--- trunk/bin/BackupPC_incPartsUpdate 2005/12/12 16:07:27 254 +++ trunk/bin/BackupPC_incPartsUpdate 2005/12/13 22:38:09 274 @@ -15,6 +15,7 @@ use Archive::Tar::Streamed; use Algorithm::Diff; use Getopt::Std; +use File::Slurp; my $bpc = BackupPC::Lib->new || die "can't create BackupPC::Lib"; my %Conf = $bpc->Conf(); @@ -22,9 +23,6 @@ use BackupPC::SearchLib; %BackupPC::SearchLib::Conf = %Conf; -# cludge: minimum .tar.gz size -my $MIN_TAR_SIZE = 80; - my $path = abs_path($0); $path =~ s#/[^/]+$#/#; my $tarIncCreate = $path .= 'BackupPC_tarIncCreate'; @@ -72,19 +70,81 @@ return strftime($t_fmt,localtime()); } +my $hsn_cache; + +sub get_backup_id($$$) { + my ($host, $share, $num) = @_; + + my $key = "$host $share $num"; + return $hsn_cache->{$key} if ($hsn_cache->{$key}); + + my $sth = $dbh->prepare(qq{ + SELECT + backups.id + FROM backups + INNER JOIN shares ON backups.shareID=shares.ID + INNER JOIN hosts ON backups.hostID = hosts.ID + where hosts.name = ? and shares.name = ? and backups.num = ? + }); + $sth->execute($host, $share, $num); + my ($id) = $sth->fetchrow_array; + + $hsn_cache->{"$host $share $num"} = $id; + + print STDERR "# $host $share $num == $id\n" if ($opt{d}); + + return $id; +} + + sub tar_check($$$$) { my ($host,$share,$num,$filename) = @_; - if ($debug) { - print STDERR " {{ CHECK: ${host}:${share}#${num} and $filename"; - } else { - print " check"; + my $t = time(); + print curr_time, " check $host:$share#$num -> $filename"; + + sub check_part { + my ($host, $share, $num, $part_nr, $tar_size, $size, $md5, $items) = @_; + my $backup_id = get_backup_id($host, $share, $num); + my $sth_md5 = $dbh->prepare(qq{ + select + id, tar_size, size, md5, items + from backup_parts + where backup_id = ? and part_nr = ? + }); + + $sth_md5->execute($backup_id, $part_nr); + + if (my $row = $sth_md5->fetchrow_hashref) { + return if ( + $row->{tar_size} >= $tar_size && + $row->{size} == $size && + $row->{md5} eq $md5 && + $row->{items} == $items + ); + print ", deleting invalid backup_parts $row->{id}"; + $dbh->do(qq{ delete from backup_parts where id = $row->{id} }); + } + print ", inserting new"; + my $sth_insert = $dbh->prepare(qq{ + insert into backup_parts ( + backup_id, + part_nr, + tar_size, + size, + md5, + items + ) values (?,?,?,?,?,?) + }); + + $sth_insert->execute($backup_id, $part_nr, $tar_size, $size, $md5, $items); + $dbh->commit; } my @tar_parts; if (-d "$tar_dir/$filename") { - print STDERR " multi-part" if ($opt{d}); + print ", multi-part"; opendir(my $dir, "$tar_dir/$filename") || die "can't readdir $tar_dir/$filename: $!"; @tar_parts = map { my $p = $_; $p =~ s#^#${filename}/#; $p } grep { !/^\./ && !/md5/ && -f "$tar_dir/$filename/$_" } readdir($dir); closedir($dir); @@ -92,37 +152,83 @@ push @tar_parts, "${filename}.tar.gz"; } + print " [parts: ",join(", ", @tar_parts),"]" if ($opt{d}); + + my $same = 1; my @tar_files; - print " [parts: ",join(", ", @tar_parts),"]" if ($opt{d}); + my $backup_part; - print " reading"; + print " reading" if ($opt{d}); foreach my $tarfilename (@tar_parts) { - print STDERR " $tarfilename" if ($debug); + print "\n\t- $tarfilename"; - my $path = "$tar_dir/$tarfilename"; - my $md5 = $path; - $md5 =~ s/\.tar\.gz$/.md5/ || die "can't create md5 filename from $md5"; - if (! -e $md5) { - print ", creating md5"; - system( $bin->{md5sum} . " $path > $md5") == 0 or die "can't create md5 $path: $!"; + my $size = (stat( "$tar_dir/$tarfilename" ))[7] || die "can't stat $tar_dir/$tarfilename"; + + if ($size > $Conf{MaxArchiveSize}) { + print ", part bigger than media $size > $Conf{MaxArchiveSize}\n"; + return 0; } + print ", $size bytes"; + + my $path = "$tar_dir/$tarfilename"; + open(my $fh, "gzip -cd $tar_dir/$tarfilename |") or die "can't open $tar_dir/$tarfilename: $!"; binmode($fh); my $tar = Archive::Tar::Streamed->new($fh); + my $tar_size = 0; + my $items = 0; + while(my $entry = $tar->next) { push @tar_files, $entry->name; + $items++; + $tar_size += $entry->size; + + if ($tar_size > $Conf{MaxArchiveFileSize}) { + print ", part $tarfilename is too big $tar_size > $Conf{MaxArchiveFileSize}\n"; + return 0; + } + + } + + print ", $items items"; + + # + # check if md5 exists, and if not, create one + # + + my $md5_path = $path; + $md5_path =~ s/\.tar\.gz$/.md5/ || die "can't create md5 filename from $md5_path"; + if (! -e $md5_path || -z $md5_path) { + print ", creating md5"; + system( $bin->{md5sum} . " $path > $md5_path") == 0 or die "can't create md5 $path: $!"; + } else { + ## FIXME check if existing md5 is valid } + + my $md5 = read_file( $md5_path ) || die "can't read md5sum file $md5_path: $!"; + $md5 =~ s#\s.*$##; + + # extract part number from filename + my $part_nr = 1; + $part_nr = $1 if ($tarfilename =~ m#/(\d+)\.tar\.gz#); + + # + # finally, check if backup_parts table in database is valid + # + + check_part($host, $share, $num, $part_nr, $tar_size, $size, $md5, $items); } - @tar_files = sort @tar_files; - print STDERR " ",($#tar_files + 1), " files" if ($debug); + # short-cut and exit; + return $same unless($same); - print STDERR ", database" if ($debug); + @tar_files = sort @tar_files; + print "\n\t",($#tar_files + 1), " tar files"; my $sth = $dbh->prepare(qq{ SELECT path,type @@ -141,14 +247,13 @@ push @db_files, $path; } - print STDERR " ",($#db_files + 1), " files, diff" if ($debug); + print " ",($#db_files + 1), " database files, diff"; @db_files = sort @db_files; - my $same = 1; if ($#tar_files != $#db_files) { $same = 0; - print STDERR " NUMBER" if ($debug); + print " NUMBER"; } else { my $diff = Algorithm::Diff->new(\@tar_files, \@db_files); while ( $diff->Next() ) { @@ -159,8 +264,8 @@ } } - print " ",($same ? 'ok' : 'DIFFERENT'); - print STDERR " }} " if ($debug); + print " ",($same ? 'ok' : 'DIFFERENT'), + ", dur: ",fmt_time(time() - $t), "\n"; return $same; } @@ -191,6 +296,9 @@ my $curr_backup = 1; while (my $row = $sth->fetchrow_hashref) { + + $curr_backup++; + my $tar_file = BackupPC::SearchLib::getGzipName($row->{'host'}, $row->{'share'}, $row->{'num'}); # this will return -1 if file doesn't exist @@ -198,12 +306,15 @@ print "# size: $size backup.size: ", $row->{inc_size},"\n" if ($opt{d}); - if ( $row->{'inc_size'} != -1 && $size != -1 && $row->{'inc_size'} == $size && ( $check && tar_check($row->{'host'}, $row->{'share'}, $row->{'num'}, $tar_file) || 1) ) { - next; + if ( $row->{'inc_size'} != -1 && $size != -1 && $row->{'inc_size'} >= $size) { + if ($check) { + tar_check($row->{'host'}, $row->{'share'}, $row->{'num'}, $tar_file) && next; + } else { + next; + } } - print curr_time, " $curr_backup/$num_backups ", $row->{'host'}, ":", $row->{'share'}, " #", $row->{'num'}, " -> $tar_file"; - $curr_backup++; + print curr_time, " creating $curr_backup/$num_backups ", $row->{'host'}, ":", $row->{'share'}, " #", $row->{'num'}, " -> $tar_file"; my $t = time();