Monitoring or managing IDP memory usage

[Wessel, Keith]

Thanks, Scott.

I'll try adding the math to subtract free memory. On my system, it seems to have helped on my hot spare servers. I'll see what it does on the active servers when I'm able to do a restart of Jetty.

I, too, was considering the alert when the out of memory exception occurred; was going to do that in Splunk, though. But I'm hoping to catch it before it happens.

Keith


-----Original Message-----
From: users [mailto:users-bounces at shibboleth.net] On Behalf Of Cantor, Scott
Sent: Friday, September 23, 2016 12:16 PM
To: Shib Users <users at shibboleth.net>
Subject: Re: Monitoring or managing IDP memory usage

On 9/23/16, 1:04 PM, "users on behalf of Wessel, Keith" <users-bounces at shibboleth.net on behalf of kwessel at illinois.edu> wrote:

> And according to the plugin, as soon as our Nagios admin turned it on, we were out of heap
> space. The status page does, in fact, report that we're using approx.. 2964 MB of a max
> 2964 MB.

You may just be hitting this:

https://issues.shibboleth.net/jira/browse/IDP-947

I don't know much about any of the heap reporting beyond the numbers.

With respect to Nagios, what I did was to implement a mailer in logback that sent me a page if it logged an out of memory exception. Admittedly that's reactive, but I didn't have more memory to give it, so it either was going to stay up or not (and it did not, even Jim's G1 GC suggestion doesn't keep a 32-bit IdP up with a max heap now).

-- Scott


-- 
To unsubscribe from this list send an email to users-unsubscribe at shibboleth.net