--On Thursday, February 01, 2007 8:20 PM -0800 Kurt Albershardt <[EMAIL 
PROTECTED]> wrote:

--On Wednesday, January 31, 2007 1:42 PM -0800 Kurt Albershardt <[EMAIL 
PROTECTED]> wrote:

--On Wednesday, January 31, 2007 10:33 PM +0100 Patrick de Ruiter <[EMAIL 
PROTECTED]> wrote:

On Wed, 2007-01-31 at 11:59 -0800, Kurt Albershardt wrote:

systems limit threads per process.  Perhaps you've exceeded that
limit?  HP-UX, for example, IIRC sets that limit fairly low initially.


Hmm very strange dude,

Maybe your out of filedescriptors,

Cheers
Patrick


My system is quite small (total DB space = 1.2G) but is on a fairly low powered box 
(1 gHz low power Intel Celeron, 512mb RAM, 2.5" drive.)

vmstat shows

procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in    cs us sy id wa
 2  0      0  18616  12000 319992    0    0    75   304  317    44  1  1 74 24

The wait queue goes as high as 4-5 at times so we're looking CPU bound to me 
(but please correct me if I'm misinterpreting this.)


OK, more info.  Wait states are after the dspam restart and postfix queue 
flush, they disappear as soon as the queue clears.  Postfix and mysql seem to 
be fine here so I am trying to understand which resource dspam is bumping up 
against.  Nothing in my syslog or any of the postfix logs other than these in 
mail.err:

Feb  1 08:40:47 localhost dspam[10669]: Thread creation failed: Cannot allocate 
memory
Feb  1 08:45:19 localhost dspam[10669]: Thread creation failed: Cannot allocate 
memory
Feb  1 08:45:33 localhost last message repeated 5 times
Feb  1 16:52:13 localhost dspam[20724]: Thread creation failed: Cannot allocate 
memory
Feb  1 16:52:13 localhost last message repeated 3 times
Feb  1 16:57:25 localhost dspam[20724]: Thread creation failed: Cannot allocate 
memory
Feb  1 16:58:29 localhost last message repeated 4 times

Gap is the time between restart and failure.

Is there something I can do to log the events leading up to this failure or 
something I should look for?


This was hitting me every 7-9 hours.  After setting up a crontab for 
'/etc/init.d/dspam restart' every 4 hours the system has been glitch free for 
14 days now.

Anyone have an idea what resource I am exhausting here?








--thanks

Reply via email to