I've had a similar problem this week as well. We use Shib 1.3 with Tomcat/Apache on Linux.
I've narrowed it down a bit that Tomcat is reporting out of memory when it fails (although it doesn't actually appear to be) and at the moment it seems to be failing within an hour of the metadata download completing. My plan for today is to shift the time of the download and see if it changes the pattern of the failure....
I've allocated Tomcat and the server (a VM) more RAM but it hasn't fixed it.
Cheers
dave
On 16/10/2009 08:56, "Heather Peake" <[log in to unmask]> wrote:
Hi
we've had a Shibboleth idp up and running for more than a year now and it
seemed all right.
Suddenly this week it is failing everyday with no indication in the log
files as to why. Restarting apache doesn't fix it, but restarting the server
is fixing it.
We are on version 1.3 using Apache and Tomcat on a Windows 2003 server and
hooking into a SQL database on a different server for the attributes. The
only difference I'm aware ofthis week is significant network slow downs, but
would that stop shibboleth working?
Anyone got any ideas on what might be going on or how I can turn additional
logging on to see if that shines a light?
Any suggestions greatfully accepted.
Regards
Heather
--
Dave Thornley
Service Support Manager
Learning and IT Services
Sheffield Hallam University
Tel: 0114 225 3822 / 07771 974349
Email: [log in to unmask]
|