Hi,

I'm running Bacula 5.0.2 in conjunction with PostgreSQL 8.4.3 (both
compiled from the "official" sources) on a Debian Lenny system. The
system runs the director, storage daemon and file daemon. The actual
storage media are barcode labeled LTO3 tapes contained in a HP Storage
Works 1/8 G2 autoloader equipped with an HP Ultrium 920 drive.

I've already followed the instructions in
                                
http://tldp.org/HOWTO/html_single/TCP-Keepalive-HOWTO/

and the corresponding FAQ entry
                        
http://wiki.bacula.org/doku.php?id=faq#my_backup_starts_but_dies_after_a_while_with_connection_reset_by_peer_error
                                                               
since I came accross this thread, which had been started before I subscribed to 
the Bacula users mailing list:
                                        
http://adsm.org/lists/html/Bacula-users/2010-04/msg00172.html

I decided to compile the libkeeplive from Source Forge and use the
LD_PRELOAD mechanism in order to make sure libkeeplive.so is always
loaded. Furthermore I changed the sysctl and Heartbeat Interval
settings as described in the referenced FAQ entry.

Disabling accurate backups is not a solution for me since I want
deleted files to be taken into account and I have to run quite a few
long running shell scripts for gathering data via SCP from other hosts
before the actual backup (writing to tape via SD) starts. So, a
workaround for the TCP keepalive problem is absolutely necessary for me.

A status inquiry on the client from within bconsole works without a
problem, but even with TCP keepalive enabled my backup stops after a very short 
period of time, as the
log from the Bacula director shows:

===

1-Mai 16:33 nathan-sd JobId 13: Job write elapsed time = 00:00:36,
Transfer rate = 20.30 M Bytes/second
11-Mai 16:34 nathan-fd JobId 13: Fatal error: backup.c:1019 Network
send error to SD. ERR=Connection reset by peer
11-Mai 16:34 nathan-dir JobId 13: Error: Bacula nathan-dir 5.0.2
(28Apr10): 11-Mai-2010 16:34:27
  Build OS:               x86_64-unknown-linux-gnu debian 5.0.4
  JobId:                  13
  Job:                    nathan_backup.2010-05-11_16.32.43_03
  Backup Level:           Full (upgraded from Incremental)
  Client:                 "nathan" 5.0.2 (28Apr10) 
x86_64-unknown-linux-gnu,debian,5.0.4
  FileSet:                "nathan fileset" 2010-05-10 23:05:00
  Pool:                   "WeeklyBackups" (From Job FullPool override)
  Catalog:                "MyCatalog" (From Client resource)
  Storage:                "nathan-sd" (From Pool resource)
  Scheduled time:         11-Mai-2010 16:32:41
  Start time:             11-Mai-2010 16:32:45
  End time:               11-Mai-2010 16:34:27
  Elapsed time:           1 min 42 secs
  Priority:               10
  FD Files Written:       4,072
  SD Files Written:       4,059
  FD Bytes Written:       734,948,632 (734.9 MB)
  SD Bytes Written:       731,138,860 (731.1 MB)
  Rate:                   7205.4 KB/s
  Software Compression:   None
  VSS:                    no
  Encryption:             no
  Accurate:               yes
  Volume name(s):         
  Volume Session Id:      1
  Volume Session Time:    1273588341
  Last Volume Bytes:      2,194,827,264 (2.194 GB)
  Non-fatal FD errors:    0
  SD Errors:              0
  FD termination status:  Error
  SD termination status:  Error
  Termination:            *** Backup Error ***

===

Any help will be greatly appreciated.

Thanks in advance & kind regards,

    Holger
    

signature.asc
Description: Digital signature

------------------------------------------------------------------------------

_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users