shibboleth daemon died with heavy apache load

[csross]

Our apache server (solaris) was overloaded with requests the other day.  Most
were NOT requests that require shibboleth.  The server was not out of memory
or CPU, but memory was low for a few hours.

We heard from a customer that SSO stopped working and noticed the shibd
daemon had stopped.  It was manually started again and SSO was available
again without issue.  I looked in the logs and did not see anything.  It
just stopped logging during that time.

Do you have any idea what could have caused shibd to just stop?  I checked
all server logins and noone had logged in and stopped the process.

We are running an old version (346) of shibboleth but have many changes
coming up and are upgrading.  I know it is old.

Is there some default setting that, when reached, kills the shibd daemon? 
If there is, is there a setting that would trigger an alert that killing
event is about to happen?

Any information would be helpful.

Thank you




--
View this message in context: http://shibboleth.1660669.n2.nabble.com/shibboleth-daemon-died-with-heavy-apache-load-tp7630732.html
Sent from the Shibboleth - Users mailing list archive at Nabble.com.