Hi,

we are running Zimbra 5.0.15 as a mailservice here and have problems with the stbillity of the http(s)/imap(s)/pop(s) proxies.
We see the nginx workers on the boxes using after a while 100% CPU and if all workers are on 100% on one of the CPU's the Nginx does not respond at all to requests.

The OS we are running is Red Hat Enterprrise Linux 64, Nginx Version 0.5.37, memcached 1.2.5.

Has somebody of you seen any similar?

Any help or any suggestions is very much appreciated.

Cheers
Markus