I noticed this morning that amavis was dieing every few mins. It would scan for a while (which i can watch in real time to tell if something is moving or not). Also i would watch the amavis proccess in top. After scanning for a roughly 3-5 mins amavis would completely die. ps shows there are still proccess

postfix 21569 0.0 0.0 5800 1664 ? S 12:51 0:00 smtp -n smtp-amavis -t unix -u -o smtp_data_done_timeout 1200 -o smtp_send_xforward_command yes -o disable_dns_lookups yes -o max_use 20
zimbra 21576 0.0 2.2 55288 46312 ? S 12:51 0:00 amavisd (ch1-avail)
zimbra 21578 0.0 2.2 55288 46312 ? S 12:51 0:00 amavisd (ch1-avail)
zimbra 21579 0.0 2.2 55188 46144 ? S 12:51 0:00 amavisd (virgin child)
zimbra 21580 0.0 2.2 55188 46140 ? S 12:51 0:00 amavisd (virgin child)
zimbra 21611 0.0 2.2 55188 46136 ? S 12:51 0:00 amavisd (virgin child)
zimbra 21638 0.0 2.2 55188 46140 ? S 12:51 0:00 amavisd (virgin child)
zimbra 21660 0.0 2.2 55188 46140 ? S 12:51 0:00 amavisd (virgin child)
zimbra 24357 0.0 0.0 2340 476 pts/2 R+ 12:56 0:00 grep amavis

However all amavis proccesses stop showing up in top and the scanning stops as well. (2 machines dedicated filter cluster) Mail delivery to mail store also stops

Errors are showing in /var/log/zimbra.log

May 9 12:51:56 black amavis[21572]: (21572-02) Requesting process rundown after fatal error
May 9 12:51:56 black amavis[21573]: (21573-02) Requesting process rundown after fatal error
May 9 12:51:57 black amavis[21577]: (21577-02) Requesting process rundown after fatal error

Also noticed when doing a zmcontrol stop

Stopping mta...FAILED
/opt/zimbra/cyrus-sasl-2.1.21.ZIMBRA/sbin/saslauthd: no process killed

Start appears to work ok, and status shows everything is running