[rsyslog] Out of memory

Rainer Gerhards rgerhards at hq.adiscon.com
Thu Jan 19 10:54:46 CET 2017


2017-01-19 10:19 GMT+01:00 mostolog--- via rsyslog <
rsyslog at lists.adiscon.com>:

> El 18/01/17 a las 17:05, Rainer Gerhards escribió:
>
>> If this is a test environment, I would strongly recommend to run it in a
>> terminal session under valgrind control. If it is an actual leak, valgrind
>> can provide very good diagnostic information.
>>
>> That would be very helpful and clarify what is going on.
>>
>> Rainer
>>
> On my way!
>
> Any valgrind special setting?
>

No, the defaults are good enough.


>
> Will testing 1 message each 5 seconds be enough?
>

Depends on how long you run it. To clearly diagnose a memleak, I think we
should inject something between 100,000 and 1 million messages. So it's
probably better to inject more.


>
> What impstats interval should I use?
>

If there is a memleak, it should show up in valgrind log, once rsyslog is
terminated. the maxrss is just extra information, especially if valgrind
doesn't show anything (there could be reasons if so, but the case sounds
like we should get a valgrind report). The most important thing is that it
would be nice to have an impstats report reflecting the state after
everything has been injected and before rsyslog is terminated.

>
> Does 1056K says anything to you?
>
not at all, except that it is 1M plus 32k ;-)

Rainer


>
>
> El 18/01/17 a las 17:06, David Lang escribió:
>
>> grep -e imrelp -e resource-usage |cut -f 4,6- -d " "
>>
> Will do.
>
> I'm less interested in the with and without ES than I am in 100m/sec vs
>> 2000m/sec for the same total number of messages. 1m messages/test would be
>> great, but that takes a long time, let's start with something small like
>> 50K messages per test.
>>
>> it doesn't need to be exact, but I want to see if ~50K messages @
>> 100m/sec and ~50K messages @ 2000m/sec result in similar maxrss sizes in
>> the pstats output.
>>
> Ok.
>
> with very few exceptions, rsyslog releases the memory as it goes, there
>> should not be any significant amount of memory freed by rsyslog after it's
>> been idle for a while.
>>
> But being idle for 15m should release memory from 512MB to a few KB if
> they aren't used, isnt it?
>
>
> _______________________________________________
> rsyslog mailing list
> http://lists.adiscon.net/mailman/listinfo/rsyslog
> http://www.rsyslog.com/professional-services/
> What's up with rsyslog? Follow https://twitter.com/rgerhards
> NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad
> of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you
> DON'T LIKE THAT.
>


More information about the rsyslog mailing list