samba - Aug 2012 - What could cause an "unclean shutdown" of process?

Hello everyone,

I've been experiencing many problems recently with windows software parsing
large text files stored on a samba share (smb version 3.6.3 on
ubuntu12.04). The software, which are running on windows XP stations, are
crashing quite often.

I notice this in the log.smbd at the time of the crash:
[2012/08/28 17:25:21.980609,  3] smbd/server.c:629(smbd_parent_housekeeping)
  Printcap cache time expired.
[2012/08/28 17:25:21.980679,  3] printing/pcap.c:138(pcap_cache_reload)
  reloading printcap cache
[2012/08/28 17:25:21.980756,  3] printing/pcap.c:189(pcap_cache_reload)
  reload status: ok
*[2012/08/28 17:37:08.715154,  3] smbd/server.c:292(remove_child_pid)*
*  smbd/server.c:292 Unclean shutdown of pid 32526*
*[2012/08/28 17:37:08.715318,  1] smbd/server.c:300(remove_child_pid)*
*  Scheduled cleanup of brl and lock database after unclean shutdown*
*[2012/08/28 17:37:28.722037,  1] smbd/server.c:272(cleanup_timeout_fn)*
*  Cleaning up brl and lock database after unclean shutdown*
[2012/08/28 17:38:22.509303,  3] smbd/server.c:629(smbd_parent_housekeeping)
  Printcap cache time expired.
[2012/08/28 17:38:22.509403,  3] printing/pcap.c:138(pcap_cache_reload)
  reloading printcap cache
[2012/08/28 17:38:22.509499,  3] printing/pcap.c:189(pcap_cache_reload)
  reload status: ok
I don't see anything that catches my eye in the other logs (log.nmbd,
log.<client_name>, log.<client_IP>, log.winbindd). The
log.<client_name>
shows a very long list of transactions that stops suddently at the time of
the 'unclean shutdown'.

Do you have any idea what could cause this 'unclean shutdown'? What
would
be the best way to further diagnose this problem? I've looked into the
servers' log without noticing anything strange?

Thank you very much for your help,

Frederic

Hi Frederic,

On Wed, 29 Aug 2012 11:04:22 -0400
Frederic Fournier <frederic.bioinfo at gmail.com> wrote:
> I notice this in the log.smbd at the time of the crash:
> [2012/08/28 17:25:21.980609,  3]
smbd/server.c:629(smbd_parent_housekeeping)
>   Printcap cache time expired.
> [2012/08/28 17:25:21.980679,  3] printing/pcap.c:138(pcap_cache_reload)
>   reloading printcap cache
> [2012/08/28 17:25:21.980756,  3] printing/pcap.c:189(pcap_cache_reload)
>   reload status: ok
> *[2012/08/28 17:37:08.715154,  3] smbd/server.c:292(remove_child_pid)*
> *  smbd/server.c:292 Unclean shutdown of pid 32526*
> *[2012/08/28 17:37:08.715318,  1] smbd/server.c:300(remove_child_pid)*
> *  Scheduled cleanup of brl and lock database after unclean shutdown*
> *[2012/08/28 17:37:28.722037,  1] smbd/server.c:272(cleanup_timeout_fn)*
> *  Cleaning up brl and lock database after unclean shutdown*
> [2012/08/28 17:38:22.509303,  3]
smbd/server.c:629(smbd_parent_housekeeping)
>   Printcap cache time expired.
> [2012/08/28 17:38:22.509403,  3] printing/pcap.c:138(pcap_cache_reload)
>   reloading printcap cache
> [2012/08/28 17:38:22.509499,  3] printing/pcap.c:189(pcap_cache_reload)
>   reload status: ok
> I don't see anything that catches my eye in the other logs (log.nmbd,
> log.<client_name>, log.<client_IP>, log.winbindd). The
log.<client_name>
> shows a very long list of transactions that stops suddently at the time of
> the 'unclean shutdown'.
These messages are the side effect of the asynchronous printcap update
process exiting normally, and should not be causing any problems.
To confirm this, you could change the "printcap cache time" setting to
0.

Cheers, David