+my $hsn_cache;
+
+sub get_backup_id($$$) {
+ my ($host, $share, $num) = @_;
+
+ my $key = "$host $share $num";
+ return $hsn_cache->{$key} if ($hsn_cache->{$key});
+
+ my $sth = $dbh->prepare(qq{
+ SELECT
+ backups.id
+ FROM backups
+ INNER JOIN shares ON backups.shareID=shares.ID
+ INNER JOIN hosts ON backups.hostID = hosts.ID
+ WHERE hosts.name = ? and shares.name = ? and backups.num = ?
+ });
+ $sth->execute($host, $share, $num);
+ my ($id) = $sth->fetchrow_array;
+
+ $hsn_cache->{"$host $share $num"} = $id;
+
+ print STDERR "# $host $share $num == $id\n" if ($opt{d});
+
+ return $id;
+}
+
+sub backup_inc_deleted($) {
+ my $backup_id = shift;
+ my $sth_inc_deleted = $dbh->prepare(qq{
+ update backups set
+ inc_deleted = true
+ where id = ?
+ });
+ $sth_inc_deleted->execute($backup_id);
+}
+
+sub tar_check($$$$) {
+ my ($host,$share,$num,$filename) = @_;
+
+ my $t = time();
+ print curr_time, " check $host:$share#$num -> $filename";
+
+ # depending on expected returned value this is used like:
+ # my $uncompress_size = get_gzip_size('/full/path/to.gz');
+ # my ($compress_size, $uncompress_size) = get_gzip_size('/path.gz');
+ sub get_gzip_size($) {
+ my $filename = shift;
+ die "file $filename problem: $!" unless (-r $filename);
+ open(my $gzip, $bin->{gzip}." -l $filename |") || die "can't gzip -l $filename: $!";
+ my $line = <$gzip>;
+ chomp($line);
+ $line = <$gzip> if ($line =~ /^\s+compressed/);
+
+ my ($comp, $uncomp) = (0,0);
+
+ if ($line =~ m/^\s+(\d+)\s+(\d+)\s+\d+\.\d+/) {
+ if (wantarray) {
+ return [ $1, $2 ];
+ } else {
+ return $2;
+ }
+ } else {
+ die "can't find size in line: $line";
+ }
+ }
+
+ sub check_part {
+ my ($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 ", 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);
+ } else {
+ push @tar_parts, "${filename}.tar.gz";
+ }
+
+ print " [parts: ",join(", ", @tar_parts),"]" if ($opt{d});
+
+ my $same = 1;
+ my @tar_files;
+
+ my $backup_part;
+
+ print " reading" if ($opt{d});
+
+ foreach my $tarfilename (@tar_parts) {
+
+ print "\n\t- $tarfilename";
+
+ my $path = "$tar_dir/$tarfilename";
+
+ my $size = (stat( $path ))[7] || die "can't stat $path: $!";
+
+ if ($size > $Conf{MaxArchiveSize}) {
+ print ", part bigger than media $size > $Conf{MaxArchiveSize}\n";
+ return 0;
+ }
+
+ print ", $size bytes";
+
+
+ open(my $fh, "gzip -cd $path |") or die "can't open $path: $!";
+ binmode($fh);
+ my $tar = Archive::Tar::Streamed->new($fh);
+
+ my $tar_size_inarc = 0;
+ my $items = 0;
+
+ while(my $entry = $tar->next) {
+ push @tar_files, $entry->name;
+ $items++;
+ $tar_size_inarc += $entry->size;
+
+ if ($tar_size_inarc > $Conf{MaxArchiveFileSize}) {
+ print ", part $tarfilename is too big $tar_size_inarc > $Conf{MaxArchiveFileSize}\n";
+ return 0;
+ }
+
+ }
+
+ close($fh);
+
+ print ", $items items";
+
+ if ($tar_size_inarc == 0 && $items == 0) {
+ print ", EMPTY tar\n";
+
+ my $backup_id = get_backup_id($host, $share, $num);
+ backup_inc_deleted( $backup_id );
+
+ $dbh->commit;
+
+ return 1;
+ }
+
+ my $tar_size = get_gzip_size( $path );
+
+ # real tar size is bigger because of padding
+ if ($tar_size_inarc > $tar_size) {
+ print ", size of files in tar ($tar_size_inarc) bigger than whole tar ($tar_size)!\n";
+ return 0;
+ }
+
+ #
+ # 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);
+ }
+
+ # short-cut and exit;
+ return $same unless($same);
+
+ @tar_files = sort @tar_files;
+ print "\n\t",($#tar_files + 1), " tar files";
+
+ my $sth = $dbh->prepare(qq{
+ SELECT path,type
+ FROM files
+ JOIN shares on shares.id = shareid
+ JOIN hosts on hosts.id = shares.hostid
+ WHERE hosts.name = ? and shares.name = ? and backupnum = ?
+ });
+ $sth->execute($host, $share, $num);
+ my @db_files;
+ while( my $row = $sth->fetchrow_hashref ) {
+
+ my $path = $row->{'path'} || die "no path?";
+ $path =~ s#^/#./#;
+ $path .= '/' if ($row->{'type'} == BPC_FTYPE_DIR);
+ push @db_files, $path;
+ }
+
+ print " ",($#db_files + 1), " database files, diff";
+
+ @db_files = sort @db_files;
+
+ if ($#tar_files != $#db_files) {
+ $same = 0;
+ print " NUMBER";
+ } else {
+ my $diff = Algorithm::Diff->new(\@tar_files, \@db_files);
+ while ( $diff->Next() ) {
+ next if $diff->Same();
+ $same = 0;
+ print "< $_\n" for $diff->Items(1);
+ print "> $_\n" for $diff->Items(2);
+ }
+ }
+
+ print " ",($same ? 'ok' : 'DIFFERENT'),
+ ", dur: ",fmt_time(time() - $t), "\n";
+
+ return $same;
+}
+
+