On 28/02/17 19:42, Daryl Richards via Exim-users wrote:
> On 28/02/2017 2:10 PM, Jeremy Harris wrote:
>> On 28/02/17 18:50, Daryl Richards via Exim-users wrote:
>>> This is happening with MANY hosts, not just that one.. So far today:
>>>
>>> root@lagoon:~# grep SSL_write /var/log/exim/mainlog | wc -l
>>> 14970469
>>> root@lagoon:~# grep SSL_write /var/log/exim/mainlog | uniq -s 20 | wc -l
>>> 414281
>>>
>>> This is all local delivery. Chunking has been disabled for both incoming
>>> and outgoing (just to be sure)...
>> Checking for SSL_write isn't necessarily conclusive; could you grep for
>> "exited on signal" ?
>>
> True enough, this is what I get from there (in my syslog actually)
>
> Feb 28 00:00:46 lagoon kernel: pid 79778 (exim), uid 26: exited on
> signal 11
[...]
> Feb 28 14:32:22 lagoon kernel: pid 6207 (exim), uid 26: exited on signal 11
> Feb 28 14:36:33 lagoon kernel: pid 6341 (exim), uid 26: exited on signal 11
>
> But then, for example, doing a grep for the last three pids that failed,
> gives this many lines per:
>
> root@lagoon:~# grep 6341 /var/log/exim/mainlog | wc -l
> 63867
> root@lagoon:~# grep 6207 /var/log/exim/mainlog | wc -l
> 64054
> root@lagoon:~# grep 6167 /var/log/exim/mainlog | wc -l
> 63879
mmpff. we could try to make it give up after the first, but we'd still
need to fix the crashing process - so probably not worth putting any
effort into.
Can you set up for getting a coredump?
--
Cheers,
Jeremy
