[BackupPC-users] archiving request backuppc process always fails

[Markus F. <mar...@xi...>]

hi!

all servers to backup are linux servers.
server1 has 100 servers to backup: 487.31GB containing 17536889 files 
and 4829 directories
server2 has 68 servers to backup: 740.42GB containing 8111272 files and 
4369 directories

backupPC processes run as user "wwwrun" - this is the apache user - 
because of the permissions making the configuration over the webinterface.
the archive request get startet over a cronjob and a small skript once a 
week:
  "sudo -u wwwrun /usr/local/BackupPC/bin/BackupPC_archiveStart 
archive-tape xadmin $XALL_HOSTS"
where the variable XALL_HOSTS contain all hosts of the backupPC server.

some of the archives get completed and later it fails and don't continue 
with the next archive.

logfile - error at the end:
............
Executing: /usr/local/BackupPC/bin/BackupPC_archiveHost 
/usr/local/BackupPC/bin/BackupPC_tarCreate /usr/bin/split  flirty2 24 
/usr/bin/gzip .gz 10737418240 /SATA/BackupPC/archive 0 *
Xfer PIDs are now 14276
Writing tar archive for host flirty2, backup #24, split to output files 
/SATA/BackupPC/archive/flirty2.24.tar.gz.*
Done: 1201 files, 97498657 bytes, 256 dirs, 445 specials, 0 errors
Executing: /usr/local/BackupPC/bin/BackupPC_archiveHost 
/usr/local/BackupPC/bin/BackupPC_tarCreate /usr/bin/split  flirty5 24 
/usr/bin/gzip .gz 10737418240 /SATA/BackupPC/archive 0 *
Xfer PIDs are now 14298
Writing tar archive for host flirty5, backup #24, split to output files 
/SATA/BackupPC/archive/flirty5.24.tar.gz.*
Done: 1196 files, 97942858 bytes, 255 dirs, 449 specials, 0 errors
Executing: /usr/local/BackupPC/bin/BackupPC_archiveHost 
/usr/local/BackupPC/bin/BackupPC_tarCreate /usr/bin/split  hosting1 24 
/usr/bin/gzip .gz 10737418240 /SATA/BackupPC/archive 0 *
Xfer PIDs are now 14304
Writing tar archive for host hosting1, backup #24, split to output files 
/SATA/BackupPC/archive/hosting1.24.tar.gz.*
exiting after signal ALRM
Archive failed: aborted by signal=ALRM

the strange thing is, that the duration shows ~1200 min  on both 
backupPC servers for the archive job, when it fails.
is there a limitation anywhere in the code?

I tested both: archiving with and without split and both fail

changing the splitsize value for a host is written into the config but 
it does not get read, when the scripts 
"/usr/local/BackupPC/bin/BackupPC_archiveStart" runs - there it is 
hardcoded!
why? I changed this by my own.

here is the DIFF:
--- /usr/local/BackupPC/bin/BackupPC_archiveStart.orig   2011-09-22 
17:53:40.000000000 +0200
+++ /usr/local/BackupPC/bin/BackupPC_archiveStart       2011-09-09 
14:39:04.000000000 +0200
@@ -100,7 +100,7 @@
      compext     => $bpc->{Conf}{ArchiveComp} eq 'none' ? ''
                      : ($bpc->{Conf}{ArchiveComp} eq 'gzip' ? '.gz' : 
'.bz2'),
      parfile     => $bpc->{Conf}{ArchivePar},
-    splitsize   => '0000000',
+    splitsize   => $bpc->{Conf}{ArchiveSplit},
      host        => $ArchiveHost,
      HostList    => \@HostList,
      BackupList  => \@BackupList,

----------

hardwarinfo:
  * server1
   - Quadcore X86-64 Intel(R) Xeon(R) CPU  2.00GHz
   - 3.8 GB memory
   - SLES11 SP1 x86_64

  * server2
   - Singlecore X86-64 Intel(R) Xeon(TM) CPU 2.80GHz
   - 5 GB memory
   - SLES10 SP2 x86_64

configuration file of archive host:
/etc/BackupPC/pc/archive-tape.pl
$Conf{ClientNameAlias} = '127.0.0.1';
$Conf{XferMethod} = 'archive';
$Conf{ArchiveSplit} = '10737418240';
$Conf{XferLogLevel} = 2;


kind regard
markus fröhlich