Re: [AMaViS-user] No available locker entries

[Paul J. <pj...@cu...>]

it may well be completely unrelated but I have spent a considerable 
amount of time trying to find the cause what sounds like a very 
similar set of symptoms.

I should have been more precise in my description - the SA helper 
apps fail to complete, while the child process seems to complete. The 
helper apps failing causes locks to not be released resulting in the 
kind of  resource exhaustion you describe. There would also be 
legitimate mail showing up in the postfix mailq. Deferred messages 
related to the problem I've described will correspond to a postfix 
log message saying something similar to "lost connection with 
127.0.0.1 port 10024; message may be delivered more than once" but 
the evidence helper app failure will only show up in the output from 
amavisd running in "debug-sa" mode.

if neither the log messages or helper app failures in debug-sa mode 
are apparent you can ignore everything I've written in this thread...

cheers
Paul

At 10:05 PM -0500 30/1/05, Michael W Cocke wrote:
>Sounds completely different from my problem.  I've not had a problem
>with SA since I upgraded to it from 2.63.  My problem is a resource
>exhaust in Amavisd (or in BerkeleyDB, the exact problem could be in
>either - I can't see a cause as yet).  SA etc. are all completing
>nicely.  I have SA maxlife set to a fairly short life (each child
>processes a dozen messages) before the child is killed and a fresh
>child is born for that very reason.  The actual issue that causes THAT
>problem looks to be in IOstringy, but that's about as far as I took
>that - when they added the maxlife in SA I had all I needed.
>
>Mike-