[GRLUG] Memory leak

[Don Ellis]

I think one thing we'd like (Lee & I) is some simple recipes for
people with less than massive experience at process monitoring. We
hear the terms mentioned, with "use this tool or that tool" but no
guidelines for some common conditions to look for. Are there specific
symptoms for specific kinds of pathology that would give hints that
something is not as it should be?

One caveat I see in text on this subject is: look for unexpected
results, as some process will be expected to show these conditions,
and can be ignored. Other process should not do that, and these
conditions would indicate these processes are outside normal expected
range. I guess part of this is measuring past history of every process
and watching for certain changes (which ones and how much?). There are
probably monitoring tools for this - I hope they are relatively easy
to set up, with standard package management files. Hassling with
varied system configurations & tools can really increase time
investment when done across multiple platforms.

--Don Ellis


On Wed, Sep 26, 2012 at 1:51 PM, L. V. Lammert <lvl at omnitec.net> wrote:
> Been having problems with a server blocking (i.e. %wa > 90%), .. I rebooted
> it this AM and it's back to normal!
>
> Memory Usage is now < 1GB (2GB physical, 4GB swap), where last night it was
>> 3GB total! As all of the processes are fairly static (it's a Nagios box),
> with some reverse ssh tunnels and a miniscule MySQL database.
>
> Does anyone know of a way to get more information about where the leak is
> occurring, if that is indeed the problem?
>
>         Thanks!
>
>         Lee