[horde] memcache session handler

[Kevin M. Myer]

Quoting liamr at umich.edu:

> Since we had to turn the log level way down to prevent the session
> handler from core dumping, I don't have a lot to go on.

The core dump with logging set to PEAR_LOG_DEBUG is still puzzling to 
me.  I didn't see a single explanation for why PEAR_LOG_NOTICE works 
but PEAR_LOG_DEBUG causes a core dump.

> Any ideas what would trigger this situation and how it could be avoided
> in the future?

Are there any host-based or network-based firewalls that would be 
terminating network connections?  Any chance you are running out of 
either available local network ports (cat 
/proc/sys/net/ipv4/ip_local_port_range to see how many you are 
configured for) or that you have a bunch of connections tied up in 
FIN_WAIT state?  Are you using persistent connections?  Finally, any 
chance that you simply exhausted the small amount of memory you 
allocated on each box?  In other words, is the amount of memory you 
allocated to memcached enough to handle the amount of data that 15 
servers need to store about sessions?  Maybe GC isn't working properly 
and the old session data isn't getting cleared out properly.

Kevin
-- 
Kevin M. Myer
Senior Systems Administrator
Lancaster-Lebanon Intermediate Unit 13  http://www.iu13.org