Evidence of Resource Leak?

[Rod Widdowson]

It’s not the service reloading, it’s the entire IdP..

 

If you look for this string “[net.shibboleth.idp.log.LogbackLoggingService:206] - Shibboleth IdP Version 3.0.0”

You’ll see it about every 10 seconds from 2015-02-06 09:29:16,322, ^& that’s what we spit out during initial load.

 

Precisely why it’s reloading is unclear, maybe failfast is involved.  And precisely what is chewing up the memory is also not certain.

 

R

 

From: dev-bounces at shibboleth.net [mailto:dev-bounces at shibboleth.net] On Behalf Of Marvin Addison
Sent: 10 February 2015 20:10
To: Shib Dev
Subject: Re: Evidence of Resource Leak?

 

could you check to see whether the AttributeResolver Service was reloading configuration all weekend?

 

I just checked and noticed something I hadn't before: the resolver started loading every ~10s at some point despite that I have the following property defined:

 

idp.service.attribute.resolver.checkInterval=PT15M

 

You can see from the log file that I cited that it's reloading every 15m as expected up until about ~0930, which is probably when I started tweaking the springResources-based config. No idea why that would trigger more frequent reloading.

 

M

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://shibboleth.net/pipermail/dev/attachments/20150211/0dee8071/attachment.html