167     economically keep several weeks of old backups.
168    
169     At some sites the disk-based backup will be adequate, without a
170     secondary tape backup. This system is robust to any single failure: if a
171     client disk fails or loses files, the BackupPC server can be used to
172     restore files. If the server disk fails, BackupPC can be restarted on a
173     fresh file system, and create new backups from the clients. The chance
174     of the server disk failing can be made very small by spending more money
175     on increasingly better RAID systems. However, there is still the risk
176     of catastrophic events like fires or earthquakes that can destroy
177     both the BackupPC server and the clients it is backing up if they
178     are physically nearby.
179    
180     Some sites might choose to do periodic backups to tape or cd/dvd.
181     This backup can be done perhaps weekly using the archive function of
182     BackupPC.
183    
184     Other users have reported success with removable disks to rotate the
185     BackupPC data drives, or using rsync to mirror the BackupPC data pool
186     offsite.
187    
188     =back
189    
190     =head2 Resources
191    
192     =over 4
193    
194     =item BackupPC home page
195    
196     The BackupPC Open Source project is hosted on SourceForge. The
197     home page can be found at:
198    
199     http://backuppc.sourceforge.net
200    
201     This page has links to the current documentation, the SourceForge
202     project page and general information.
203    
204     =item SourceForge project
205    
206     The SourceForge project page is at:
207    
208     http://sourceforge.net/projects/backuppc
209    
210     This page has links to the current releases of BackupPC.
211    
212     =item BackupPC FAQ
213    
214     BackupPC has a FAQ at L<http://backuppc.sourceforge.net/faq>.
215    
216     =item Mail lists
217    
218     Three BackupPC mailing lists exist for announcements (backuppc-announce),
219     developers (backuppc-devel), and a general user list for support, asking
220     questions or any other topic relevant to BackupPC (backuppc-users).
221    
222     The lists are archived on SourceForge and Gmane. The SourceForge lists
223     are not always up to date and the searching is limited, so Gmane is
224     a good alternative. See:
225    
226     http://news.gmane.org/index.php?prefix=gmane.comp.sysutils.backup.backuppc
227     http://sourceforge.net/mailarchive/forum.php?forum_id=503
228    
229     You can subscribe to these lists by visiting:
230    
231     http://lists.sourceforge.net/lists/listinfo/backuppc-announce
232     http://lists.sourceforge.net/lists/listinfo/backuppc-users
233     http://lists.sourceforge.net/lists/listinfo/backuppc-devel
234    
235     The backuppc-announce list is moderated and is used only for
236     important announcements (eg: new versions). It is low traffic.
237     You only need to subscribe to one of backuppc-announce and
238     backuppc-users: backuppc-users also receives any messages on
239     backuppc-announce.
240    
241     The backuppc-devel list is only for developers who are working on BackupPC.
242     Do not post questions or support requests there. But detailed technical
243     discussions should happen on this list.
244    
245     To post a message to the backuppc-users list, send an email to
246    
247     backuppc-users@lists.sourceforge.net
248    
249     Do not send subscription requests to this address!
250    
251     =item Other Programs of Interest
252    
253     If you want to mirror linux or unix files or directories to a remote server
254     you should consider rsync, L<http://rsync.samba.org>. BackupPC now uses
255     rsync as a transport mechanism; if you are already an rsync user you
256     can think of BackupPC as adding efficient storage (compression and
257     pooling) and a convenient user interface to rsync.
258    
259     Unison is a utility that can do two-way, interactive, synchronization.
260     See L<http://www.cis.upenn.edu/~bcpierce/unison>.
261    
262     Three popular open source packages that do tape backup are
263     Amanda (L<http://www.amanda.org>),
264     afbackup (L<http://sourceforge.net/projects/afbackup>), and
265     Bacula (L<http://www.bacula.org>).
266     Amanda can also backup WinXX machines to tape using samba.
267     These packages can be used as back ends to BackupPC to backup the
268     BackupPC server data to tape.
269    
270     Various programs and scripts use rsync to provide hardlinked backups.
271     See, for example, Mike Rubel's site (L<http://www.mikerubel.org/computers/rsync_snapshots>),
272     JW Schultz's dirvish (L<http://www.pegasys.ws/dirvish> (although as of
273     June 2004 this link doesn't work)),
274     Ben Escoto's rdiff-backup (L<http://rdiff-backup.stanford.edu>),
275     and John Bowman's rlbackup (L<http://www.math.ualberta.ca/imaging/rlbackup>).
276    
277     BackupPC provides many additional features, such as compressed storage,
278     hardlinking any matching files (rather than just files with the same name),
279     and storing special files without root privileges. But these other scripts
280     provide simple and effective solutions and are worthy of consideration.
281    
282     =back
283    
284     =head2 Road map
285    
286     The new features planned for future releases of BackupPC
287     are at L<http://backuppc.sourceforge.net/faq/roadMap.html>.
288    
289     Comments and suggestions are welcome.
290    
291     =head2 You can help
292    
293     BackupPC is free. I work on BackupPC because I enjoy doing it and I like
294     to contribute to the open source community.
295    
296     BackupPC already has more than enough features for my own needs. The
297     main compensation for continuing to work on BackupPC is knowing that
298     more and more people find it useful. So feedback is certainly
299     appreciated, both positive and negative.
300    
301     Beyond being a satisfied user and telling other people about it, everyone
302     is encouraged to add links to L<http://backuppc.sourceforge.net>
303     (I'll see then via Google) or otherwise publicize BackupPC. Unlike
304     the commercial products in this space, I have a zero budget (in both
305     time and money) for marketing, PR and advertising, so it's up to
306     all of you! Feel free to vote for BackupPC at
307     L<http://freshmeat.net/projects/backuppc>.
308    
309     Also, everyone is encouraged to contribute patches, bug reports, feature
310     and design suggestions, new code, FAQs, and documentation corrections or
311     improvements. Answering questions on the mail list is a big help too.
312    
313     =head1 Installing BackupPC
314    
315     =head2 Requirements
316    
317     BackupPC requires:
318    
319     =over 4
320    
321     =item *
322    
323     A linux, solaris, or unix based server with a substantial amount of free
324     disk space (see the next section for what that means). The CPU and disk
325     performance on this server will determine how many simultaneous backups
326     you can run. You should be able to run 4-8 simultaneous backups on a
327     moderately configured server.
328    
329     Several users have reported significantly better performance using
330     reiser compared to ext3 for the BackupPC data file system. It is
331     also recommended you consider either an LVM or raid setup (either
332     in HW or SW; eg: 3Ware RAID5) so that you can expand the
333     file system as necessary.
334    
335     When BackupPC starts with an empty pool, all the backup data will be
336     written to the pool on disk. After more backups are done, a higher
337     percentage of incoming files will already be in the pool. BackupPC is
338     able to avoid writing to disk new files that are already in the pool.
339     So over time disk writes will reduce significantly (by perhaps a factor
340     of 20 or more), since eventually 95% or more of incoming backup files
341     are typically in the pool. Disk reads from the pool are still needed to
342     do file compares to verify files are an exact match. So, with a mature
343     pool, if a relatively fast client generates data at say 1MB/sec, and you
344     run 4 simultaneous backups, there will be an average server disk load of
345     about 4MB/sec reads and 0.2MB/sec writes (assuming 95% of the incoming
346     files are in the pool). These rates will be perhaps 40% lower if
347     compression is on.
348    
349     =item *
350    
351     Perl version 5.6.0 or later. BackupPC has been tested with
352     version 5.6.x, and 5.8.x. If you don't have perl, please
353     see L<http://www.cpan.org>.
354    
355     =item *
356    
357     Perl modules Compress::Zlib, Archive::Zip and File::RsyncP. Try "perldoc
358     Compress::Zlib" and "perldoc Archive::Zip" to see if you have these
359     modules. If not, fetch them from L<http://www.cpan.org> and see the
360     instructions below for how to build and install them.
361    
362     The File::RsyncP module is available from L<http://perlrsync.sourceforge.net>
363     or CPAN. You'll need to install the File::RsyncP module if you want to use
364     Rsync as a transport method.
365    
366     =item *
367    
368     If you are using smb to backup WinXX machines you need smbclient and
369     nmblookup from the samba package. You will also need nmblookup if
370     you are backing up linux/unix DHCP machines. See L<http://www.samba.org>.
371     Version 2.2.0 or later of Samba is required.
372     Samba versions 3.x are stable and now recommended instead of 2.x.
373    
374     See L<http://www.samba.org> for source and binaries. It's pretty easy to
375     fetch and compile samba, and just grab smbclient and nmblookup, without
376     doing the installation. Alternatively, L<http://www.samba.org> has binary
377     distributions for most platforms.
378    
379     =item *
380    
381     If you are using tar to backup linux/unix machines you should have version
382     1.13.7 at a minimum, with version 1.13.20 or higher recommended. Use
383     "tar --version" to check your version. Various GNU mirrors have the newest
384     versions of tar, see for example L<http://www.funet.fi/pub/gnu/alpha/gnu/tar>.
385     As of June 2003 the latest version is 1.13.25.
386    
387     =item *
388    
389     If you are using rsync to backup linux/unix machines you should have
390     version 2.5.5 or higher on each client machine. See
391     L<http://rsync.samba.org>. Use "rsync --version" to check your version.
392    
393     For BackupPC to use Rsync you will also need to install the perl
394     File::RsyncP module, which is available from
395     L<http://perlrsync.sourceforge.net>.
396     Version 0.52 or later is required.
397    
398     =item *
399    
400     The Apache web server, see L<http://www.apache.org>, preferably built
401     with mod_perl support.
402    
403     =back
404    
405     =head2 How much disk space do I need?
406    
407     Here's one real example for an environment that is backing up 65 laptops
408     with compression off. Each full backup averages 3.2GB. Each incremental
409     backup averages about 0.2GB. Storing one full backup and two incremental
410     backups per laptop is around 240GB of raw data. But because of the
411     pooling of identical files, only 87GB is used. This is without
412     compression.
413    
414     Another example, with compression on: backing up 95 laptops, where
415     each backup averages 3.6GB and each incremental averages about 0.3GB.
416     Keeping three weekly full backups, and six incrementals is around
417     1200GB of raw data. Because of pooling and compression, only 150GB
418     is needed.
419    
420     Here's a rule of thumb. Add up the disk usage of all the machines you
421     want to backup (210GB in the first example above). This is a rough
422     minimum space estimate that should allow a couple of full backups and at
423     least half a dozen incremental backups per machine. If compression is on
424     you can reduce the storage requirements by maybe 30-40%. Add some margin
425     in case you add more machines or decide to keep more old backups.
426    
427     Your actual mileage will depend upon the types of clients, operating
428     systems and applications you have. The more uniform the clients and
429     applications the bigger the benefit from pooling common files.
430    
431     For example, the Eudora email tool stores each mail folder in a separate
432     file, and attachments are extracted as separate files. So in the sadly
433     common case of a large attachment emailed to many recipients, Eudora
434     will extract the attachment into a new file. When these machines are
435     backed up, only one copy of the file will be stored on the server, even
436     though the file appears in many different full or incremental backups. In
437     this sense Eudora is a "friendly" application from the point of view of
438     backup storage requirements.
439    
440     An example at the other end of the spectrum is Outlook. Everything
441     (email bodies, attachments, calendar, contact lists) is stored in a
442     single file, which often becomes huge. Any change to this file requires
443     a separate copy of the file to be saved during backup. Outlook is even
444     more troublesome, since it keeps this file locked all the time, so it
445     cannot be read by smbclient whenever Outlook is running. See the
446     L<Limitations|limitations> section for more discussion of this problem.
447    
448     In addition to total disk space, you shold make sure you have
449     plenty of inodes on your BackupPC data partition. Some users have
450     reported running out of inodes on their BackupPC data partition.
451     So even if you have plenty of disk space, BackupPC will report
452     failures when the inodes are exhausted. This is a particular
453     problem with ext2/ext3 file systems that have a fixed number of
454     inodes when the file system is built. Use "df -i" to see your
455     inode usage.
456    
457     =head2 Step 1: Getting BackupPC
458    
459     Some linux distributions now include BackupPC. The Debian
460     distribution, supprted by Ludovic Drolez, can be found at
461     L<http://packages.debian.org/backuppc>; it should be included
462     in the next stable Debian release. On Debian, BackupPC can
463     be installed with the command:
464    
465     apt-get install backuppc
466    
467     In the future there might be packages for Gentoo and other
468     linux flavors. If the packaged version is older than the
469     released version then you will probably want to install the
470     lastest version as described below.
471    
472     Otherwise, manually fetching and installing BackupPC is easy.
473     Start by downloading the latest version from
474     L<http://backuppc.sourceforge.net>. Hit the "Code" button,
475     then select the "backuppc" or "backuppc-beta" package and
476     download the latest version.
477    
478     =head2 Step 2: Installing the distribution
479    
480     First off, there are three perl modules you should install.
481     These are all optional, but highly recommended:
482    
483     =over 4
484    
485     =item Compress::Zlib
486    
487     To enable compression, you will need to install Compress::Zlib
488     from L<http://www.cpan.org>.
489     You can run "perldoc Compress::Zlib" to see if this module is installed.
490    
491     =item Archive::Zip
492    
493     To support restore via Zip archives you will need to install
494     Archive::Zip, also from L<http://www.cpan.org>.
495     You can run "perldoc Archive::Zip" to see if this module is installed.
496    
497     =item File::RsyncP
498    
499     To use rsync and rsyncd with BackupPC you will need to install File::RsyncP.
500     You can run "perldoc File::RsyncP" to see if this module is installed.
501     File::RsyncP is available from L<http://perlrsync.sourceforge.net>.
502     Version 0.52 or later is required.
503    
504     =back
505    
506     To build and install these packages, fetch the tar.gz file and
507     then run these commands:
508    
509     tar zxvf Archive-Zip-1.01.tar.gz
510     cd Archive-Zip-1.01
511     perl Makefile.PL
512     make
513     make test
514     make install
515    
516     The same sequence of commands can be used for each module.
517    
518     Now let's move onto BackupPC itself. After fetching
519     BackupPC-2.1.0.tar.gz, run these commands as root:
520    
521     tar zxf BackupPC-2.1.0.tar.gz
522     cd BackupPC-2.1.0
523     perl configure.pl
524    
525     In the future this release might also have patches available on the
526     SourceForge site. These patch files are text files, with a name of
527     the form
528    
529     BackupPC-2.1.0plN.diff
530    
531     where N is the patch level, eg: pl5 is patch-level 5. These
532     patch files are cumulative: you only need apply the last patch
533     file, not all the earlier patch files. If a patch file is
534     available, eg: BackupPC-2.1.0pl5.diff, you should apply
535     the patch after extracting the tar file:
536    
537     # fetch BackupPC-2.1.0.tar.gz
538     # fetch BackupPC-2.1.0pl5.diff
539     tar zxf BackupPC-2.1.0.tar.gz
540     cd BackupPC-2.1.0
541     patch -p0 < ../BackupPC-2.1.0pl5.diff
542     perl configure.pl
543    
544     A patch file includes comments that describe that bug fixes
545     and changes. Feel free to review it before you apply the patch.
546    
547     The configure.pl script also accepts command-line options if you
548     wish to run it in a non-interactive manner. It has self-contained
549     documentation for all the command-line options, which you can
550     read with perldoc:
551    
552     perldoc configure.pl
553    
554     When you run configure.pl you will be prompted for the full paths
555     of various executables, and you will be prompted for the following
556     information:
557    
558     =over 4
559    
560     =item BackupPC User
561    
562     It is best if BackupPC runs as a special user, eg backuppc, that has
563     limited privileges. It is preferred that backuppc belongs to a system
564     administrator group so that sys admin members can browse backuppc files,
565     edit the configuration files and so on. Although configurable, the
566     default settings leave group read permission on pool files, so make
567     sure the BackupPC user's group is chosen restrictively.
568    
569     On this installation, this is __BACKUPPCUSER__.
570    
571     =item Data Directory
572    
573     You need to decide where to put the data directory, below which
574     all the BackupPC data is stored. This needs to be a big file system.
575    
576     On this installation, this is __TOPDIR__.
577    
578     =item Install Directory
579    
580     You should decide where the BackupPC scripts, libraries and documentation
581     should be installed, eg: /opt/local/BackupPC.
582    
583     On this installation, this is __INSTALLDIR__.
584    
585     =item CGI bin Directory
586    
587     You should decide where the BackupPC CGI script resides. This will
588     usually below Apache's cgi-bin directory.
589    
590     On this installation, this is __CGIDIR__.
591    
592     =item Apache image directory
593    
594     A directory where BackupPC's images are stored so that Apache can
595     serve them. This should be somewhere under Apache's DocumentRoot
596     directory.
597    
598     =back
599    
600     =head2 Step 3: Setting up config.pl
601    
602     After running configure.pl, browse through the config file,
603     __INSTALLDIR__/conf/config.pl, and make sure all the default settings
604     are correct. In particular, you will need to decide whether to use
605     smb, tar or rsync transport (or whether to set it on a per-PC basis)
606     and set the relevant parameters for that transport method.
607     See the section L<Client Setup|step 5: client setup> for more details.
608    
609     =head2 Step 4: Setting up the hosts file
610    
611     The file __TOPDIR__/conf/hosts contains the list of clients to backup.
612     BackupPC reads this file in three cases:
613    
614     =over 4
615    
616     =item *
617    
618     Upon startup.
619    
620     =item *
621    
622     When BackupPC is sent a HUP (-1) signal. Assuming you installed the
623     init.d script, you can also do this with "/etc/init.d/backuppc reload".
624    
625     =item *
626    
627     When the modification time of the hosts file changes. BackupPC
628     checks the modification time once during each regular wakeup.
629    
630     =back
631    
632     Whenever you change the hosts file (to add or remove a host) you can
633     either do a kill -HUP BackupPC_pid or simply wait until the next regular
634     wakeup period.
635    
636     Each line in the hosts file contains three fields, separated
637     by white space:
638    
639     =over 4
640    
641     =item Host name
642    
643     This is typically the host name or NetBios name of the client machine
644     and should be in lower case. The host name can contain spaces (escape
645     with a backslash), but it is not recommended.
646    
647     Please read the section L<How BackupPC Finds Hosts|how backuppc finds hosts>.
648    
649     In certain cases you might want several distinct clients to refer
650     to the same physical machine. For example, you might have a database
651     you want to backup, and you want to bracket the backup of the database
652     with shutdown/restart using $Conf{DumpPreUserCmd} and $Conf{DumpPostUserCmd}.
653     But you also want to backup the rest of the machine while the database
654     is still running. In the case you can specify two different clients in
655     the host file, using any mnemonic name (eg: myhost_mysql and myhost), and
656     use $Conf{ClientNameAlias} in myhost_mysql's config.pl to specify the
657     real host name of the machine.
658    
659     =item DHCP flag
660    
661     Starting with v2.0.0 the way hosts are discovered has changed and now
662     in most cases you should specify 0 for the DHCP flag, even if the host
663     has a dynamically assigned IP address.
664     Please read the section L<How BackupPC Finds Hosts|how backuppc finds hosts>
665     to understand whether you need to set the DHCP flag.
666    
667     You only need to set DHCP to 1 if your client machine doesn't
668     respond to the NetBios multicast request:
669    
670     nmblookup myHost
671    
672     but does respond to a request directed to its IP address:
673    
674     nmblookup -A W.X.Y.Z
675    
676     If you do set DHCP to 1 on any client you will need to specify the range of
677     DHCP addresses to search is specified in $Conf{DHCPAddressRanges}.
678    
679     Note also that the $Conf{ClientNameAlias} feature does not work for
680     clients with DHCP set to 1.
681    
682     =item User name
683    
684     This should be the unix login/email name of the user who "owns" or uses
685     this machine. This is the user who will be sent email about this
686     machine, and this user will have permission to stop/start/browse/restore
687     backups for this host. Leave this blank if no specific person should
688     receive email or be allowed to stop/start/browse/restore backups
689     for this host. Administrators will still have full permissions.
690    
691     =item More users
692    
693     Additional user names, separate by commas and with no white space,
694     can be specified. These users will also have full permission in
695     the CGI interface to stop/start/browse/restore backups for this host.
696     These users will not be sent email about this host.
697    
698     =back
699    
700     The first non-comment line of the hosts file is special: it contains
701     the names of the columns and should not be edited.
702    
703     Here's a simple example of a hosts file:
704    
705     host dhcp user moreUsers
706     farside 0 craig jim,dave
707     larson 1 gary andy
708    
709     =head2 Step 5: Client Setup
710    
711     Two methods for getting backup data from a client are supported: smb and
712     tar. Smb or rsync are the preferred methods for WinXX clients and rsync or
713     tar are the preferred methods for linux/unix clients.
714    
715     The transfer method is set using the $Conf{XferMethod} configuration
716     setting. If you have a mixed environment (ie: you will use smb for some
717     clients and tar for others), you will need to pick the most common
718     choice for $Conf{XferMethod} for the main config.pl file, and then
719     override it in the per-PC config file for those hosts that will use
720     the other method. (Or you could run two completely separate instances
721     of BackupPC, with different data directories, one for WinXX and the
722     other for linux/unix, but then common files between the different
723     machine types will duplicated.)
724    
725     Here are some brief client setup notes:
726    
727     =over 4
728    
729     =item WinXX
730    
731     The preferred setup for WinXX clients is to set $Conf{XferMethod} to "smb".
732     (Actually, for v2.0.0, rsyncd is the better method for WinXX if you are
733     prepared to run rsync/cygwin on your WinXX client. More information
734     about this will be provided via the FAQ.)
735    
736     If you want to use rsyncd for WinXX clients you can find a pre-packaged
737     zip file on L<http://backuppc.sourceforge.net>. The package is called
738     cygwin-rsync. It contains rsync.exe, template setup files and the
739     minimal set of cygwin libraries for everything to run. The README file
740     contains instructions for running rsync as a service, so it starts
741     automatically everytime you boot your machine.
742    
743     If you build your own rsync, for rsync 2.6.2 it is strongly
744     recommended you apply the patch in the cygwin-rsync package on
745     L<http://backuppc.sourceforge.net>. This patch adds the --checksum-seed
746     option for checksum caching, and also sends all errors to the client,
747     which is important so BackupPC can log all file access errors.
748    
749     Otherwise, to use SMB, you need to create shares for the data you want
750     to backup. Open "My Computer", right click on the drive (eg: C), and
751     select "Sharing..." (or select "Properties" and select the "Sharing"
752     tab). In this dialog box you can enable sharing, select the share name
753     and permissions. Many machines will be configured by default to share
754     the entire C drive as C$ using the administrator password.
755    
756     If this machine uses DHCP you will also need to make sure the
757     NetBios name is set. Go to Control Panel|System|Network Identification
758     (on Win2K) or Control Panel|System|Computer Name (on WinXP).
759     Also, you should go to Control Panel|Network Connections|Local Area
760     Connection|Properties|Internet Protocol (TCP/IP)|Properties|Advanced|WINS
761     and verify that NetBios is not disabled.
762    
763     The relevant configuration settings are $Conf{SmbShareName},
764     $Conf{SmbShareUserName}, $Conf{SmbSharePasswd}, $Conf{SmbClientPath},
765     $Conf{SmbClientFullCmd}, $Conf{SmbClientIncrCmd} and
766     $Conf{SmbClientRestoreCmd}.
767    
768     BackupPC needs to know the smb share user name and password for a
769     client machine that uses smb. The user name is specified in
770     $Conf{SmbShareUserName}. There are four ways to tell BackupPC the
771     smb share password:
772    
773     =over 4
774    
775     =item *
776    
777     As an environment variable BPC_SMB_PASSWD set before BackupPC starts.
778     If you start BackupPC manually the BPC_SMB_PASSWD variable must be set
779     manually first. For backward compatibility for v1.5.0 and prior, the
780     environment variable PASSWD can be used if BPC_SMB_PASSWD is not set.
781     Warning: on some systems it is possible to see environment variables of
782     running processes.
783    
784     =item *
785    
786     Alternatively the BPC_SMB_PASSWD setting can be included in
787     /etc/init.d/backuppc, in which case you must make sure this file
788     is not world (other) readable.
789    
790     =item *
791    
792     As a configuration variable $Conf{SmbSharePasswd} in
793     __TOPDIR__/conf/config.pl. If you put the password
794     here you must make sure this file is not world (other) readable.
795    
796     =item *
797    
798     As a configuration variable $Conf{SmbSharePasswd} in the per-PC
799     configuration file, __TOPDIR__/pc/$host/config.pl. You will have to
800     use this option if the smb share password is different for each host.
801     If you put the password here you must make sure this file is not
802     world (other) readable.
803    
804     =back
805    
806     Placement and protection of the smb share password is a possible
807     security risk, so please double-check the file and directory
808     permissions. In a future version there might be support for
809     encryption of this password, but a private key will still have to
810     be stored in a protected place. Suggestions are welcome.
811    
812     As an alternative to setting $Conf{XferMethod} to "smb" (using
813     smbclient) for WinXX clients, you can use an smb network filesystem (eg:
814     ksmbfs or similar) on your linux/unix server to mount the share,
815     and then set $Conf{XferMethod} to "tar" (use tar on the network
816     mounted file system).
817    
818     Also, to make sure that file names with 8-bit characters are correctly
819     transferred by smbclient you should add this to samba's smb.conf file
820     for samba 2.x:
821    
822     [global]
823     # Accept the windows charset
824     client code page = 850
825     character set = ISO8859-1
826    
827     For samba 3.x this should instead be:
828    
829     [global]
830     unix charset = ISO8859-1
831    
832     This setting should work for western europe.
833     See L<http://www.oreilly.com/catalog/samba/chapter/book/ch08_03.html>
834     for more information about settings for other languages.
835    
836     =item Linux/Unix
837    
838     The preferred setup for linux/unix clients is to set $Conf{XferMethod}
839     to "rsync", "rsyncd" or "tar".
840    
841     You can use either rsync, smb, or tar for linux/unix machines. Smb requires
842     that the Samba server (smbd) be run to provide the shares. Since the smb
843     protocol can't represent special files like symbolic links and fifos,
844     tar and rsync are the better transport methods for linux/unix machines.
845     (In fact, by default samba makes symbolic links look like the file or
846     directory that they point to, so you could get an infinite loop if a
847     symbolic link points to the current or parent directory. If you really
848     need to use Samba shares for linux/unix backups you should turn off the
849     "follow symlinks" samba config setting. See the smb.conf manual page.)
850    
851     The requirements for each Xfer Method are:
852    
853     =over 4
854    
855     =item tar
856    
857     You must have GNU tar on the client machine. Use "tar --version"
858     or "gtar --version" to verify. The version should be at least
859     1.13.7, and 1.13.20 or greater is recommended. Tar is run on
860     the client machine via rsh or ssh.
861    
862     The relevant configuration settings are $Conf{TarClientPath},
863     $Conf{TarShareName}, $Conf{TarClientCmd}, $Conf{TarFullArgs},
864     $Conf{TarIncrArgs}, and $Conf{TarClientRestoreCmd}.
865    
866     =item rsync
867    
868     You should have at least rsync 2.5.5, and the latest version 2.5.6
869     is recommended. Rsync is run on the remote client via rsh or ssh.
870    
871     The relevant configuration settings are $Conf{RsyncClientPath},
872     $Conf{RsyncClientCmd}, $Conf{RsyncClientRestoreCmd}, $Conf{RsyncShareName},
873     $Conf{RsyncArgs}, and $Conf{RsyncRestoreArgs}.
874    
875     =item rsyncd
876    
877     You should have at least rsync 2.5.5, and the latest version 2.6.2
878     is recommended. In this case the rsync daemon should be running on
879     the client machine and BackupPC connects directly to it.
880    
881     The relevant configuration settings are $Conf{RsyncdClientPort},
882     $Conf{RsyncdUserName}, $Conf{RsyncdPasswd}, $Conf{RsyncdAuthRequired},
883     $Conf{RsyncShareName}, $Conf{RsyncArgs}, and $Conf{RsyncRestoreArgs}.
884     $Conf{RsyncShareName} is the name of an rsync module (ie: the thing
885     in square brackets in rsyncd's conf file -- see rsyncd.conf), not a
886     file system path.
887    
888     Be aware that rsyncd will remove the leading '/' from path names in
889     symbolic links if you specify "use chroot = no" in the rsynd.conf file.
890     See the rsyncd.conf manual page for more information.
891    
892     =back
893    
894     For linux/unix machines you should not backup "/proc". This directory
895     contains a variety of files that look like regular files but they are
896     special files that don't need to be backed up (eg: /proc/kcore is a
897     regular file that contains physical memory). See $Conf{BackupFilesExclude}.
898     It is safe to back up /dev since it contains mostly character-special
899     and block-special files, which are correctly handed by BackupPC
900     (eg: backing up /dev/hda5 just saves the block-special file information,
901     not the contents of the disk).
902    
903     Alternatively, rather than backup all the file systems as a single
904     share ("/"), it is easier to restore a single file system if you backup
905     each file system separately. To do this you should list each file system
906     mount point in $Conf{TarShareName} or $Conf{RsyncShareName}, and add the
907     --one-file-system option to $Conf{TarClientCmd} or add --one-file-system
908     (note the different punctuation) to $Conf{RsyncArgs}. In this case there
909     is no need to exclude /proc explicitly since it looks like a different
910     file system.
911    
912     Next you should decide whether to run tar over ssh, rsh or nfs. Ssh is
913     the preferred method. Rsh is not secure and therefore not recommended.
914     Nfs will work, but you need to make sure that the BackupPC user (running
915     on the server) has sufficient permissions to read all the files below
916     the nfs mount.
917    
918     Ssh allows BackupPC to run as a privileged user on the client (eg:
919     root), since it needs sufficient permissions to read all the backup
920     files. Ssh is setup so that BackupPC on the server (an otherwise low
921     privileged user) can ssh as root on the client, without being prompted
922     for a password. There are two common versions of ssh: v1 and v2. Here
923     are some instructions for one way to setup ssh. (Check which version
924     of SSH you have by typing "ssh" or "man ssh".)
925    
926     =item Mac OS X
927    
928     In general this should be similar to Linux/Unix machines.
929     Mark Stosberg reports that you can also use hfstar.
930     See L<http://fink.sourceforge.net/pdb/package.php/hfstar>.
931    
932     =item SSH Setup
933    
934     SSH is a secure way to run tar or rsync on a backup client to extract
935     the data. SSH provides strong authentication and encryption of
936     the network data.
937    
938     Note that if you run rsyncd (rsync daemon), ssh is not used.
939     In this case, rsyncd provides its own authentication, but there
940     is no encryption of network data. If you want encryption of
941     network data you can use ssh to create a tunnel, or use a
942     program like stunnel. If someone submits instructions I
943    
944     Setup instructions for ssh are at
945     L<http://backuppc.sourceforge.net/faq/ssh.html>.
946    
947     =item Clients that use DHCP
948    
949     If a client machine uses DHCP BackupPC needs some way to find the
950     IP address given the host name. One alternative is to set dhcp
951     to 1 in the hosts file, and BackupPC will search a pool of IP
952     addresses looking for hosts. More efficiently, it is better to
953     set dhcp = 0 and provide a mechanism for BackupPC to find the
954     IP address given the host name.
955    
956     For WinXX machines BackupPC uses the NetBios name server to determine
957     the IP address given the host name.
958     For unix machines you can run nmbd (the NetBios name server) from
959     the Samba distribution so that the machine responds to a NetBios
960     name request. See the manual page and Samba documentation for more
961     information.
962    
963     Alternatively, you can set $Conf{NmbLookupFindHostCmd} to any command
964     that returns the IP address given the host name.
965    
966     Please read the section L<How BackupPC Finds Hosts|how backuppc finds hosts>
967     for more details.
968    
969     =back
970    
971     =head2 Step 6: Running BackupPC
972    
973     The installation contains an init.d backuppc script that can be copied
974     to /etc/init.d so that BackupPC can auto-start on boot.
975     See init.d/README for further instructions.
976    
977     BackupPC should be ready to start. If you installed the init.d script,
978     then you should be able to run BackupPC with:
979    
980     /etc/init.d/backuppc start
981    
982     (This script can also be invoked with "stop" to stop BackupPC and "reload"
983     to tell BackupPC to reload config.pl and the hosts file.)
984    
985     Otherwise, just run
986    
987     __INSTALLDIR__/bin/BackupPC -d
988    
989     as user __BACKUPPCUSER__. The -d option tells BackupPC to run as a daemon
990     (ie: it does an additional fork).
991    
992     Any immediate errors will be printed to stderr and BackupPC will quit.
993     Otherwise, look in __TOPDIR__/log/LOG and verify that BackupPC reports
994     it has started and all is ok.
995    
996     =head2 Step 7: Talking to BackupPC
997    
998     Note: as of version 1.5.0, BackupPC no longer supports telnet
999     to its TCP port. First off, a unix domain socket is used
1000     instead of a TCP port. (The TCP port can still be re-enabled
1001     if your installation has apache and BackupPC running on different
1002     machines.) Secondly, even if you still use the TCP port, the
1003     messages exchanged over this interface are now protected by
1004     an MD5 digest based on a shared secret (see $Conf{ServerMesgSecret})
1005     as well as sequence numbers and per-session unique keys, preventing
1006     forgery and replay attacks.
1007    
1008     You should verify that BackupPC is running by using BackupPC_serverMesg.
1009     This sends a message to BackupPC via the unix (or TCP) socket and prints
1010     the response.
1011    
1012     You can request status information and start and stop backups using this
1013     interface. This socket interface is mainly provided for the CGI interface
1014     (and some of the BackupPC sub-programs use it too). But right now we just
1015     want to make sure BackupPC is happy. Each of these commands should
1016     produce some status output:
1017    
1018     __INSTALLDIR__/bin/BackupPC_serverMesg status info
1019     __INSTALLDIR__/bin/BackupPC_serverMesg status jobs
1020     __INSTALLDIR__/bin/BackupPC_serverMesg status hosts
1021    
1022     The output should be some hashes printed with Data::Dumper. If it
1023     looks cryptic and confusing, and doesn't look like an error message,
1024     then all is ok.
1025    
1026     The jobs status should initially show just BackupPC_trashClean.
1027     The hosts status should produce a list of every host you have listed
1028     in __TOPDIR__/conf/hosts as part of a big cryptic output line.
1029    
1030     You can also request that all hosts be queued:
1031    
1032     __INSTALLDIR__/bin/BackupPC_serverMesg backup all
1033    
1034     At this point you should make sure the CGI interface works since
1035     it will be much easier to see what is going on. That's our
1036     next subject.
1037    
1038     =head2 Step 8: CGI interface
1039    
1040     The CGI interface script, BackupPC_Admin, is a powerful and flexible
1041     way to see and control what BackupPC is doing. It is written for an
1042     Apache server. If you don't have Apache, see L<http://www.apache.org>.
1043    
1044     There are two options for setting up the CGI interface: standard
1045     mode and using mod_perl. Mod_perl provides much higher performance
1046     (around 15x) and is the best choice if your Apache was built with
1047     mod_perl support. To see if your apache was built with mod_perl
1048     run this command:
1049    
1050     httpd -l | egrep mod_perl
1051    
1052     If this prints mod_perl.c then your Apache supports mod_perl.
1053    
1054     Using mod_perl with BackupPC_Admin requires a dedicated Apache
1055     to be run as the BackupPC user (__BACKUPPCUSER__). This is
1056     because BackupPC_Admin needs permission to access various files
1057     in BackupPC's data directories. In contrast, the standard
1058     installation (without mod_perl) solves this problem by having
1059     BackupPC_Admin installed as setuid to the BackupPC user, so that
1060     BackupPC_Admin runs as the BackuPC user.
1061    
1062     Here are some specifics for each setup:
1063    
1064     =over 4
1065    
1066     =item Standard Setup
1067    
1068     The CGI interface should have been installed by the configure.pl script
1069     in __CGIDIR__/BackupPC_Admin. BackupPC_Admin should have been installed
1070     as setuid to the BackupPC user (__BACKUPPCUSER__), in addition to user
1071     and group execute permission.
1072    
1073     You should be very careful about permissions on BackupPC_Admin and
1074     the directory __CGIDIR__: it is important that normal users cannot
1075     directly execute or change BackupPC_Admin, otherwise they can access
1076     backup files for any PC. You might need to change the group ownership
1077     of BackupPC_Admin to a group that Apache belongs to so that Apache
1078     can execute it (don't add "other" execute permission!).
1079     The permissions should look like this:
1080    
1081     ls -l __CGIDIR__/BackupPC_Admin
1082     -swxr-x--- 1 __BACKUPPCUSER__ web 82406 Jun 17 22:58 __CGIDIR__/BackupPC_Admin
1083    
1084     The setuid script won't work unless perl on your machine was installed
1085     with setuid emulation. This is likely the problem if you get an error
1086     saying such as "Wrong user: my userid is 25, instead of 150", meaning
1087     the script is running as the httpd user, not the BackupPC user.
1088     This is because setuid scripts are disabled by the kernel in most
1089     flavors of unix and linux.
1090    
1091     To see if your perl has setuid emulation, see if there is a program
1092     called sperl5.6.0 (or sperl5.8.2 etc, based on your perl version)
1093     in the place where perl is installed. If you can't find this program,
1094     then you have two options: rebuild and reinstall perl with the setuid
1095     emulation turned on (answer "y" to the question "Do you want to do
1096     setuid/setgid emulation?" when you run perl's configure script), or
1097     switch to the mod_perl alternative for the CGI script (which doesn't
1098     need setuid to work).
1099    
1100     =item Mod_perl Setup
1101    
1102     The advantage of the mod_perl setup is that no setuid script is needed,
1103     and there is a huge performance advantage. Not only does all the perl
1104     code need to be parsed just once, the config.pl and hosts files, plus
1105     the connection to the BackupPC server are cached between requests. The
1106     typical speedup is around 15 times.
1107    
1108     To use mod_perl you need to run Apache as user __BACKUPPCUSER__.
1109     If you need to run multiple Apache's for different services then
1110     you need to create multiple top-level Apache directories, each
1111     with their own config file. You can make copies of /etc/init.d/httpd
1112     and use the -d option to httpd to point each http to a different
1113     top-level directory. Or you can use the -f option to explicitly
1114     point to the config file. Multiple Apache's will run on different
1115     Ports (eg: 80 is standard, 8080 is a typical alternative port accessed
1116     via http://yourhost.com:8080).
1117    
1118     Inside BackupPC's Apache http.conf file you should check the
1119     settings for ServerRoot, DocumentRoot, User, Group, and Port. See
1120     L<http://httpd.apache.org/docs/server-wide.html> for more details.
1121    
1122     For mod_perl, BackupPC_Admin should not have setuid permission, so
1123     you should turn it off:
1124    
1125     chmod u-s __CGIDIR__/BackupPC_Admin
1126    
1127     To tell Apache to use mod_perl to execute BackupPC_Admin, add this
1128     to Apache's 1.x httpd.conf file:
1129    
1130     <IfModule mod_perl.c>
1131     PerlModule Apache::Registry
1132     PerlTaintCheck On
1133     <Location /cgi-bin/BackupPC/BackupPC_Admin> # <--- change path as needed
1134     SetHandler perl-script
1135     PerlHandler Apache::Registry
1136     Options ExecCGI
1137     PerlSendHeader On
1138     </Location>
1139     </IfModule>
1140    
1141     Apache 2.0.44 with Perl 5.8.0 on RedHat 7.1, Don Silvia reports that
1142     this works (with tweaks from Michael Tuzi):
1143    
1144     LoadModule perl_module modules/mod_perl.so
1145     PerlModule Apache2
1146    
1147     <Directory /path/to/cgi/>
1148     SetHandler perl-script
1149     PerlResponseHandler ModPerl::Registry
1150     PerlOptions +ParseHeaders
1151     Options +ExecCGI
1152     Order deny,allow
1153     Deny from all
1154     Allow from 192.168.0
1155     AuthName "Backup Admin"
1156     AuthType Basic
1157     AuthUserFile /path/to/user_file
1158     Require valid-user
1159     </Directory>
1160    
1161     There are other optimizations and options with mod_perl. For
1162     example, you can tell mod_perl to preload various perl modules,
1163     which saves memory compared to loading separate copies in every
1164     Apache process after they are forked. See Stas's definitive
1165     mod_perl guide at L<http://perl.apache.org/guide>.
1166    
1167     =back
1168    
1169     BackupPC_Admin requires that users are authenticated by Apache.
1170     Specifically, it expects that Apache sets the REMOTE_USER environment
1171     variable when it runs. There are several ways to do this. One way
1172     is to create a .htaccess file in the cgi-bin directory that looks like:
1173    
1174     AuthGroupFile /etc/httpd/conf/group # <--- change path as needed
1175     AuthUserFile /etc/http/conf/passwd # <--- change path as needed
1176     AuthType basic
1177     AuthName "access"
1178     require valid-user
1179    
1180     You will also need "AllowOverride Indexes AuthConfig" in the Apache
1181     httpd.conf file to enable the .htaccess file. Alternatively, everything
1182     can go in the Apache httpd.conf file inside a Location directive. The
1183     list of users and password file above can be extracted from the NIS
1184     passwd file.
1185    
1186     One alternative is to use LDAP. In Apache's http.conf add these lines:
1187    
1188     LoadModule auth_ldap_module modules/auth_ldap.so
1189     AddModule auth_ldap.c
1190    
1191     # cgi-bin - auth via LDAP (for BackupPC)
1192     <Location /cgi-binBackupPC/BackupPC_Admin> # <--- change path as needed
1193     AuthType Basic
1194     AuthName "BackupPC login"
1195     # replace MYDOMAIN, PORT, ORG and CO as needed
1196     AuthLDAPURL ldap://ldap.MYDOMAIN.com:PORT/o=ORG,c=CO?uid?sub?(objectClass=*)
1197     require valid-user
1198     </Location>
1199    
1200     If you want to disable the user authentication you can set
1201     $Conf{CgiAdminUsers} to '*', which allows any user to have
1202     full access to all hosts and backups. In this case the REMOTE_USER
1203     environment variable does not have to be set by Apache.
1204    
1205     Alternatively, you can force a particular user name by getting Apache
1206     to set REMOTE_USER, eg, to hardcode the user to www you could add
1207     this to Apache's httpd.conf:
1208    
1209     <Location /cgi-bin/BackupPC/BackupPC_Admin> # <--- change path as needed
1210     Setenv REMOTE_USER www
1211     </Location>
1212    
1213     Finally, you should also edit the config.pl file and adjust, as necessary,
1214     the CGI-specific settings. They're near the end of the config file. In
1215     particular, you should specify which users or groups have administrator
1216     (privileged) access: see the config settings $Conf{CgiAdminUserGroup}
1217     and $Conf{CgiAdminUsers}. Also, the configure.pl script placed various
1218     images into $Conf{CgiImageDir} that BackupPC_Admin needs to serve
1219     up. You should make sure that $Conf{CgiImageDirURL} is the correct
1220     URL for the image directory.
1221    
1222     See the section L<Fixing installation problems|fixing installation problems> for suggestions on debugging the Apache authentication setup.
1223    
1224     =head2 How BackupPC Finds Hosts
1225    
1226     Starting with v2.0.0 the way hosts are discovered has changed. In most
1227     cases you should specify 0 for the DHCP flag in the conf/hosts file,
1228     even if the host has a dynamically assigned IP address.
1229    
1230     BackupPC (starting with v2.0.0) looks up hosts with DHCP = 0 in this manner:
1231    
1232     =over 4
1233    
1234     =item *
1235    
1236     First DNS is used to lookup the IP address given the client's name
1237     using perl's gethostbyname() function. This should succeed for machines
1238     that have fixed IP addresses that are known via DNS. You can manually
1239     see whether a given host have a DNS entry according to perls'
1240     gethostbyname function with this command:
1241    
1242     perl -e 'print(gethostbyname("myhost") ? "ok\n" : "not found\n");'
1243    
1244     =item *
1245    
1246     If gethostbyname() fails, BackupPC then attempts a NetBios multicast to
1247     find the host. Provided your client machine is configured properly,
1248     it should respond to this NetBios multicast request. Specifically,
1249     BackupPC runs a command of this form:
1250    
1251     nmblookup myhost
1252    
1253     If this fails you will see output like:
1254    
1255     querying myhost on 10.10.255.255
1256     name_query failed to find name myhost
1257    
1258     If this success you will see output like:
1259    
1260     querying myhost on 10.10.255.255
1261     10.10.1.73 myhost<00>
1262    
1263     Depending on your netmask you might need to specify the -B option to
1264     nmblookup. For example:
1265    
1266     nmblookup -B 10.10.1.255 myhost
1267    
1268     If necessary, experiment on the nmblookup command that will return the
1269     IP address of the client given its name. Then update
1270     $Conf{NmbLookupFindHostCmd} with any necessary options to nmblookup.
1271    
1272     =back
1273    
1274     For hosts that have the DHCP flag set to 1, these machines are
1275     discovered as follows:
1276    
1277     =over 4
1278    
1279     =item *
1280    
1281     A DHCP address pool ($Conf{DHCPAddressRanges}) needs to be specified.
1282     BackupPC will check the NetBIOS name of each machine in the range using
1283     a command of the form:
1284    
1285     nmblookup -A W.X.Y.Z
1286    
1287     where W.X.Y.Z is each candidate address from $Conf{DHCPAddressRanges}.
1288     Any host that has a valid NetBIOS name returned by this command (ie:
1289     matching an entry in the hosts file) will be backed up. You can
1290     modify the specific nmblookup command if necessary via $Conf{NmbLookupCmd}.
1291    
1292     =item *
1293    
1294     You only need to use this DHCP feature if your client machine doesn't
1295     respond to the NetBios multicast request:
1296    
1297     nmblookup myHost
1298    
1299     but does respond to a request directed to its IP address:
1300    
1301     nmblookup -A W.X.Y.Z
1302    
1303     =back
1304    
1305     =head2 Other installation topics
1306    
1307     =over 4
1308    
1309     =item Removing a client
1310    
1311     If there is a machine that no longer needs to be backed up (eg: a retired
1312     machine) you have two choices. First, you can keep the backups accessible
1313     and browsable, but disable all new backups. Alternatively, you can
1314     completely remove the client and all its backups.
1315    
1316     To disable backups for a client there are two special values for
1317     $Conf{FullPeriod} in that client's per-PC config.pl file:
1318    
1319     =over 4
1320    
1321     =item -1
1322    
1323     Don't do any regular backups on this machine. Manually
1324     requested backups (via the CGI interface) will still occur.
1325    
1326     =item -2
1327    
1328     Don't do any backups on this machine. Manually requested
1329     backups (via the CGI interface) will be ignored.
1330    
1331     =back
1332    
1333     This will still allow that client's old backups to be browsable
1334     and restorable.
1335    
1336     To completely remove a client and all its backups, you should remove its
1337     entry in the conf/hosts file, and then delete the __TOPDIR__/pc/$host
1338     directory. Whenever you change the hosts file, you should send
1339     BackupPC a HUP (-1) signal so that it re-reads the hosts file.
1340     If you don't do this, BackupPC will automatically re-read the
1341     hosts file at the next regular wakeup.
1342    
1343     Note that when you remove a client's backups you won't initially recover
1344     a lot of disk space. That's because the client's files are still in
1345     the pool. Overnight, when BackupPC_nightly next runs, all the unused
1346     pool files will be deleted and this will recover the disk space used
1347     by the client's backups.
1348    
1349     =item Copying the pool
1350    
1351     If the pool disk requirements grow you might need to copy the entire
1352     data directory to a new (bigger) file system. Hopefully you are lucky
1353     enough to avoid this by having the data directory on a RAID file system
1354     or LVM that allows the capacity to be grown in place by adding disks.
1355    
1356     The backup data directories contain large numbers of hardlinks. If
1357     you try to copy the pool the target directory will occupy a lot more
1358     space if the hardlinks aren't re-established.
1359    
1360     The GNU cp program with the -a option is aware of hardlinks and knows
1361     to re-establish them. So GNU cp -a is the recommended way to copy
1362     the data directory and pool. Don't forget to stop BackupPC while
1363     the copy runs.
1364    
1365     =item Compressing an existing pool
1366    
1367     If you are upgrading BackupPC and want to turn compression on you have
1368     two choices:
1369    
1370     =over 4
1371    
1372     =item *
1373    
1374     Simply turn on compression. All new backups will be compressed. Both old
1375     (uncompressed) and new (compressed) backups can be browsed and viewed.
1376     Eventually, the old backups will expire and all the pool data will be
1377     compressed. However, until the old backups expire, this approach could
1378     require 60% or more additional pool storage space to store both
1379     uncompressed and compressed versions of the backup files.
1380    
1381     =item *
1382    
1383     Convert all the uncompressed pool files and backups to compressed.
1384     The script __INSTALLDIR__/bin/BackupPC_compressPool does this.
1385     BackupPC must not be running when you run BackupPC_compressPool.
1386     Also, there must be no existing compressed backups when you
1387     run BackupPC_compressPool.
1388    
1389     BackupPC_compressPool compresses all the files in the uncompressed pool
1390     (__TOPDIR__/pool) and moves them to the compressed pool
1391     (__TOPDIR__/cpool). It rewrites the files in place, so that the
1392     existing hardlinks are not disturbed.
1393    
1394     =back
1395    
1396     The rest of this section discusses how to run BackupPC_compressPool.
1397    
1398     BackupPC_compressPool takes three command line options:
1399    
1400     =over 4
1401    
1402     =item -t
1403    
1404     Test mode: do everything except actually replace the pool files.
1405     Useful for estimating total run time without making any real
1406     changes.
1407    
1408     =item -r
1409    
1410     Read check: re-read the compressed file and compare it against
1411     the original uncompressed file. Can only be used in test mode.
1412    
1413     =item -c #
1414    
1415     Number of children to fork. BackupPC_compressPool can take a long time
1416     to run, so to speed things up it spawns four children, each working on a
1417     different part of the pool. You can change the number of children with
1418     the -c option.
1419    
1420     =back
1421    
1422     Here are the recommended steps for running BackupPC_compressPool:
1423    
1424     =over 4
1425    
1426     =item *
1427    
1428     Stop BackupPC (eg: "/etc/init.d/backuppc stop").
1429    
1430     =item *
1431    
1432     Set $Conf{CompressLevel} to a non-zero number (eg: 3).
1433    
1434     =item *
1435    
1436     Do a dry run of BackupPC_compressPool. Make sure you run this as
1437     the BackupPC user (__BACKUPPCUSER__):
1438    
1439     BackupPC_compressPool -t -r
1440    
1441     The -t option (test mode) makes BackupPC_compressPool do all the steps,
1442     but not actually change anything. The -r option re-reads the compressed
1443     file and compares it against the original.
1444    
1445     BackupPC_compressPool gives a status as it completes each 1% of the job.
1446     It also shows the cumulative compression ratio and estimated completion
1447     time. Once you are comfortable that things look ok, you can kill
1448     BackupPC_compressPool or wait for it to finish.
1449    
1450     =item *
1451    
1452     Now you are ready to run BackupPC_compressPool for real. Once again,
1453     as the BackupPC user (__BACKUPPCUSER__), run:
1454    
1455     BackupPC_compressPool
1456    
1457     You should put the output into a file and tail this file. (The running
1458     time could be twice as long as the test mode since the test mode file
1459     writes are immediately followed by an unlink, so in test mode it is
1460     likely the file writes never make it to disk.)
1461    
1462     It is B<critical> that BackupPC_compressPool runs to completion before
1463     re-starting BackupPC. Before BackupPC_compressPool completes, none of
1464     the existing backups will be in a consistent state. If you must stop
1465     BackupPC_compressPool for some reason, send it an INT or TERM signal
1466     and give it several seconds (or more) to clean up gracefully.
1467     After that, you can re-run BackupPC_compressPool and it will start
1468     again where it left off. Once again, it is critical that it runs
1469     to 100% completion.
1470    
1471     =back
1472    
1473     After BackupPC_compressPool completes you should have a complete set
1474     of compressed backups (and your disk usage should be lower). You
1475     can now re-start BackupPC.
1476    
1477     =back
1478    
1479     =head2 Fixing installation problems
1480    
1481     Please see the FAQ at L<http://backuppc.sourceforge.net/faq> for
1482     debugging suggestions.
1483    
1484     =head1 Restore functions
1485    
1486     BackupPC supports several different methods for restoring files. The
1487     most convenient restore options are provided via the CGI interface.
1488     Alternatively, backup files can be restored using manual commands.
1489    
1490     =head2 CGI restore options
1491    
1492     By selecting a host in the CGI interface, a list of all the backups
1493     for that machine will be displayed. By selecting the backup number
1494     you can navigate the shares and directory tree for that backup.
1495    
1496     BackupPC's CGI interface automatically fills incremental backups
1497     with the corresponding full backup, which means each backup has
1498     a filled appearance. Therefore, there is no need to do multiple
1499     restores from the incremental and full backups: BackupPC does all
1500     the hard work for you. You simply select the files and directories
1501     you want from the correct backup vintage in one step.
1502    
1503     You can download a single backup file at any time simply by selecting
1504     it. Your browser should prompt you with the file name and ask you
1505     whether to open the file or save it to disk.
1506    
1507     Alternatively, you can select one or more files or directories in
1508     the currently selected directory and select "Restore selected files".
1509     (If you need to restore selected files and directories from several
1510     different parent directories you will need to do that in multiple
1511     steps.)
1512    
1513     If you select all the files in a directory, BackupPC will replace
1514     the list of files with the parent directory. You will be presented
1515     with a screen that has three options:
1516    
1517     =over 4
1518    
1519     =item Option 1: Direct Restore
1520    
1521     With this option the selected files and directories are restored
1522     directly back onto the host, by default in their original location.
1523     Any old files with the same name will be overwritten, so use caution.
1524     You can optionally change the target host name, target share name,
1525     and target path prefix for the restore, allowing you to restore the
1526     files to a different location.
1527    
1528     Once you select "Start Restore" you will be prompted one last time
1529     with a summary of the exact source and target files and directories
1530     before you commit. When you give the final go ahead the restore
1531     operation will be queued like a normal backup job, meaning that it
1532     will be deferred if there is a backup currently running for that host.
1533     When the restore job is run, smbclient, tar, rsync or rsyncd is used
1534     (depending upon $Conf{XferMethod}) to actually restore the files.
1535     Sorry, there is currently no option to cancel a restore that has been
1536     started.
1537    
1538     A record of the restore request, including the result and list of
1539     files and directories, is kept. It can be browsed from the host's
1540     home page. $Conf{RestoreInfoKeepCnt} specifies how many old restore
1541     status files to keep.
1542    
1543     Note that for direct restore to work, the $Conf{XferMethod} must
1544     be able to write to the client. For example, that means an SMB
1545     share for smbclient needs to be writable, and the rsyncd module
1546     needs "read only" set to "yes". This creates additional security
1547     risks. If you only create read-only SMB shares (which is a good
1548     idea), then the direct restore will fail. You can disable the
1549     direct restore option by setting $Conf{SmbClientRestoreCmd},
1550     $Conf{TarClientRestoreCmd} and $Conf{RsyncRestoreArgs} to undef.
1551    
1552     =item Option 2: Download Zip archive
1553    
1554     With this option a zip file containing the selected files and directories
1555     is downloaded. The zip file can then be unpacked or individual files
1556     extracted as necessary on the host machine. The compression level can be
1557     specified. A value of 0 turns off compression.
1558    
1559     When you select "Download Zip File" you should be prompted where to
1560     save the restore.zip file.
1561    
1562     BackupPC does not consider downloading a zip file as an actual
1563     restore operation, so the details are not saved for later browsing
1564     as in the first case. However, a mention that a zip file was
1565     downloaded by a particular user, and a list of the files, does
1566     appear in BackupPC's log file.
1567    
1568     =item Option 3: Download Tar archive
1569    
1570     This is identical to the previous option, except a tar file is downloaded
1571     rather than a zip file (and there is currently no compression option).
1572    
1573     =back
1574    
1575     =head2 Command-line restore options
1576    
1577     Apart from the CGI interface, BackupPC allows you to restore files
1578     and directories from the command line. The following programs can
1579     be used:
1580    
1581     =over 4
1582    
1583     =item BackupPC_zcat
1584    
1585     For each file name argument it inflates (uncompresses) the file and
1586     writes it to stdout. To use BackupPC_zcat you could give it the
1587     full file name, eg:
1588    
1589     __INSTALLDIR__/bin/BackupPC_zcat __TOPDIR__/pc/host/5/fc/fcraig/fexample.txt > example.txt
1590    
1591     It's your responsibility to make sure the file is really compressed:
1592     BackupPC_zcat doesn't check which backup the requested file is from.
1593     BackupPC_zcat returns a non-zero status if it fails to uncompress
1594     a file.
1595    
1596     =item BackupPC_tarCreate
1597    
1598     BackupPC_tarCreate creates a tar file for any files or directories in
1599     a particular backup. Merging of incrementals is done automatically,
1600     so you don't need to worry about whether certain files appear in the
1601     incremental or full backup.
1602    
1603     The usage is:
1604    
1605     BackupPC_tarCreate [-t] [-h host] [-n dumpNum] [-s shareName]
1606     [-r pathRemove] [-p pathAdd] [-b BLOCKS] [-w writeBufSz]
1607     files/directories...
1608    
1609     The command-line files and directories are relative to the specified
1610     shareName. The tar file is written to stdout.
1611    
1612     The required options are:
1613    
1614     =over 4
1615    
1616     =item -h host
1617    
1618     host from which the tar archive is created
1619    
1620     =item -n dumpNum
1621    
1622     dump number from which the tar archive is created
1623    
1624     =item -s shareName
1625    
1626     share name from which the tar archive is created
1627    
1628     =back
1629    
1630     Other options are:
1631    
1632     =over 4
1633    
1634     =item -t
1635    
1636     print summary totals
1637    
1638     =item -r pathRemove
1639    
1640     path prefix that will be replaced with pathAdd
1641    
1642     =item -p pathAdd
1643    
1644     new path prefix
1645    
1646     =item -b BLOCKS
1647    
1648     the tar block size, default is 20, meaning tar writes data in 20 * 512
1649     bytes chunks.
1650    
1651     =item -w writeBufSz
1652    
1653     write buffer size, default 1048576 (1MB). You can increase this if
1654     you are trying to stream to a fast tape device.
1655    
1656     =back
1657    
1658     The -h, -n and -s options specify which dump is used to generate
1659     the tar archive. The -r and -p options can be used to relocate
1660     the paths in the tar archive so extracted files can be placed
1661     in a location different from their original location.
1662    
1663     =item BackupPC_zipCreate
1664    
1665     BackupPC_zipCreate creates a zip file for any files or directories in
1666     a particular backup. Merging of incrementals is done automatically,
1667     so you don't need to worry about whether certain files appear in the
1668     incremental or full backup.
1669    
1670     The usage is:
1671    
1672     BackupPC_zipCreate [-t] [-h host] [-n dumpNum] [-s shareName]
1673     [-r pathRemove] [-p pathAdd] [-c compressionLevel]
1674     files/directories...
1675    
1676     The command-line files and directories are relative to the specified
1677     shareName. The zip file is written to stdout.
1678    
1679     The required options are:
1680    
1681     =over 4
1682    
1683     =item -h host
1684    
1685     host from which the zip archive is created
1686    
1687     =item -n dumpNum
1688    
1689     dump number from which the zip archive is created
1690    
1691     =item -s shareName
1692    
1693     share name from which the zip archive is created
1694    
1695     =back
1696    
1697     Other options are:
1698    
1699     =over 4
1700    
1701     =item -t
1702    
1703     print summary totals
1704    
1705     =item -r pathRemove
1706    
1707     path prefix that will be replaced with pathAdd
1708    
1709     =item -p pathAdd
1710    
1711     new path prefix
1712    
1713     =item -c level
1714    
1715     compression level (default is 0, no compression)
1716    
1717     =back
1718    
1719     The -h, -n and -s options specify which dump is used to generate
1720     the zip archive. The -r and -p options can be used to relocate
1721     the paths in the zip archive so extracted files can be placed
1722     in a location different from their original location.
1723    
1724     =back
1725    
1726     Each of these programs reside in __INSTALLDIR__/bin.
1727    
1728     =head1 Archive functions
1729    
1730     BackupPC supports archiving to removable media. For users that require
1731     offsite backups, BackupPC can create archives that stream to tape
1732     devices, or create files of specified sizes to fit onto cd or dvd media.
1733    
1734     Each archive type is specified by a BackupPC host with its XferMethod
1735     set to 'archive'. This allows for multiple configurations at sites where
1736     there might be a combination of tape and cd/dvd backups being made.
1737    
1738     BackupPC provides a menu that allows one or more hosts to be archived.
1739     The most recent backup of each host is archived using BackupPC_tarCreate,
1740     and the output is optionally compressed and split into fixed-sized
1741     files (eg: 650MB).
1742    
1743     The archive for each host is done by default using
1744     __INSTALLDIR__/BackupPC_archiveHost. This script can be copied
1745     and customized as needed.
1746    
1747     =head2 Configuring an Archive Host
1748    
1749     To create an Archive Host, add it to the hosts file just as any other host
1750     and call it a name that best describes the type of archive, e.g. ArchiveDLT
1751    
1752     To tell BackupPC that the Host is for Archives, create a config.pl file in
1753     the Archive Hosts's pc directory, adding the following line:
1754    
1755     $Conf{XferMethod} = 'archive';
1756    
1757     To further customise the archive's parameters you can adding the changed
1758     parameters in the host's config.pl file. The parameters are explained in
1759     the config.pl file. Parameters may be fixed or the user can be allowed
1760     to change them (eg: output device).
1761    
1762     The per-host archive command is $Conf{ArchiveClientCmd}. By default
1763     this invokes
1764    
1765     __INSTALLDIR__/BackupPC_archiveHost
1766    
1767     which you can copy and customize as necessary.
1768    
1769     =head2 Starting an Archive
1770    
1771     In the web interface, click on the Archive Host you wish to use. You will see a
1772     list of previous archives and a summary on each. By clicking the "Start Archive"
1773     button you are presented with the list of hosts and the approximate backup size
1774     (note this is raw size, not projected compressed size) Select the hosts you wish
1775     to archive and press the "Archive Selected Hosts" button.
1776    
1777     The next screen allows you to adjust the parameters for this archive run.
1778     Press the "Start the Archive" to start archiving the selected hosts with the
1779     parameters displayed.
1780    
1781     =head1 BackupPC Design
1782    
1783     =head2 Some design issues
1784    
1785     =over 4
1786    
1787     =item Pooling common files
1788    
1789     To quickly see if a file is already in the pool, an MD5 digest of the
1790     file length and contents is used as the file name in the pool. This
1791     can't guarantee a file is identical: it just reduces the search to
1792     often a single file or handful of files. A complete file comparison
1793     is always done to verify if two files are really the same.
1794    
1795     Identical files on multiples backups are represented by hard links.
1796     Hardlinks are used so that identical files all refer to the same
1797     physical file on the server's disk. Also, hard links maintain
1798     reference counts so that BackupPC knows when to delete unused files
1799     from the pool.
1800    
1801     For the computer-science majors among you, you can think of the pooling
1802     system used by BackupPC as just a chained hash table stored on a (big)
1803     file system.
1804    
1805     =item The hashing function
1806    
1807     There is a tradeoff between how much of file is used for the MD5 digest
1808     and the time taken comparing all the files that have the same hash.
1809    
1810     Using the file length and just the first 4096 bytes of the file for the
1811     MD5 digest produces some repetitions. One example: with 900,000 unique
1812     files in the pool, this hash gives about 7,000 repeated files, and in
1813     the worst case 500 files have the same hash. That's not bad: we only
1814     have to do a single file compare 99.2% of the time. But in the worst
1815     case we have to compare as many as 500 files checking for a match.
1816    
1817     With a modest increase in CPU time, if we use the file length and the
1818     first 256K of the file we now only have 500 repeated files and in the
1819     worst case around 20 files have the same hash. Furthermore, if we
1820     instead use the first and last 128K of the file (more specifically, the
1821     first and eighth 128K chunks for files larger than 1MB) we get only 300
1822     repeated files and in the worst case around 20 files have the same hash.
1823    
1824     Based on this experimentation, this is the hash function used by BackupPC.
1825     It is important that you don't change the hash function after files
1826     are already in the pool. Otherwise your pool will grow to twice the
1827     size until all the old backups (and all the old files with old hashes)
1828     eventually expire.
1829    
1830     =item Compression
1831    
1832     BackupPC supports compression. It uses the deflate and inflate methods
1833     in the Compress::Zlib module, which is based on the zlib compression
1834     library (see L<http://www.gzip.org/zlib/>).
1835    
1836     The $Conf{CompressLevel} setting specifies the compression level to use.
1837     Zero (0) means no compression. Compression levels can be from 1 (least
1838     cpu time, slightly worse compression) to 9 (most cpu time, slightly
1839     better compression). The recommended value is 3. Changing it to 5, for
1840     example, will take maybe 20% more cpu time and will get another 2-3%
1841     additional compression. Diminishing returns set in above 5. See the zlib
1842     documentation for more information about compression levels.
1843    
1844     BackupPC implements compression with minimal CPU load. Rather than
1845     compressing every incoming backup file and then trying to match it
1846     against the pool, BackupPC computes the MD5 digest based on the
1847     uncompressed file, and matches against the candidate pool files by
1848     comparing each uncompressed pool file against the incoming backup file.
1849     Since inflating a file takes roughly a factor of 10 less CPU time than
1850     deflating there is a big saving in CPU time.
1851    
1852     The combination of pooling common files and compression can yield
1853     a factor of 8 or more overall saving in backup storage.
1854    
1855     =back
1856    
1857     =head2 BackupPC operation
1858    
1859     BackupPC reads the configuration information from
1860     __TOPDIR__/conf/config.pl. It then runs and manages all the backup
1861     activity. It maintains queues of pending backup requests, user backup
1862     requests and administrative commands. Based on the configuration various
1863     requests will be executed simultaneously.
1864    
1865     As specified by $Conf{WakeupSchedule}, BackupPC wakes up periodically
1866     to queue backups on all the PCs. This is a four step process:
1867    
1868     =over 4
1869    
1870     =item 1
1871    
1872     For each host and DHCP address backup requests are queued on the
1873     background command queue.
1874    
1875     =item 2
1876    
1877     For each PC, BackupPC_dump is forked. Several of these may be run in
1878     parallel, based on the configuration. First a ping is done to see if
1879     the machine is alive. If this is a DHCP address, nmblookup is run to
1880     get the netbios name, which is used as the host name. If DNS lookup
1881     fails, $Conf{NmbLookupFindHostCmd} is run to find the IP address from
1882     the host name. The file __TOPDIR__/pc/$host/backups is read to decide
1883     whether a full or incremental backup needs to be run. If no backup is
1884     scheduled, or the ping to $host fails, then BackupPC_dump exits.
1885    
1886     The backup is done using the specified XferMethod. Either samba's smbclient
1887     or tar over ssh/rsh/nfs piped into BackupPC_tarExtract, or rsync over ssh/rsh
1888     is run, or rsyncd is connected to, with the incoming data
1889     extracted to __TOPDIR__/pc/$host/new. The XferMethod output is put
1890     into __TOPDIR__/pc/$host/XferLOG.
1891    
1892     The letter in the XferLOG file shows the type of object, similar to the
1893     first letter of the modes displayed by ls -l:
1894    
1895     d -> directory
1896     l -> symbolic link
1897     b -> block special file
1898     c -> character special file
1899     p -> pipe file (fifo)
1900     nothing -> regular file
1901    
1902     The words mean:
1903    
1904     =over 4
1905    
1906     =item create
1907    
1908     new for this backup (ie: directory or file not in pool)
1909    
1910     =item pool
1911    
1912     found a match in the pool
1913    
1914     =item same
1915    
1916     file is identical to previous backup (contents were
1917     checksummed and verified during full dump).
1918    
1919     =item skip
1920    
1921     file skipped in incremental because attributes are the
1922     same (only displayed if $Conf{XferLogLevel} >= 2).
1923    
1924     =back
1925    
1926     As BackupPC_tarExtract extracts the files from smbclient or tar, or as
1927     rsync runs, it checks each file in the backup to see if it is identical
1928     to an existing file from any previous backup of any PC. It does this
1929     without needed to write the file to disk. If the file matches an
1930     existing file, a hardlink is created to the existing file in the pool.
1931     If the file does not match any existing files, the file is written to
1932     disk and the file name is saved in __TOPDIR__/pc/$host/NewFileList for
1933     later processing by BackupPC_link. BackupPC_tarExtract and rsync can handle
1934     arbitrarily large files and multiple candidate matching files without
1935     needing to write the file to disk in the case of a match. This
1936     significantly reduces disk writes (and also reads, since the pool file
1937     comparison is done disk to memory, rather than disk to disk).
1938    
1939     Based on the configuration settings, BackupPC_dump checks each
1940     old backup to see if any should be removed. Any expired backups
1941     are moved to __TOPDIR__/trash for later removal by BackupPC_trashClean.
1942    
1943     =item 3
1944    
1945     For each complete, good, backup, BackupPC_link is run.
1946     To avoid race conditions as new files are linked into the
1947     pool area, only a single BackupPC_link program runs
1948     at a time and the rest are queued.
1949    
1950     BackupPC_link reads the NewFileList written by BackupPC_dump and
1951     inspects each new file in the backup. It re-checks if there is a
1952     matching file in the pool (another BackupPC_link
1953     could have added the file since BackupPC_dump checked). If so, the file
1954     is removed and replaced by a hard link to the existing file. If the file
1955     is new, a hard link to the file is made in the pool area, so that this
1956     file is available for checking against each new file and new backup.
1957    
1958     Then, if $Conf{IncrFill} is set (note that the default setting is
1959     off), for each incremental backup, hard links are made in the new
1960     backup to all files that were not extracted during the incremental
1961     backups. The means the incremental backup looks like a complete
1962     image of the PC (with the exception that files that were removed on
1963     the PC since the last full backup will still appear in the backup
1964     directory tree).
1965    
1966     The CGI interface knows how to merge unfilled incremental backups will
1967     the most recent prior filled (full) backup, giving the incremental
1968     backups a filled appearance. The default for $Conf{IncrFill} is off,
1969     since there is no need to fill incremental backups. This saves
1970     some level of disk activity, since lots of extra hardlinks are no
1971     longer needed (and don't have to be deleted when the backup expires).
1972    
1973     =item 4
1974    
1975     BackupPC_trashClean is always run in the background to remove any
1976     expired backups. Every 5 minutes it wakes up and removes all the files
1977     in __TOPDIR__/trash.
1978    
1979     Also, once each night, BackupPC_nightly is run to complete some additional
1980     administrative tasks, such as cleaning the pool. This involves removing
1981     any files in the pool that only have a single hard link (meaning no backups
1982     are using that file). Again, to avoid race conditions, BackupPC_nightly
1983     is only run when there are no BackupPC_dump or BackupPC_link processes
1984     running. Therefore, when it is time to run BackupPC_nightly, no new
1985     backups are started and BackupPC waits until all backups have finished.
1986     Then BackupPC_nightly is run, and until it finishes no new backups are
1987     started. If BackupPC_nightly is slow, the settings
1988    
1989     =back
1990    
1991     BackupPC also listens for TCP connections on $Conf{ServerPort}, which
1992     is used by the CGI script BackupPC_Admin for status reporting and
1993     user-initiated backup or backup cancel requests.
1994    
1995     =head2 Storage layout
1996    
1997     BackupPC resides in three directories:
1998    
1999     =over 4
2000    
2001     =item __INSTALLDIR__
2002    
2003     Perl scripts comprising BackupPC reside in __INSTALLDIR__/bin,
2004     libraries are in __INSTALLDIR__/lib and documentation
2005     is in __INSTALLDIR__/doc.
2006    
2007     =item __CGIDIR__
2008    
2009     The CGI script BackupPC_Admin resides in this cgi binary directory.
2010    
2011     =item __TOPDIR__
2012    
2013     All of BackupPC's data (PC backup images, logs, configuration information)
2014     is stored below this directory.
2015    
2016     =back
2017    
2018     Below __TOPDIR__ are several directories:
2019    
2020     =over 4
2021    
2022     =item __TOPDIR__/conf
2023    
2024     The directory __TOPDIR__/conf contains:
2025    
2026     =over 4
2027    
2028     =item config.pl
2029    
2030     Configuration file. See L<Configuration file|configuration file>
2031     below for more details.
2032    
2033     =item hosts
2034    
2035     Hosts file, which lists all the PCs to backup.
2036    
2037     =back
2038    
2039     =item __TOPDIR__/log
2040    
2041     The directory __TOPDIR__/log contains:
2042    
2043     =over 4
2044    
2045     =item LOG
2046    
2047     Current (today's) log file output from BackupPC.
2048    
2049     =item LOG.0 or LOG.0.z
2050    
2051     Yesterday's log file output. Log files are aged daily and compressed
2052     (if compression is enabled), and old LOG files are deleted.
2053    
2054     =item BackupPC.pid
2055    
2056     Contains BackupPC's process id.
2057    
2058     =item status.pl
2059    
2060     A summary of BackupPC's status written periodically by BackupPC so
2061     that certain state information can be maintained if BackupPC is
2062     restarted. Should not be edited.
2063    
2064     =item UserEmailInfo.pl
2065    
2066     A summary of what email was last sent to each user, and when the
2067     last email was sent. Should not be edited.
2068    
2069     =back
2070    
2071     =item __TOPDIR__/trash
2072    
2073     Any directories and files below this directory are periodically deleted
2074     whenever BackupPC_trashClean checks. When a backup is aborted or when an
2075     old backup expires, BackupPC_dump simply moves the directory to
2076     __TOPDIR__/trash for later removal by BackupPC_trashClean.
2077    
2078     =item __TOPDIR__/pool
2079    
2080     All uncompressed files from PC backups are stored below __TOPDIR__/pool.
2081     Each file's name is based on the MD5 hex digest of the file contents.
2082     Specifically, for files less than 256K, the file length and the entire
2083     file is used. For files up to 1MB, the file length and the first and
2084     last 128K are used. Finally, for files longer than 1MB, the file length,
2085     and the first and eighth 128K chunks for the file are used.
2086    
2087     Each file is stored in a subdirectory X/Y/Z, where X, Y, Z are the
2088     first 3 hex digits of the MD5 digest.
2089    
2090     For example, if a file has an MD5 digest of 123456789abcdef0,
2091     the file is stored in __TOPDIR__/pool/1/2/3/123456789abcdef0.
2092    
2093     The MD5 digest might not be unique (especially since not all the file's
2094     contents are used for files bigger than 256K). Different files that have
2095     the same MD5 digest are stored with a trailing suffix "_n" where n is
2096     an incrementing number starting at 0. So, for example, if two additional
2097     files were identical to the first, except the last byte was different,
2098     and assuming the file was larger than 1MB (so the MD5 digests are the
2099     same but the files are actually different), the three files would be
2100     stored as:
2101    
2102     __TOPDIR__/pool/1/2/3/123456789abcdef0
2103     __TOPDIR__/pool/1/2/3/123456789abcdef0_0
2104     __TOPDIR__/pool/1/2/3/123456789abcdef0_1
2105    
2106     Both BackupPC_dump (actually, BackupPC_tarExtract) and BackupPC_link are
2107     responsible for checking newly backed up files against the pool. For
2108     each file, the MD5 digest is used to generate a file name in the pool
2109     directory. If the file exists in the pool, the contents are compared.
2110     If there is no match, additional files ending in "_n" are checked.
2111     (Actually, BackupPC_tarExtract compares multiple candidate files in
2112     parallel.) If the file contents exactly match, the file is created by
2113     simply making a hard link to the pool file (this is done by
2114     BackupPC_tarExtract as the backup proceeds). Otherwise,
2115     BackupPC_tarExtract writes the new file to disk and a new hard link is
2116     made in the pool to the file (this is done later by BackupPC_link).
2117    
2118     Therefore, every file in the pool will have at least 2 hard links
2119     (one for the pool file and one for the backup file below __TOPDIR__/pc).
2120     Identical files from different backups or PCs will all be linked to
2121     the same file. When old backups are deleted, some files in the pool
2122     might only have one link. BackupPC_nightly checks the entire pool
2123     and removes all files that have only a single link, thereby recovering
2124     the storage for that file.
2125    
2126     One other issue: zero length files are not pooled, since there are a lot
2127     of these files and on most file systems it doesn't save any disk space
2128     to turn these files into hard links.
2129    
2130     =item __TOPDIR__/cpool
2131    
2132     All compressed files from PC backups are stored below __TOPDIR__/cpool.
2133     Its layout is the same as __TOPDIR__/pool, and the hashing function
2134     is the same (and, importantly, based on the uncompressed file, not
2135     the compressed file).
2136    
2137     =item __TOPDIR__/pc/$host
2138    
2139     For each PC $host, all the backups for that PC are stored below
2140     the directory __TOPDIR__/pc/$host. This directory contains the
2141     following files:
2142    
2143     =over 4
2144    
2145     =item LOG
2146    
2147     Current log file for this PC from BackupPC_dump.
2148    
2149     =item LOG.0 or LOG.0.z
2150    
2151     Last month's log file. Log files are aged monthly and compressed
2152     (if compression is enabled), and old LOG files are deleted.
2153    
2154     =item XferERR or XferERR.z
2155    
2156     Output from the transport program (ie: smbclient, tar or rsync)
2157     for the most recent failed backup.
2158    
2159     =item new
2160    
2161     Subdirectory in which the current backup is stored. This
2162     directory is renamed if the backup succeeds.
2163    
2164     =item XferLOG or XferLOG.z
2165    
2166     Output from the transport program (ie: smbclient, tar or rsync)
2167     for the current backup.
2168    
2169     =item nnn (an integer)
2170    
2171     Successful backups are in directories numbered sequentially starting at 0.
2172    
2173     =item XferLOG.nnn or XferLOG.nnn.z
2174    
2175     Output from the transport program (ie: smbclient, tar or rsync)
2176     corresponding to backup number nnn.
2177    
2178     =item RestoreInfo.nnn
2179    
2180     Information about restore request #nnn including who, what, when, and
2181     why. This file is in Data::Dumper format. (Note that the restore
2182     numbers are not related to the backup number.)
2183    
2184     =item RestoreLOG.nnn.z
2185    
2186     Output from smbclient, tar or rsync during restore #nnn. (Note that the restore
2187     numbers are not related to the backup number.)
2188    
2189     =item ArchiveInfo.nnn
2190    
2191     Information about archive request #nnn including who, what, when, and
2192     why. This file is in Data::Dumper format. (Note that the archive
2193     numbers are not related to the restore or backup number.)
2194    
2195     =item ArchiveLOG.nnn.z
2196    
2197     Output from archive #nnn. (Note that the archive numbers are not related
2198     to the backup or restore number.)
2199    
2200     =item config.pl
2201    
2202     Optional configuration settings specific to this host. Settings in this
2203     file override the main configuration file.
2204    
2205     =item backups
2206    
2207     A tab-delimited ascii table listing information about each successful
2208     backup, one per row. The columns are:
2209    
2210     =over 4
2211    
2212     =item num
2213    
2214     The backup number, an integer that starts at 0 and increments
2215     for each successive backup. The corresponding backup is stored
2216     in the directory num (eg: if this field is 5, then the backup is
2217     stored in __TOPDIR__/pc/$host/5).
2218    
2219     =item type
2220    
2221     Set to "full" or "incr" for full or incremental backup.
2222    
2223     =item startTime
2224    
2225     Start time of the backup in unix seconds.
2226    
2227     =item endTime
2228    
2229     Stop time of the backup in unix seconds.
2230    
2231     =item nFiles
2232    
2233     Number of files backed up (as reported by smbclient, tar or rsync).
2234    
2235     =item size
2236    
2237     Total file size backed up (as reported by smbclient, tar or rsync).
2238    
2239     =item nFilesExist
2240    
2241     Number of files that were already in the pool
2242     (as determined by BackupPC_dump and BackupPC_link).
2243    
2244     =item sizeExist
2245    
2246     Total size of files that were already in the pool
2247     (as determined by BackupPC_dump and BackupPC_link).
2248    
2249     =item nFilesNew
2250    
2251     Number of files that were not in the pool
2252     (as determined by BackupPC_link).
2253    
2254     =item sizeNew
2255    
2256     Total size of files that were not in the pool
2257     (as determined by BackupPC_link).
2258    
2259     =item xferErrs
2260    
2261     Number of errors or warnings from smbclient, tar or rsync.
2262    
2263     =item xferBadFile
2264    
2265     Number of errors from smbclient that were bad file errors (zero otherwise).
2266    
2267     =item xferBadShare
2268    
2269     Number of errors from smbclient that were bad share errors (zero otherwise).
2270    
2271     =item tarErrs
2272    
2273     Number of errors from BackupPC_tarExtract.
2274    
2275     =item compress
2276    
2277     The compression level used on this backup. Zero or empty means no
2278     compression.
2279    
2280     =item sizeExistComp
2281    
2282     Total compressed size of files that were already in the pool
2283     (as determined by BackupPC_dump and BackupPC_link).
2284    
2285     =item sizeNewComp
2286    
2287     Total compressed size of files that were not in the pool
2288     (as determined by BackupPC_link).
2289    
2290     =item noFill
2291    
2292     Set if this backup has not been filled in with the most recent
2293     previous filled or full backup. See $Conf{IncrFill}.
2294    
2295     =item fillFromNum
2296    
2297     If this backup was filled (ie: noFill is 0) then this is the
2298     number of the backup that it was filled from
2299    
2300     =item mangle
2301    
2302     Set if this backup has mangled file names and attributes. Always
2303     true for backups in v1.4.0 and above. False for all backups prior
2304     to v1.4.0.
2305    
2306     =item xferMethod
2307    
2308     Set to the value of $Conf{XferMethod} when this dump was done.
2309    
2310     =item level
2311    
2312     The level of this dump. A full dump is level 0. Currently incrementals
2313     are 1. But when multi-level incrementals are supported this will reflect
2314     each dump's incremental level.
2315    
2316     =back
2317    
2318     =item restores
2319    
2320     A tab-delimited ascii table listing information about each requested
2321     restore, one per row. The columns are:
2322    
2323     =over 4
2324    
2325     =item num
2326    
2327     Restore number (matches the suffix of the RestoreInfo.nnn and
2328     RestoreLOG.nnn.z file), unrelated to the backup number.
2329    
2330     =item startTime
2331    
2332     Start time of the restore in unix seconds.
2333    
2334     =item endTime
2335    
2336     End time of the restore in unix seconds.
2337    
2338     =item result
2339    
2340     Result (ok or failed).
2341    
2342     =item errorMsg
2343    
2344     Error message if restore failed.
2345    
2346     =item nFiles
2347    
2348     Number of files restored.
2349    
2350     =item size
2351    
2352     Size in bytes of the restored files.
2353    
2354     =item tarCreateErrs
2355    
2356     Number of errors from BackupPC_tarCreate during restore.
2357    
2358     =item xferErrs
2359    
2360     Number of errors from smbclient, tar or rsync during restore.
2361    
2362     =back
2363    
2364     =item archives
2365    
2366     A tab-delimited ascii table listing information about each requested
2367     archive, one per row. The columns are:
2368    
2369     =over 4
2370    
2371     =item num
2372    
2373     Archive number (matches the suffix of the ArchiveInfo.nnn and
2374     ArchiveLOG.nnn.z file), unrelated to the backup or restore number.
2375    
2376     =item startTime
2377    
2378     Start time of the restore in unix seconds.
2379    
2380     =item endTime
2381    
2382     End time of the restore in unix seconds.
2383    
2384     =item result
2385    
2386     Result (ok or failed).
2387    
2388     =item errorMsg
2389    
2390     Error message if archive failed.
2391    
2392     =back
2393    
2394     =back
2395    
2396     =back
2397    
2398     =head2 Compressed file format
2399    
2400     The compressed file format is as generated by Compress::Zlib::deflate
2401     with one minor, but important, tweak. Since Compress::Zlib::inflate
2402     fully inflates its argument in memory, it could take large amounts of
2403     memory if it was inflating a highly compressed file. For example, a
2404     200MB file of 0x0 bytes compresses to around 200K bytes. If
2405     Compress::Zlib::inflate was called with this single 200K buffer, it
2406     would need to allocate 200MB of memory to return the result.
2407    
2408     BackupPC watches how efficiently a file is compressing. If a big file
2409     has very high compression (meaning it will use too much memory when it
2410     is inflated), BackupPC calls the flush() method, which gracefully
2411     completes the current compression. BackupPC then starts another
2412     deflate and simply appends the output file. So the BackupPC compressed
2413     file format is one or more concatenated deflations/flushes. The specific
2414     ratios that BackupPC uses is that if a 6MB chunk compresses to less
2415     than 64K then a flush will be done.
2416    
2417     Back to the example of the 200MB file of 0x0 bytes. Adding flushes
2418     every 6MB adds only 200 or so bytes to the 200K output. So the
2419     storage cost of flushing is negligible.
2420    
2421     To easily decompress a BackupPC compressed file, the script
2422     BackupPC_zcat can be found in __INSTALLDIR__/bin. For each
2423     file name argument it inflates the file and writes it to stdout.
2424    
2425     =head2 Rsync checksum caching
2426    
2427     An incremental backup with rsync compares attributes on the client
2428     with the last full backup. Any files with identical attributes
2429     are skipped. A full backup with rsync sets the --ignore-times
2430     option, which causes every file to be examined independent of
2431     attributes.
2432    
2433     Each file is examined by generating block checksums (default 2K
2434     blocks) on the receiving side (that's the BackupPC side), sending
2435     those checksums to the client, where the remote rsync matches those
2436     checksums with the corresponding file. The matching blocks and new
2437     data is sent back, allowing the client file to be reassembled.
2438     A checksum for the entire file is sent to as an extra check the
2439     the reconstructed file is correct.
2440    
2441     This results in significant disk IO and computation for BackupPC:
2442     every file in a full backup, or any file with non-matching attributes
2443     in an incremental backup, needs to be uncompressed, block checksums
2444     computed and sent. Then the receiving side reassembles the file and
2445     has to verify the whole-file checksum. Even if the file is identical,
2446     prior to 2.1.0, BackupPC had to read and uncompress the file twice,
2447     once to compute the block checksums and later to verify the whole-file
2448     checksum.
2449    
2450     Starting in 2.1.0, BackupPC supports optional checksum caching,
2451     which means the block and file checksums only need to be computed
2452     once for each file. This results in a significant performance
2453     improvement. This only works for compressed pool files.
2454     It is enabled by adding
2455    
2456     '--checksum-seed=32761',
2457    
2458     to $Conf{RsyncArgs} and $Conf{RsyncRestoreArgs}.
2459    
2460     Rsync versions prior to and including rsync-2.6.2 need a small patch to
2461     add support for the --checksum-seed option. This patch is available in
2462     the cygwin-rsyncd package at L<http://backuppc.sourceforge.net>.
2463     This patch is already included in rsync CVS, so it will be standard
2464     in future versions of rsync.
2465    
2466     When this option is present, BackupPC will add block and file checksums
2467     to the compressed pool file the next time a pool file is used and it
2468     doesn't already have cached checksums. The first time a new file is
2469     written to the pool, the checksums are not appended. The next time
2470     checksums are needed for a file, they are computed and added. So the
2471     full performance benefit of checksum caching won't be noticed until the
2472     third time a pool file is used (eg: the third full backup).
2473    
2474     With checksum caching enabled, there is a risk that should a file's contents
2475     in the pool be corrupted due to a disk problem, but the cached checksums
2476     are still correct, the corruption will not be detected by a full backup,
2477     since the file contents are no longer read and compared. To reduce the
2478     chance that this remains undetected, BackupPC can recheck cached checksums
2479     for a fraction of the files. This fraction is set with the
2480     $Conf{RsyncCsumCacheVerifyProb} setting. The default value of 0.01 means
2481     that 1% of the time a file's checksums are read, the checksums are verified.
2482     This reduces performance slightly, but, over time, ensures that files
2483     contents are in sync with the cached checksums.
2484    
2485     The format of the cached checksum data can be discovered by looking at
2486     the code. Basically, the first byte of the compressed file is changed
2487     to denote that checksums are appended. The block and file checksum
2488     data, plus some other information and magic word, are appended to the
2489     compressed file. This allows the cache update to be done in-place.
2490    
2491     =head2 File name mangling
2492    
2493     Backup file names are stored in "mangled" form. Each node of
2494     a path is preceded by "f" (mnemonic: file), and special characters
2495     (\n, \r, % and /) are URI-encoded as "%xx", where xx is the ascii
2496     character's hex value. So c:/craig/example.txt is now stored as
2497     fc/fcraig/fexample.txt.
2498    
2499     This was done mainly so meta-data could be stored alongside the backup
2500     files without name collisions. In particular, the attributes for the
2501     files in a directory are stored in a file called "attrib", and mangling
2502     avoids file name collisions (I discarded the idea of having a duplicate
2503     directory tree for every backup just to store the attributes). Other
2504     meta-data (eg: rsync checksums) could be stored in file names preceded
2505     by, eg, "c". There are two other benefits to mangling: the share name
2506     might contain "/" (eg: "/home/craig" for tar transport), and I wanted
2507     that represented as a single level in the storage tree. Secondly, as
2508     files are written to NewFileList for later processing by BackupPC_link,
2509     embedded newlines in the file's path will cause problems which are
2510     avoided by mangling.
2511    
2512     The CGI script undoes the mangling, so it is invisible to the user.
2513     Old (unmangled) backups are still supported by the CGI
2514     interface.
2515    
2516     =head2 Special files
2517    
2518     Linux/unix file systems support several special file types: symbolic
2519     links, character and block device files, fifos (pipes) and unix-domain
2520     sockets. All except unix-domain sockets are supported by BackupPC
2521     (there's no point in backing up or restoring unix-domain sockets since
2522     they only have meaning after a process creates them). Symbolic links are
2523     stored as a plain file whose contents are the contents of the link (not
2524     the file it points to). This file is compressed and pooled like any
2525     normal file. Character and block device files are also stored as plain
2526     files, whose contents are two integers separated by a comma; the numbers
2527     are the major and minor device number. These files are compressed and
2528     pooled like any normal file. Fifo files are stored as empty plain files
2529     (which are not pooled since they have zero size). In all cases, the
2530     original file type is stored in the attrib file so it can be correctly
2531     restored.
2532    
2533     Hardlinks are also supported. When GNU tar first encounters a file with
2534     more than one link (ie: hardlinks) it dumps it as a regular file. When
2535     it sees the second and subsequent hardlinks to the same file, it dumps
2536     just the hardlink information. BackupPC correctly recognizes these
2537     hardlinks and stores them just like symlinks: a regular text file
2538     whose contents is the path of the file linked to. The CGI script
2539     will download the original file when you click on a hardlink.
2540    
2541     Also, BackupPC_tarCreate has enough magic to re-create the hardlinks
2542     dynamically based on whether or not the original file and hardlinks
2543     are both included in the tar file. For example, imagine a/b/x is a
2544     hardlink to a/c/y. If you use BackupPC_tarCreate to restore directory
2545     a, then the tar file will include a/b/x as the original file and a/c/y
2546     will be a hardlink to a/b/x. If, instead you restore a/c, then the
2547     tar file will include a/c/y as the original file, not a hardlink.
2548    
2549     =head2 Attribute file format
2550    
2551     The unix attributes for the contents of a directory (all the files and
2552     directories in that directory) are stored in a file called attrib.
2553     There is a single attrib file for each directory in a backup.
2554     For example, if c:/craig contains a single file c:/craig/example.txt,
2555     that file would be stored as fc/fcraig/fexample.txt and there would be an
2556     attribute file in fc/fcraig/attrib (and also fc/attrib and ./attrib).
2557     The file fc/fcraig/attrib would contain a single entry containing the
2558     attributes for fc/fcraig/fexample.txt.
2559    
2560     The attrib file starts with a magic number, followed by the
2561     concatenation of the following information for each file:
2562    
2563     =over 4
2564    
2565     =item *
2566    
2567     File name length in perl's pack "w" format (variable length base 128).
2568    
2569     =item *
2570    
2571     File name.
2572    
2573     =item *
2574    
2575     The unix file type, mode, uid, gid and file size divided by 4GB and
2576     file size modulo 4GB (type mode uid gid sizeDiv4GB sizeMod4GB),
2577     in perl's pack "w" format (variable length base 128).
2578    
2579     =item *
2580    
2581     The unix mtime (unix seconds) in perl's pack "N" format (32 bit integer).
2582    
2583     =back
2584    
2585     The attrib file is also compressed if compression is enabled.
2586     See the lib/BackupPC/Attrib.pm module for full details.
2587    
2588     Attribute files are pooled just like normal backup files. This saves
2589     space if all the files in a directory have the same attributes across
2590     multiple backups, which is common.
2591    
2592     =head2 Optimizations
2593    
2594     BackupPC doesn't care about the access time of files in the pool
2595     since it saves attribute meta-data separate from the files. Since
2596     BackupPC mostly does reads from disk, maintaining the access time of
2597     files generates a lot of unnecessary disk writes. So, provided
2598     BackupPC has a dedicated data disk, you should consider mounting
2599     BackupPC's data directory with the noatime attribute (see mount(1)).
2600    
2601     =head2 Limitations
2602    
2603     BackupPC isn't perfect (but it is getting better). Please see
2604     L<http://backuppc.sourceforge.net/faq/limitations.html> for a
2605     discussion of some of BackupPC's limitations.
2606    
2607     =head2 Security issues
2608    
2609     Please see L<http://backuppc.sourceforge.net/faq/security.html> for a
2610     discussion of some of various security issues.
2611    
2612     =head1 Configuration File
2613    
2614     The BackupPC configuration file resides in __TOPDIR__/conf/config.pl.
2615     Optional per-PC configuration files reside in __TOPDIR__/pc/$host/config.pl.
2616     This file can be used to override settings just for a particular PC.
2617    
2618     =head2 Modifying the main configuration file
2619    
2620     The configuration file is a perl script that is executed by BackupPC, so
2621     you should be careful to preserve the file syntax (punctuation, quotes
2622     etc) when you edit it. It is recommended that you use CVS, RCS or some
2623     other method of source control for changing config.pl.
2624    
2625     BackupPC reads or re-reads the main configuration file and
2626     the hosts file in three cases:
2627    
2628     =over 4
2629    
2630     =item *
2631    
2632     Upon startup.
2633    
2634     =item *
2635    
2636     When BackupPC is sent a HUP (-1) signal. Assuming you installed the
2637     init.d script, you can also do this with "/etc/init.d/backuppc reload".
2638    
2639     =item *
2640    
2641     When the modification time of config.pl file changes. BackupPC
2642     checks the modification time once during each regular wakeup.
2643    
2644     =back
2645    
2646     Whenever you change the configuration file you can either do
2647     a kill -HUP BackupPC_pid or simply wait until the next regular
2648     wakeup period.
2649    
2650     Each time the configuration file is re-read a message is reported in the
2651     LOG file, so you can tail it (or view it via the CGI interface) to make
2652     sure your kill -HUP worked. Errors in parsing the configuration file are
2653     also reported in the LOG file.
2654    
2655     The optional per-PC configuration file (__TOPDIR__/pc/$host/config.pl)
2656     is read whenever it is needed by BackupPC_dump, BackupPC_link and others.
2657    
2658     =head2 Configuration file includes
2659    
2660     If you have a heterogeneous set of clients (eg: a variety of WinXX and
2661     linux/unix machines) you will need to create host-specific config.pl files
2662     for some or all of these machines to customize the default settings from
2663     the master config.pl file (at a minimum to set $Conf{XferMethod}).
2664    
2665     Since the config.pl file is just regular perl code, you can include
2666     one config file from another. For example, imagine you had three general
2667     classes of machines: WinXX desktops, linux machines in the DMZ and
2668     linux desktops. You could create three config files in __TOPDIR__/conf:
2669    
2670     __TOPDIR__/conf/ConfigWinDesktop.pl
2671     __TOPDIR__/conf/ConfigLinuxDMZ.pl
2672     __TOPDIR__/conf/ConfigLinuxDesktop.pl
2673    
2674     From each client's directory you can either add a symbolic link to
2675     the appropriate config file:
2676    
2677     cd __TOPDIR__/pc/$host
2678     ln -s ../../conf/ConfigWinDesktop.pl config.pl
2679    
2680     or, better yet, create a config.pl file in __TOPDIR__/pc/$host
2681     that includes the default config.pl file using perl's "do"
2682     command:
2683    
2684     do "__TOPDIR__/conf/ConfigWinDesktop.pl";
2685    
2686     This alternative allows you to set other configuration options
2687     specific to each host after the "do" command (perhaps even
2688     overriding the settings in the included file).
2689    
2690     Note that you could also include snippets of configuration settings
2691     from the main configuration file. However, be aware that the
2692     modification-time checking that BackupPC does only applies to the
2693     main configuration file: if you change one of the included files,
2694     BackupPC won't notice. You will need to either touch the main
2695     configuration file too, or send BackupPC a HUP (-1) signal.
2696    
2697     =head1 Configuration Parameters
2698    
2699     The configuration parameters are divided into five general groups.
2700     The first group (general server configuration) provides general
2701     configuration for BackupPC. The next two groups describe what to
2702     backup, when to do it, and how long to keep it. The fourth group
2703     are settings for email reminders, and the final group contains
2704     settings for the CGI interface.
2705    
2706     All configuration settings in the second through fifth groups can
2707     be overridden by the per-PC config.pl file.
2708    
2709     =head2 General server configuration
2710    
2711     =over 4
2712    
2713     =item $Conf{ServerHost} = '';
2714    
2715     Host name on which the BackupPC server is running.
2716    
2717     =item $Conf{ServerPort} = -1;
2718    
2719     TCP port number on which the BackupPC server listens for and accepts
2720     connections. Normally this should be disabled (set to -1). The TCP
2721     port is only needed if apache runs on a different machine from BackupPC.
2722     In that case, set this to any spare port number over 1024 (eg: 2359).
2723     If you enable the TCP port, make sure you set $Conf{ServerMesgSecret}
2724     too!
2725    
2726     =item $Conf{ServerMesgSecret} = '';
2727    
2728     Shared secret to make the TCP port secure. Set this to a hard to guess
2729     string if you enable the TCP port (ie: $Conf{ServerPort} > 0).
2730    
2731     To avoid possible attacks via the TCP socket interface, every client
2732     message is protected by an MD5 digest. The MD5 digest includes four
2733     items:
2734     - a seed that is sent to the client when the connection opens
2735     - a sequence number that increments for each message
2736     - a shared secret that is stored in $Conf{ServerMesgSecret}
2737     - the message itself.
2738    
2739     The message is sent in plain text preceded by the MD5 digest. A
2740     snooper can see the plain-text seed sent by BackupPC and plain-text
2741     message from the client, but cannot construct a valid MD5 digest since
2742     the secret $Conf{ServerMesgSecret} is unknown. A replay attack is
2743     not possible since the seed changes on a per-connection and
2744     per-message basis.
2745    
2746     =item $Conf{MyPath} = '/bin';
2747    
2748     PATH setting for BackupPC. An explicit value is necessary
2749     for taint mode. Value shouldn't matter too much since
2750     all execs use explicit paths. However, taint mode in perl
2751     will complain if this directory is world writable.
2752    
2753     =item $Conf{UmaskMode} = 027;
2754    
2755     Permission mask for directories and files created by BackupPC.
2756     Default value prevents any access from group other, and prevents
2757     group write.
2758    
2759     =item $Conf{WakeupSchedule} = [1..23];
2760    
2761     Times at which we wake up, check all the PCs, and schedule necessary
2762     backups. Times are measured in hours since midnight. Can be
2763     fractional if necessary (eg: 4.25 means 4:15am).
2764    
2765     If the hosts you are backing up are always connected to the network
2766     you might have only one or two wakeups each night. This will keep
2767     the backup activity after hours. On the other hand, if you are backing
2768     up laptops that are only intermittently connected to the network you
2769     will want to have frequent wakeups (eg: hourly) to maximized the chance
2770     that each laptop is backed up.
2771    
2772     Examples:
2773    
2774     $Conf{WakeupSchedule} = [22.5]; # once per day at 10:30 pm.
2775     $Conf{WakeupSchedule} = [1..23]; # every hour except midnight
2776     $Conf{WakeupSchedule} = [2,4,6,8,10,12,14,16,18,20,22]; # every 2 hours
2777    
2778     The default value is every hour except midnight.
2779    
2780     The first entry of $Conf{WakeupSchedule} is when BackupPC_nightly
2781     is run. No other backups can run while BackupPC_nightly is
2782     running. You might want to re-arrange the entries in
2783     $Conf{WakeupSchedule} (they don't have to be ascending) so that
2784     the first entry is when you want BackupPC_nightly to run
2785     (eg: when you don't expect a lot of regular backups to run).
2786    
2787     =item $Conf{MaxBackups} = 4;
2788    
2789     Maximum number of simultaneous backups to run. If there
2790     are no user backup requests then this is the maximum number
2791     of simultaneous backups.
2792    
2793     =item $Conf{MaxUserBackups} = 4;
2794    
2795     Additional number of simultaneous backups that users can run.
2796     As many as $Conf{MaxBackups} + $Conf{MaxUserBackups} requests can
2797     run at the same time.
2798    
2799     =item $Conf{MaxPendingCmds} = 10;
2800    
2801     Maximum number of pending link commands. New backups will only be
2802     started if there are no more than $Conf{MaxPendingCmds} plus
2803     $Conf{MaxBackups} number of pending link commands, plus running jobs.
2804     This limit is to make sure BackupPC doesn't fall too far behind in
2805     running BackupPC_link commands.
2806    
2807     =item $Conf{MaxBackupPCNightlyJobs} = 2;
2808    
2809     How many BackupPC_nightly processes to run in parallel.
2810    
2811     Each night, at the first wakeup listed in $Conf{WakeupSchedule},
2812     BackupPC_nightly is run. Its job is to remove unneeded files
2813     in the pool, ie: files that only have one link. To avoid race
2814     conditions, BackupPC_nightly runs only when there are no backups
2815     running, and no backups will start while it runs.
2816    
2817     So to reduce the elapsed time, you might want to increase this
2818     setting to run several BackupPC_nightly processes in parallel
2819     (eg: 4, or even 8).
2820    
2821     =item $Conf{BackupPCNightlyPeriod} = 1;
2822    
2823     How many days (runs) it takes BackupPC_nightly to traverse the
2824     entire pool. Normally this is 1, which means every night it runs,
2825     it does traverse the entire pool removing unused pool files.
2826    
2827     Other valid values are 2, 4, 8, 16. This causes BackupPC_nightly to
2828     traverse 1/2, 1/4, 1/8 or 1/16th of the pool each night, meaning it
2829     takes 2, 4, 8 or 16 days to completely traverse the pool. The
2830     advantage is that each night the running time of BackupPC_nightly
2831     is reduced roughly in proportion, since the total job is split
2832     over multiple days. The disadvantage is that unused pool files
2833     take longer to get deleted, which will slightly increase disk
2834     usage.
2835    
2836     Note that even when $Conf{BackupPCNightlyPeriod} > 1, BackupPC_nightly
2837     still runs every night. It just does less work each time it runs.
2838    
2839     Examples:
2840    
2841    
2842     $Conf{BackupPCNightlyPeriod} = 1; # entire pool is checked every night
2843    
2844     $Conf{BackupPCNightlyPeriod} = 2; # two days to complete pool check
2845     # (different half each night)
2846    
2847     $Conf{BackupPCNightlyPeriod} = 4; # four days to complete pool check
2848     # (different quarter each night)
2849    
2850     =item $Conf{MaxOldLogFiles} = 14;
2851    
2852     Maximum number of log files we keep around in log directory.
2853     These files are aged nightly. A setting of 14 means the log
2854     directory will contain about 2 weeks of old log files, in
2855     particular at most the files LOG, LOG.0, LOG.1, ... LOG.13
2856     (except today's LOG, these files will have a .z extension if
2857     compression is on).
2858    
2859     If you decrease this number after BackupPC has been running for a
2860     while you will have to manually remove the older log files.
2861    
2862     =item $Conf{DfPath} = '/bin/df';
2863    
2864     Full path to the df command. Security caution: normal users
2865     should not allowed to write to this file or directory.
2866    
2867     =item $Conf{DfCmd} = '$dfPath $topDir';
2868    
2869     Command to run df. The following variables are substituted at run-time:
2870    
2871     $dfPath path to df ($Conf{DfPath})
2872     $topDir top-level BackupPC data directory
2873    
2874     =item $Conf{DfMaxUsagePct} = 95;
2875    
2876     Maximum threshold for disk utilization on the __TOPDIR__ filesystem.
2877     If the output from $Conf{DfPath} reports a percentage larger than
2878     this number then no new regularly scheduled backups will be run.
2879     However, user requested backups (which are usually incremental and
2880     tend to be small) are still performed, independent of disk usage.
2881     Also, currently running backups will not be terminated when the disk
2882     usage exceeds this number.
2883    
2884     =item $Conf{TrashCleanSleepSec} = 300;
2885    
2886     How long BackupPC_trashClean sleeps in seconds between each check
2887     of the trash directory. Once every 5 minutes should be reasonable.
2888    
2889     =item $Conf{DHCPAddressRanges} = [];
2890    
2891     List of DHCP address ranges we search looking for PCs to backup.
2892     This is an array of hashes for each class C address range.
2893     This is only needed if hosts in the conf/hosts file have the
2894     dhcp flag set.
2895    
2896     Examples:
2897    
2898     # to specify 192.10.10.20 to 192.10.10.250 as the DHCP address pool
2899     $Conf{DHCPAddressRanges} = [
2900     {
2901     ipAddrBase => '192.10.10',
2902     first => 20,
2903     last => 250,
2904     },
2905     ];
2906     # to specify two pools (192.10.10.20-250 and 192.10.11.10-50)
2907     $Conf{DHCPAddressRanges} = [
2908     {
2909     ipAddrBase => '192.10.10',
2910     first => 20,
2911     last => 250,
2912     },
2913     {
2914     ipAddrBase => '192.10.11',
2915     first => 10,
2916     last => 50,
2917     },
2918     ];
2919    
2920     =item $Conf{BackupPCUser} = '';
2921    
2922     =item $Conf{CgiDir} = '';
2923    
2924     =item $Conf{InstallDir} = '';
2925    
2926     These configuration settings aren't used by BackupPC, but simply
2927     remember a few settings used by configure.pl during installation.
2928     These are used by configure.pl when upgrading to new versions of
2929     BackupPC.
2930    
2931     =item $Conf{BackupPCUserVerify} = 1;
2932    
2933     Whether BackupPC and the CGI script BackupPC_Admin verify that they
2934     are really running as user $Conf{BackupPCUser}. If this flag is set
2935     and the effective user id (euid) differs from $Conf{BackupPCUser}
2936     then both scripts exit with an error. This catches cases where
2937     BackupPC might be accidently started as root or the wrong user,
2938     or if the CGI script is not installed correctly.
2939    
2940     =item $Conf{HardLinkMax} = 31999;
2941    
2942     Maximum number of hardlinks supported by the $TopDir file system
2943     that BackupPC uses. Most linux or unix file systems should support
2944     at least 32000 hardlinks per file, or 64000 in other cases. If a pool
2945     file already has this number of hardlinks, a new pool file is created
2946     so that new hardlinks can be accommodated. This limit will only
2947     be hit if an identical file appears at least this number of times
2948     across all the backups.
2949    
2950     =back
2951    
2952     =head2 What to backup and when to do it
2953    
2954     =over 4
2955    
2956     =item $Conf{SmbShareName} = 'C$';
2957    
2958     Name of the host share that is backed up when using SMB. This can be a
2959     string or an array of strings if there are multiple shares per host.
2960     Examples:
2961    
2962    
2963     $Conf{SmbShareName} = 'c'; # backup 'c' share
2964     $Conf{SmbShareName} = ['c', 'd']; # backup 'c' and 'd' shares
2965    
2966     This setting only matters if $Conf{XferMethod} = 'smb'.
2967    
2968     =item $Conf{SmbShareUserName} = '';
2969    
2970     Smbclient share user name. This is passed to smbclient's -U argument.
2971    
2972     This setting only matters if $Conf{XferMethod} = 'smb'.
2973    
2974     =item $Conf{SmbSharePasswd} = '';
2975    
2976     Smbclient share password. This is passed to smbclient via its PASSWD
2977     environment variable. There are several ways you can tell BackupPC
2978     the smb share password. In each case you should be very careful about
2979     security. If you put the password here, make sure that this file is
2980     not readable by regular users! See the "Setting up config.pl" section
2981     in the documentation for more information.
2982    
2983     This setting only matters if $Conf{XferMethod} = 'smb'.
2984    
2985     =item $Conf{TarShareName} = '/';
2986    
2987     Which host directories to backup when using tar transport. This can be a
2988     string or an array of strings if there are multiple directories to
2989     backup per host. Examples:
2990    
2991    
2992     $Conf{TarShareName} = '/'; # backup everything
2993     $Conf{TarShareName} = '/home'; # only backup /home
2994     $Conf{TarShareName} = ['/home', '/src']; # backup /home and /src
2995    
2996     The fact this parameter is called 'TarShareName' is for historical
2997     consistency with the Smb transport options. You can use any valid
2998     directory on the client: there is no need for it to correspond to
2999     any Smb share or device mount point.
3000    
3001     Note also that you can also use $Conf{BackupFilesOnly} to specify
3002     a specific list of directories to backup. It's more efficient to
3003     use this option instead of $Conf{TarShareName} since a new tar is
3004     run for each entry in $Conf{TarShareName}.
3005    
3006     On the other hand, if you add --one-file-system to $Conf{TarClientCmd}
3007     you can backup each file system separately, which makes restoring one
3008     bad file system easier. In this case you would list all of the mount
3009     points here, since you can't get the same result with
3010     $Conf{BackupFilesOnly}:
3011    
3012     $Conf{TarShareName} = ['/', '/var', '/data', '/boot'];
3013    
3014     This setting only matters if $Conf{XferMethod} = 'tar'.
3015    
3016     =item $Conf{FullPeriod} = 6.97;
3017    
3018     Minimum period in days between full backups. A full dump will only be
3019     done if at least this much time has elapsed since the last full dump,
3020     and at least $Conf{IncrPeriod} days has elapsed since the last
3021     successful dump.
3022    
3023     Typically this is set slightly less than an integer number of days. The
3024     time taken for the backup, plus the granularity of $Conf{WakeupSchedule}
3025     will make the actual backup interval a bit longer.
3026    
3027     There are two special values for $Conf{FullPeriod}:
3028    
3029     -1 Don't do any regular backups on this machine. Manually
3030     requested backups (via the CGI interface) will still occur.
3031    
3032     -2 Don't do any backups on this machine. Manually requested
3033     backups (via the CGI interface) will be ignored.
3034    
3035     These special settings are useful for a client that is no longer
3036     being backed up (eg: a retired machine), but you wish to keep the
3037     last backups available for browsing or restoring to other machines.
3038    
3039     =item $Conf{IncrPeriod} = 0.97;
3040    
3041     Minimum period in days between incremental backups (a user requested
3042     incremental backup will be done anytime on demand).
3043    
3044     Typically this is set slightly less than an integer number of days. The
3045     time taken for the backup, plus the granularity of $Conf{WakeupSchedule}
3046     will make the actual backup interval a bit longer.
3047    
3048     =item $Conf{FullKeepCnt} = 1;
3049    
3050     Number of full backups to keep. Must be >= 1.
3051    
3052     In the steady state, each time a full backup completes successfully
3053     the oldest one is removed. If this number is decreased, the
3054     extra old backups will be removed.
3055    
3056     If filling of incremental dumps is off the oldest backup always
3057     has to be a full (ie: filled) dump. This might mean one or two
3058     extra full dumps are kept until the oldest incremental backups expire.
3059    
3060     Exponential backup expiry is also supported. This allows you to specify:
3061    
3062     - num fulls to keep at intervals of 1 * $Conf{FullPeriod}, followed by
3063     - num fulls to keep at intervals of 2 * $Conf{FullPeriod},
3064     - num fulls to keep at intervals of 4 * $Conf{FullPeriod},
3065     - num fulls to keep at intervals of 8 * $Conf{FullPeriod},
3066     - num fulls to keep at intervals of 16 * $Conf{FullPeriod},
3067    
3068     and so on. This works by deleting every other full as each expiry
3069     boundary is crossed.
3070    
3071     Exponential expiry is specified using an array for $Conf{FullKeepCnt}:
3072    
3073     $Conf{FullKeepCnt} = [4, 2, 3];
3074    
3075     Entry #n specifies how many fulls to keep at an interval of
3076     2^n * $Conf{FullPeriod} (ie: 1, 2, 4, 8, 16, 32, ...).
3077    
3078     The example above specifies keeping 4 of the most recent full backups
3079     (1 week interval) two full backups at 2 week intervals, and 3 full
3080     backups at 4 week intervals, eg:
3081    
3082     full 0 19 weeks old \
3083     full 1 15 weeks old >--- 3 backups at 4 * $Conf{FullPeriod}
3084     full 2 11 weeks old /
3085     full 3 7 weeks old \____ 2 backups at 2 * $Conf{FullPeriod}
3086     full 4 5 weeks old /
3087     full 5 3 weeks old \
3088     full 6 2 weeks old \___ 4 backups at 1 * $Conf{FullPeriod}
3089     full 7 1 week old /
3090     full 8 current /
3091    
3092     On a given week the spacing might be less than shown as each backup
3093     ages through each expiry period. For example, one week later, a
3094     new full is completed and the oldest is deleted, giving:
3095    
3096     full 0 16 weeks old \
3097     full 1 12 weeks old >--- 3 backups at 4 * $Conf{FullPeriod}
3098     full 2 8 weeks old /
3099     full 3 6 weeks old \____ 2 backups at 2 * $Conf{FullPeriod}
3100     full 4 4 weeks old /
3101     full 5 3 weeks old \
3102     full 6 2 weeks old \___ 4 backups at 1 * $Conf{FullPeriod}
3103     full 7 1 week old /
3104     full 8 current /
3105    
3106     You can specify 0 as a count (except in the first entry), and the
3107     array can be as long as you wish. For example:
3108    
3109    
3110     $Conf{FullKeepCnt} = [4, 0, 4, 0, 0, 2];
3111    
3112     This will keep 10 full dumps, 4 most recent at 1 * $Conf{FullPeriod},
3113     followed by 4 at an interval of 4 * $Conf{FullPeriod} (approx 1 month
3114     apart), and then 2 at an interval of 32 * $Conf{FullPeriod} (approx
3115     7-8 months apart).
3116    
3117     Example: these two settings are equivalent and both keep just
3118     the four most recent full dumps:
3119    
3120     $Conf{FullKeepCnt} = 4;
3121     $Conf{FullKeepCnt} = [4];
3122    
3123     =item $Conf{FullKeepCntMin} = 1;
3124    
3125     =item $Conf{FullAgeMax} = 90;
3126    
3127     Very old full backups are removed after $Conf{FullAgeMax} days. However,
3128     we keep at least $Conf{FullKeepCntMin} full backups no matter how old
3129     they are.
3130    
3131     Note that $Conf{FullAgeMax} will be increased to $Conf{FullAgeMax}
3132     times $Conf{FullPeriod} if $Conf{FullAgeMax} specifies enough
3133     full backups to exceed $Conf{FullAgeMax}.
3134    
3135     =item $Conf{IncrKeepCnt} = 6;
3136    
3137     Number of incremental backups to keep. Must be >= 1.
3138    
3139     In the steady state, each time an incr backup completes successfully
3140     the oldest one is removed. If this number is decreased, the
3141     extra old backups will be removed.
3142    
3143     =item $Conf{IncrKeepCntMin} = 1;
3144    
3145     =item $Conf{IncrAgeMax} = 30;
3146    
3147     Very old incremental backups are removed after $Conf{IncrAgeMax} days.
3148     However, we keep at least $Conf{IncrKeepCntMin} incremental backups no
3149     matter how old they are.
3150    
3151     =item $Conf{PartialAgeMax} = 3;
3152    
3153     A failed full backup is saved as a partial backup. The rsync
3154     XferMethod can take advantage of the partial full when the next
3155     backup is run. This parameter sets the age of the partial full
3156     in days: if the partial backup is older than this number of
3157     days, then rsync will ignore (not use) the partial full when
3158     the next backup is run. If you set this to a negative value
3159     then no partials will be saved. If you set this to 0, partials
3160     will be saved, but will not be used by the next backup.
3161    
3162     The default setting of 3 days means that a partial older than
3163     3 days is ignored when the next full backup is done.
3164    
3165     =item $Conf{IncrFill} = 0;
3166    
3167     Whether incremental backups are filled. "Filling" means that the
3168     most recent full (or filled) dump is merged into the new incremental
3169     dump using hardlinks. This makes an incremental dump look like a
3170     full dump. Prior to v1.03 all incremental backups were filled.
3171     In v1.4.0 and later the default is off.
3172    
3173     BackupPC, and the cgi interface in particular, do the right thing on
3174     un-filled incremental backups. It will correctly display the merged
3175     incremental backup with the most recent filled backup, giving the
3176     un-filled incremental backups a filled appearance. That means it
3177     invisible to the user whether incremental dumps are filled or not.
3178    
3179     Filling backups takes a little extra disk space, and it does cost
3180     some extra disk activity for filling, and later removal. Filling
3181     is no longer useful, since file mangling and compression doesn't
3182     make a filled backup very useful. It's likely the filling option
3183     will be removed from future versions: filling will be delegated to
3184     the display and extraction of backup data.
3185    
3186     If filling is off, BackupPC makes sure that the oldest backup is
3187     a full, otherwise the following incremental backups will be
3188     incomplete. This might mean an extra full backup has to be
3189     kept until the following incremental backups expire.
3190    
3191     The default is off. You can turn this on or off at any
3192     time without affecting existing backups.
3193    
3194     =item $Conf{RestoreInfoKeepCnt} = 10;
3195    
3196     Number of restore logs to keep. BackupPC remembers information about
3197     each restore request. This number per client will be kept around before
3198     the oldest ones are pruned.
3199    
3200     Note: files/dirs delivered via Zip or Tar downloads don't count as
3201     restores. Only the first restore option (where the files and dirs
3202     are written to the host) count as restores that are logged.
3203    
3204     =item $Conf{ArchiveInfoKeepCnt} = 10;
3205    
3206     Number of archive logs to keep. BackupPC remembers information
3207     about each archive request. This number per archive client will
3208     be kept around before the oldest ones are pruned.
3209    
3210     =item $Conf{BackupFilesOnly} = undef;
3211    
3212     List of directories or files to backup. If this is defined, only these
3213     directories or files will be backed up.
3214    
3215     For Smb, only one of $Conf{BackupFilesExclude} and $Conf{BackupFilesOnly}
3216     can be specified per share. If both are set for a particular share, then
3217     $Conf{BackupFilesOnly} takes precedence and $Conf{BackupFilesExclude}
3218     is ignored.
3219    
3220     This can be set to a string, an array of strings, or, in the case
3221     of multiple shares, a hash of strings or arrays. A hash is used
3222     to give a list of directories or files to backup for each share
3223     (the share name is the key). If this is set to just a string or
3224     array, and $Conf{SmbShareName} contains multiple share names, then
3225     the setting is assumed to apply all shares.
3226    
3227     Examples:
3228    
3229     $Conf{BackupFilesOnly} = '/myFiles';
3230     $Conf{BackupFilesOnly} = ['/myFiles']; # same as first example
3231     $Conf{BackupFilesOnly} = ['/myFiles', '/important'];
3232     $Conf{BackupFilesOnly} = {
3233     'c' => ['/myFiles', '/important'], # these are for 'c' share
3234     'd' => ['/moreFiles', '/archive'], # these are for 'd' share
3235     };
3236    
3237     =item $Conf{BackupFilesExclude} = undef;
3238    
3239     List of directories or files to exclude from the backup. For Smb,
3240     only one of $Conf{BackupFilesExclude} and $Conf{BackupFilesOnly}
3241     can be specified per share. If both are set for a particular share,
3242     then $Conf{BackupFilesOnly} takes precedence and
3243     $Conf{BackupFilesExclude} is ignored.
3244    
3245     This can be set to a string, an array of strings, or, in the case
3246     of multiple shares, a hash of strings or arrays. A hash is used
3247     to give a list of directories or files to exclude for each share
3248     (the share name is the key). If this is set to just a string or
3249     array, and $Conf{SmbShareName} contains multiple share names, then
3250     the setting is assumed to apply to all shares.
3251    
3252     The exact behavior is determined by the underlying transport program,
3253     smbclient or tar. For smbclient the exlclude file list is passed into
3254     the X option. Simple shell wild-cards using "*" or "?" are allowed.
3255    
3256     For tar, if the exclude file contains a "/" it is assumed to be anchored
3257     at the start of the string. Since all the tar paths start with "./",
3258     BackupPC prepends a "." if the exclude file starts with a "/". Note
3259     that GNU tar version >= 1.13.7 is required for the exclude option to
3260     work correctly. For linux or unix machines you should add
3261     "/proc" to $Conf{BackupFilesExclude} unless you have specified
3262     --one-file-system in $Conf{TarClientCmd} or --one-file-system in
3263     $Conf{RsyncArgs}. Also, for tar, do not use a trailing "/" in
3264     the directory name: a trailing "/" causes the name to not match
3265     and the directory will not be excluded.
3266    
3267     Users report that for smbclient you should specify a directory
3268     followed by "/*", eg: "/proc/*", instead of just "/proc".
3269    
3270     Examples:
3271    
3272     $Conf{BackupFilesExclude} = '/temp';
3273     $Conf{BackupFilesExclude} = ['/temp']; # same as first example
3274     $Conf{BackupFilesExclude} = ['/temp', '/winnt/tmp'];
3275     $Conf{BackupFilesExclude} = {
3276     'c' => ['/temp', '/winnt/tmp'], # these are for 'c' share
3277     'd' => ['/junk', '/dont_back_this_up'], # these are for 'd' share
3278     };
3279    
3280     =item $Conf{BlackoutBadPingLimit} = 3;
3281    
3282     =item $Conf{BlackoutGoodCnt} = 7;
3283    
3284     PCs that are always or often on the network can be backed up after
3285     hours, to reduce PC, network and server load during working hours. For
3286     each PC a count of consecutive good pings is maintained. Once a PC has
3287     at least $Conf{BlackoutGoodCnt} consecutive good pings it is subject
3288     to "blackout" and not backed up during hours and days specified by
3289     $Conf{BlackoutPeriods}.
3290    
3291     To allow for periodic rebooting of a PC or other brief periods when a
3292     PC is not on the network, a number of consecutive bad pings is allowed
3293     before the good ping count is reset. This parameter is
3294     $Conf{BlackoutBadPingLimit}.
3295    
3296     Note that bad and good pings don't occur with the same interval. If a
3297     machine is always on the network, it will only be pinged roughly once
3298     every $Conf{IncrPeriod} (eg: once per day). So a setting for
3299     $Conf{BlackoutGoodCnt} of 7 means it will take around 7 days for a
3300     machine to be subject to blackout. On the other hand, if a ping is
3301     failed, it will be retried roughly every time BackupPC wakes up, eg,
3302     every one or two hours. So a setting for $Conf{BlackoutBadPingLimit} of
3303     3 means that the PC will lose its blackout status after 3-6 hours of
3304     unavailability.
3305    
3306     To disable the blackout feature set $Conf{BlackoutGoodCnt} to a negative
3307     value. A value of 0 will make all machines subject to blackout. But
3308     if you don't want to do any backups during the day it would be easier
3309     to just set $Conf{WakeupSchedule} to a restricted schedule.
3310    
3311     =item $Conf{BlackoutPeriods} = [ ... ];
3312    
3313     One or more blackout periods can be specified. If a client is
3314     subject to blackout then no regular (non-manual) backups will
3315     be started during any of these periods. hourBegin and hourEnd
3316     specify hours fro midnight and weekDays is a list of days of
3317     the week where 0 is Sunday, 1 is Monday etc.
3318    
3319     For example:
3320    
3321    
3322     $Conf{BlackoutPeriods} = [
3323     {
3324     hourBegin => 7.0,
3325     hourEnd => 19.5,
3326     weekDays => [1, 2, 3, 4, 5],
3327     },
3328     ];
3329    
3330     specifies one blackout period from 7:00am to 7:30pm local time
3331     on Mon-Fri.
3332    
3333     The blackout period can also span midnight by setting
3334     hourBegin > hourEnd, eg:
3335    
3336     $Conf{BlackoutPeriods} = [
3337     {
3338     hourBegin => 7.0,
3339     hourEnd => 19.5,
3340     weekDays => [1, 2, 3, 4, 5],
3341     },
3342     {
3343     hourBegin => 23,
3344     hourEnd => 5,
3345     weekDays => [5, 6],
3346     },
3347     ];
3348    
3349     This specifies one blackout period from 7:00am to 7:30pm local time
3350     on Mon-Fri, and a second period from 11pm to 5am on Friday and
3351     Saturday night.
3352    
3353     =item $Conf{BackupZeroFilesIsFatal} = 1;
3354    
3355     A backup of a share that has zero files is considered fatal. This is
3356     used to catch miscellaneous Xfer errors that result in no files being
3357     backed up. If you have shares that might be empty (and therefore an
3358     empty backup is valid) you should set this flag to 0.
3359    
3360     =back
3361    
3362     =head2 General per-PC configuration settings
3363    
3364     =over 4
3365    
3366     =item $Conf{XferMethod} = 'smb';
3367    
3368     What transport method to use to backup each host. If you have
3369     a mixed set of WinXX and linux/unix hosts you will need to override
3370     this in the per-PC config.pl.
3371    
3372     The valid values are:
3373    
3374     - 'smb': backup and restore via smbclient and the SMB protocol.
3375     Easiest choice for WinXX.
3376    
3377     - 'rsync': backup and restore via rsync (via rsh or ssh).
3378     Best choice for linux/unix. Good choice also for WinXX.
3379    
3380     - 'rsyncd': backup and restre via rsync daemon on the client.
3381     Best choice for linux/unix if you have rsyncd running on
3382     the client. Good choice also for WinXX.
3383    
3384     - 'tar': backup and restore via tar, tar over ssh, rsh or nfs.
3385     Good choice for linux/unix.
3386    
3387     - 'archive': host is a special archive host. Backups are not done.
3388     An archive host is used to archive other host's backups
3389     to permanent media, such as tape, CDR or DVD.
3390    
3391    
3392     =item $Conf{XferLogLevel} = 1;
3393    
3394     Level of verbosity in Xfer log files. 0 means be quiet, 1 will give
3395     will give one line per file, 2 will also show skipped files on
3396     incrementals, higher values give more output.
3397    
3398     =item $Conf{SmbClientPath} = '/usr/bin/smbclient';
3399    
3400     Full path for smbclient. Security caution: normal users should not
3401     allowed to write to this file or directory.
3402    
3403     smbclient is from the Samba distribution. smbclient is used to
3404     actually extract the incremental or full dump of the share filesystem
3405     from the PC.
3406    
3407     This setting only matters if $Conf{XferMethod} = 'smb'.
3408    
3409     =item $Conf{SmbClientFullCmd} = '$smbClientPath \\\\$host\\$shareName' ...
3410    
3411     Command to run smbclient for a full dump.
3412     This setting only matters if $Conf{XferMethod} = 'smb'.
3413    
3414     The following variables are substituted at run-time:
3415    
3416     $smbClientPath same as $Conf{SmbClientPath}
3417     $host host to backup/restore
3418     $hostIP host IP address
3419     $shareName share name
3420     $userName user name
3421     $fileList list of files to backup (based on exclude/include)
3422     $I_option optional -I option to smbclient
3423     $X_option exclude option (if $fileList is an exclude list)
3424     $timeStampFile start time for incremental dump
3425    
3426     =item $Conf{SmbClientIncrCmd} = '$smbClientPath \\\\$host\\$shareName' ...
3427    
3428     Command to run smbclient for an incremental dump.
3429     This setting only matters if $Conf{XferMethod} = 'smb'.
3430    
3431     Same variable substitutions are applied as $Conf{SmbClientFullCmd}.
3432    
3433     =item $Conf{SmbClientRestoreCmd} = '$smbClientPath \\\\$host\\$shareName' ...
3434    
3435     Command to run smbclient for a restore.
3436     This setting only matters if $Conf{XferMethod} = 'smb'.
3437    
3438     Same variable substitutions are applied as $Conf{SmbClientFullCmd}.
3439    
3440     If your smb share is read-only then direct restores will fail.
3441     You should set $Conf{SmbClientRestoreCmd} to undef and the
3442     corresponding CGI restore option will be removed.
3443    
3444     =item $Conf{TarClientCmd} = '$sshPath -q -x -n -l root $host' ...
3445    
3446     Full command to run tar on the client. GNU tar is required. You will
3447     need to fill in the correct paths for ssh2 on the local host (server)
3448     and GNU tar on the client. Security caution: normal users should not
3449     allowed to write to these executable files or directories.
3450    
3451     See the documentation for more information about setting up ssh2 keys.
3452    
3453     If you plan to use NFS then tar just runs locally and ssh2 is not needed.
3454     For example, assuming the client filesystem is mounted below /mnt/hostName,
3455     you could use something like:
3456    
3457     $Conf{TarClientCmd} = '$tarPath -c -v -f - -C /mnt/$host/$shareName'
3458     . ' --totals';
3459    
3460     In the case of NFS or rsh you need to make sure BackupPC's privileges
3461     are sufficient to read all the files you want to backup. Also, you
3462     will probably want to add "/proc" to $Conf{BackupFilesExclude}.
3463    
3464     The following variables are substituted at run-time:
3465    
3466     $host host name
3467     $hostIP host's IP address
3468     $incrDate newer-than date for incremental backups
3469     $shareName share name to backup (ie: top-level directory path)
3470     $fileList specific files to backup or exclude
3471     $tarPath same as $Conf{TarClientPath}
3472     $sshPath same as $Conf{SshPath}
3473    
3474     If a variable is followed by a "+" it is shell escaped. This is
3475     necessary for the command part of ssh or rsh, since it ends up
3476     getting passed through the shell.
3477    
3478     This setting only matters if $Conf{XferMethod} = 'tar'.
3479    
3480     =item $Conf{TarFullArgs} = '$fileList+';
3481    
3482     Extra tar arguments for full backups. Several variables are substituted at
3483     run-time. See $Conf{TarClientCmd} for the list of variable substitutions.
3484    
3485     If you are running tar locally (ie: without rsh or ssh) then remove the
3486     "+" so that the argument is no longer shell escaped.
3487    
3488     This setting only matters if $Conf{XferMethod} = 'tar'.
3489    
3490     =item $Conf{TarIncrArgs} = '--newer=$incrDate+ $fileList+';
3491    
3492     Extra tar arguments for incr backups. Several variables are substituted at
3493     run-time. See $Conf{TarClientCmd} for the list of variable substitutions.
3494    
3495     Note that GNU tar has several methods for specifying incremental backups,
3496     including:
3497    
3498     --newer-mtime $incrDate+
3499     This causes a file to be included if the modification time is
3500     later than $incrDate (meaning its contents might have changed).
3501     But changes in the ownership or modes will not qualify the
3502     file to be included in an incremental.
3503    
3504     --newer=$incrDate+
3505     This causes the file to be included if any attribute of the
3506     file is later than $incrDate, meaning either attributes or
3507     the modification time. This is the default method. Do
3508     not use --atime-preserve in $Conf{TarClientCmd} above,
3509     otherwise resetting the atime (access time) counts as an
3510     attribute change, meaning the file will always be included
3511     in each new incremental dump.
3512    
3513     If you are running tar locally (ie: without rsh or ssh) then remove the
3514     "+" so that the argument is no longer shell escaped.
3515    
3516     This setting only matters if $Conf{XferMethod} = 'tar'.
3517    
3518     =item $Conf{TarClientRestoreCmd} = '$sshPath -q -x -l root $host' ...
3519    
3520     Full command to run tar for restore on the client. GNU tar is required.
3521     This can be the same as $Conf{TarClientCmd}, with tar's -c replaced by -x
3522     and ssh's -n removed.
3523    
3524     See $Conf{TarClientCmd} for full details.
3525    
3526     This setting only matters if $Conf{XferMethod} = "tar".
3527    
3528     If you want to disable direct restores using tar, you should set
3529     $Conf{TarClientRestoreCmd} to undef and the corresponding CGI
3530     restore option will be removed.
3531    
3532     =item $Conf{TarClientPath} = '/bin/tar';
3533    
3534     Full path for tar on the client. Security caution: normal users should not
3535     allowed to write to this file or directory.
3536    
3537     This setting only matters if $Conf{XferMethod} = 'tar'.
3538    
3539     =item $Conf{RsyncClientPath} = '/bin/rsync';
3540    
3541     Path to rsync executable on the client
3542    
3543     =item $Conf{RsyncClientCmd} = '$sshPath -q -x -l root $host $rsyncPath $argList+';
3544    
3545     Full command to run rsync on the client machine. The following variables
3546     are substituted at run-time:
3547    
3548     $host host name being backed up
3549     $hostIP host's IP address
3550     $shareName share name to backup (ie: top-level directory path)
3551     $rsyncPath same as $Conf{RsyncClientPath}
3552     $sshPath same as $Conf{SshPath}
3553     $argList argument list, built from $Conf{RsyncArgs},
3554     $shareName, $Conf{BackupFilesExclude} and
3555     $Conf{BackupFilesOnly}
3556    
3557     This setting only matters if $Conf{XferMethod} = 'rsync'.
3558    
3559     =item $Conf{RsyncClientRestoreCmd} = '$sshPath -q -x -l root $host $rsyncPath $argList+';
3560    
3561     Full command to run rsync for restore on the client. The following
3562     variables are substituted at run-time:
3563    
3564     $host host name being backed up
3565     $hostIP host's IP address
3566     $shareName share name to backup (ie: top-level directory path)
3567     $rsyncPath same as $Conf{RsyncClientPath}
3568     $sshPath same as $Conf{SshPath}
3569     $argList argument list, built from $Conf{RsyncArgs},
3570     $shareName, $Conf{BackupFilesExclude} and
3571     $Conf{BackupFilesOnly}
3572    
3573     This setting only matters if $Conf{XferMethod} = 'rsync'.
3574    
3575     =item $Conf{RsyncShareName} = '/';
3576    
3577     Share name to backup. For $Conf{XferMethod} = "rsync" this should
3578     be a file system path, eg '/' or '/home'.
3579    
3580     For $Conf{XferMethod} = "rsyncd" this should be the name of the module
3581     to backup (ie: the name from /etc/rsynd.conf).
3582    
3583     This can also be a list of multiple file system paths or modules.
3584     For example, by adding --one-file-system to $Conf{RsyncArgs} you
3585     can backup each file system separately, which makes restoring one
3586     bad file system easier. In this case you would list all of the mount
3587     points:
3588    
3589     $Conf{RsyncShareName} = ['/', '/var', '/data', '/boot'];
3590    
3591     =item $Conf{RsyncdClientPort} = 873;
3592    
3593     Rsync daemon port on the client, for $Conf{XferMethod} = "rsyncd".
3594    
3595     =item $Conf{RsyncdUserName} = '';
3596    
3597     Rsync daemon user name on client, for $Conf{XferMethod} = "rsyncd".
3598     The user name and password are stored on the client in whatever file
3599     the "secrets file" parameter in rsyncd.conf points to
3600     (eg: /etc/rsyncd.secrets).
3601    
3602     =item $Conf{RsyncdPasswd} = '';
3603    
3604     Rsync daemon user name on client, for $Conf{XferMethod} = "rsyncd".
3605     The user name and password are stored on the client in whatever file
3606     the "secrets file" parameter in rsyncd.conf points to
3607     (eg: /etc/rsyncd.secrets).
3608    
3609     =item $Conf{RsyncdAuthRequired} = 1;
3610    
3611     Whether authentication is mandatory when connecting to the client's
3612     rsyncd. By default this is on, ensuring that BackupPC will refuse to
3613     connect to an rsyncd on the client that is not password protected.
3614     Turn off at your own risk.
3615    
3616     =item $Conf{RsyncCsumCacheVerifyProb} = 0.01;
3617    
3618     When rsync checksum caching is enabled (by adding the
3619     --checksum-seed=32761 option to $Conf{RsyncArgs}), the cached
3620     checksums can be occaisonally verified to make sure the file
3621     contents matches the cached checksums. This is to avoid the
3622     risk that disk problems might cause the pool file contents to
3623     get corrupted, but the cached checksums would make BackupPC
3624     think that the file still matches the client.
3625    
3626     This setting is the probability (0 means never and 1 means always)
3627     that a file will be rechecked. Setting it to 0 means the checksums
3628     will not be rechecked (unless there is a phase 0 failure). Setting
3629     it to 1 (ie: 100%) means all files will be checked, but that is
3630     not a desirable setting since you are better off simply turning
3631     caching off (ie: remove the --checksum-seed option).
3632    
3633     The default of 0.01 means 1% (on average) of the files during a full
3634     backup will have their cached checksum re-checked.
3635    
3636     This setting has no effect unless checksum caching is turned on.
3637    
3638     =item $Conf{RsyncArgs} = [ ... ];
3639    
3640     Arguments to rsync for backup. Do not edit the first set unless you
3641     have a thorough understanding of how File::RsyncP works.
3642    
3643     Examples of additional arguments that should work are --exclude/--include,
3644     eg:
3645    
3646     $Conf{RsyncArgs} = [
3647     # original arguments here
3648     '-v',
3649     '--exclude', '/proc',
3650     '--exclude', '*.tmp',
3651     ];
3652    
3653     =item $Conf{RsyncRestoreArgs} = [ ... ];
3654    
3655     Arguments to rsync for restore. Do not edit the first set unless you
3656     have a thorough understanding of how File::RsyncP works.
3657    
3658     If you want to disable direct restores using rsync (eg: is the module
3659     is read-only), you should set $Conf{RsyncRestoreArgs} to undef and
3660     the corresponding CGI restore option will be removed.
3661    
3662     =item $Conf{ArchiveDest} = '/tmp';
3663    
3664     Archive Destination
3665    
3666     The Destination of the archive
3667     e.g. /tmp for file archive or /dev/nst0 for device archive
3668    
3669     =item $Conf{ArchiveComp} = 'gzip';
3670    
3671     Archive Compression type
3672    
3673     The valid values are:
3674    
3675     - 'none': No Compression
3676    
3677     - 'gzip': Medium Compression. Recommended.
3678    
3679     - 'bzip2': High Compression but takes longer.
3680    
3681     =item $Conf{ArchivePar} = 0;
3682    
3683     Archive Parity Files
3684    
3685     The amount of Parity data to generate, as a percentage
3686     of the archive size.
3687     Uses the commandline par2 (par2cmdline) available from
3688     http://parchive.sourceforge.net
3689    
3690     Only useful for file dumps.
3691    
3692     Set to 0 to disable this feature.
3693    
3694     =item $Conf{ArchiveSplit} = 0;
3695    
3696     Archive Size Split
3697    
3698     Only for file archives. Splits the output into
3699     the specified size * 1,000,000.
3700     e.g. to split into 650,000,000 bytes, specify 650 below.
3701    
3702     If the value is 0, or if $Conf{ArchiveDest} is an existing file or
3703     device (e.g. a streaming tape drive), this feature is disabled.
3704    
3705     =item $Conf{ArchiveClientCmd} = '$Installdir/bin/BackupPC_archiveHost' ...
3706    
3707     Archive Command
3708    
3709     This is the command that is called to actually run the archive process
3710     for each host. The following variables are substituted at run-time:
3711    
3712     $Installdir The installation directory of BackupPC
3713     $tarCreatePath The path to BackupPC_tarCreate
3714     $splitpath The path to the split program
3715     $parpath The path to the par2 program
3716     $host The host to archive
3717     $backupnumber The backup number of the host to archive
3718     $compression The path to the compression program
3719     $compext The extension assigned to the compression type
3720     $splitsize The number of bytes to split archives into
3721     $archiveloc The location to put the archive
3722     $parfile The amount of parity data to create (percentage)
3723    
3724     =item $Conf{SshPath} = '/usr/bin/ssh';
3725    
3726     Full path for ssh. Security caution: normal users should not
3727     allowed to write to this file or directory.
3728    
3729     =item $Conf{NmbLookupPath} = '/usr/bin/nmblookup';
3730    
3731     Full path for nmblookup. Security caution: normal users should not
3732     allowed to write to this file or directory.
3733    
3734     nmblookup is from the Samba distribution. nmblookup is used to get the
3735     netbios name, necessary for DHCP hosts.
3736    
3737     =item $Conf{NmbLookupCmd} = '$nmbLookupPath -A $host';
3738    
3739     NmbLookup command. Given an IP address, does an nmblookup on that
3740     IP address. The following variables are substituted at run-time:
3741    
3742     $nmbLookupPath path to nmblookup ($Conf{NmbLookupPath})
3743     $host IP address
3744    
3745     This command is only used for DHCP hosts: given an IP address, this
3746     command should try to find its NetBios name.
3747    
3748     =item $Conf{NmbLookupFindHostCmd} = '$nmbLookupPath $host';
3749    
3750     NmbLookup command. Given a netbios name, finds that host by doing
3751     a NetBios lookup. Several variables are substituted at run-time:
3752    
3753     $nmbLookupPath path to nmblookup ($Conf{NmbLookupPath})
3754     $host NetBios name
3755    
3756     In some cases you might need to change the broadcast address, for
3757     example if nmblookup uses 192.168.255.255 by default and you find
3758     that doesn't work, try 192.168.1.255 (or your equivalent class C
3759     address) using the -B option:
3760    
3761     $Conf{NmbLookupFindHostCmd} = '$nmbLookupPath -B 192.168.1.255 $host';
3762    
3763     If you use a WINS server and your machines don't respond to
3764     multicast NetBios requests you can use this (replace 1.2.3.4
3765     with the IP address of your WINS server):
3766    
3767     $Conf{NmbLookupFindHostCmd} = '$nmbLookupPath -R -U 1.2.3.4 $host';
3768    
3769     This is preferred over multicast since it minimizes network traffic.
3770    
3771     Experiment manually for your site to see what form of nmblookup command
3772     works.
3773    
3774     =item $Conf{FixedIPNetBiosNameCheck} = 0;
3775    
3776     For fixed IP address hosts, BackupPC_dump can also verify the netbios
3777     name to ensure it matches the host name. An error is generated if
3778     they do not match. Typically this flag is off. But if you are going
3779     to transition a bunch of machines from fixed host addresses to DHCP,
3780     setting this flag is a great way to verify that the machines have
3781     their netbios name set correctly before turning on DCHP.
3782    
3783     =item $Conf{PingPath} = '/bin/ping';
3784    
3785     Full path to the ping command. Security caution: normal users
3786     should not be allowed to write to this file or directory.
3787    
3788     If you want to disable ping checking, set this to some program
3789     that exits with 0 status, eg:
3790    
3791     $Conf{PingPath} = '/bin/echo';
3792    
3793     =item $Conf{PingCmd} = '$pingPath -c 1 $host';
3794    
3795     Ping command. The following variables are substituted at run-time:
3796    
3797     $pingPath path to ping ($Conf{PingPath})
3798     $host host name
3799    
3800     Wade Brown reports that on solaris 2.6 and 2.7 ping -s returns the wrong
3801     exit status (0 even on failure). Replace with "ping $host 1", which
3802     gets the correct exit status but we don't get the round-trip time.
3803    
3804     =item $Conf{ServerInitdPath} = '';
3805    
3806     =item $Conf{ServerInitdStartCmd} = '';
3807    
3808     Path to init.d script and command to use that script to start the
3809     server from the CGI interface. The following variables are substituted
3810     at run-time:
3811    
3812     $sshPath path to ssh ($Conf{SshPath})
3813     $serverHost same as $Conf{ServerHost}
3814     $serverInitdPath path to init.d script ($Conf{ServerInitdPath})
3815    
3816     Example:
3817    
3818    
3819     $Conf{ServerInitdPath} = '/etc/init.d/backuppc';
3820     $Conf{ServerInitdStartCmd} = '$sshPath -q -x -l root $serverHost'
3821     . ' $serverInitdPath start'
3822     . ' < /dev/null >& /dev/null';
3823    
3824     =item $Conf{CompressLevel} = 0;
3825    
3826     Compression level to use on files. 0 means no compression. Compression
3827     levels can be from 1 (least cpu time, slightly worse compression) to
3828     9 (most cpu time, slightly better compression). The recommended value
3829     is 3. Changing to 5, for example, will take maybe 20% more cpu time
3830     and will get another 2-3% additional compression. See the zlib
3831     documentation for more information about compression levels.
3832    
3833     Changing compression on or off after backups have already been done
3834     will require both compressed and uncompressed pool files to be stored.
3835     This will increase the pool storage requirements, at least until all
3836     the old backups expire and are deleted.
3837    
3838     It is ok to change the compression value (from one non-zero value to
3839     another non-zero value) after dumps are already done. Since BackupPC
3840     matches pool files by comparing the uncompressed versions, it will still
3841     correctly match new incoming files against existing pool files. The
3842     new compression level will take effect only for new files that are
3843     newly compressed and added to the pool.
3844    
3845     If compression was off and you are enabling compression for the first
3846     time you can use the BackupPC_compressPool utility to compress the
3847     pool. This avoids having the pool grow to accommodate both compressed
3848     and uncompressed backups. See the documentation for more information.
3849