That's kernel space, not application space.  Your most likely culprits
are bad memory or excess heat.  Try to log/graph lmsensors for the 
latter, and run memtest86 (for at least one full run) for the former.
Also look at any other hardware monitors you might have available, 
including smartd.

Even if you were running out of memory with the original 4G, you shouldn't
see that kind of behavior.  You could expect to see paging, then thrashing,
then the kernel killing off processes, but you shouldn't see that type
of error.  So suspect a hardware fault.

In fact, your swap space is barely touched so (at least with 16G) you're
probably not seeing any memory pressure at all.  (What is used could 
just be related to tmpfs or the filesystem cache.)

Devin
-- 
Shirt, Shoes, Sober... --Pick Two
                                                        - Chuck Yerkes


------------------------------------------------------------------------------
Learn how Oracle Real Application Clusters (RAC) One Node allows customers
to consolidate database storage, standardize their database environment, and, 
should the need arise, upgrade to a full multi-node Oracle RAC database 
without downtime or disruption
http://p.sf.net/sfu/oracle-sfdevnl
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users