Re: Opendkim number of threads growing abnormaly under Solaris 10

From: Alessandro Vesely <vesely_at_tana.it>
Date: Sat, 16 Oct 2010 12:41:55 +0200

Pardon me please for inadvertently stumbling into this thread...

On 14/Oct/10 03:58, John Wood wrote:
>> This confirms that it's stuck waiting for a DNS reply. There are many
>> possible causes:
>>
>> - the resolver against which the filter was linked is not thread-safe,
>> [...]
>
> I had asked for the output of ldd on the binary earlier specifically
> because I thought this was worth verifying. That being said, libresolv
> in Solaris 10 is thread safe:

I agree with Murray that the likely cause is DNS queries. However, I
don't think that a 75-15 difference in 12 hours is enough to infer a
systematic resolver breakage (misconfigured firewall or packet filter,
or some other reason). It obviously depends on the rate of incoming
mail, but couldn't it be that /some/ domains deserve a timeout that
never arrives? The excheck function is involved in checking their
A/AAAA/MX existence...

I'd be curious to know how does the MTA react in this case --taking
more than 10 minutes for body filtering. Should it be the
responsibility of each milter not to exceed eom timeout?
Received on Sat Oct 16 2010 - 10:42:06 PST