I have following issue with one installation.
I have implemented a hot/cold stategy to rsync /opt/zimbra to a dr site.
Every night, when restarting zimbra, due to peering band issues (they have 2 megabits with colt and band generally is not an issue, maybe clamav mirrors have a poor peering with colt in italy), freshclam is not able to refresh clam daily db, clamav does not start, mta does not deliver messages and so on.
What I would is that when i restart zimbra, only clamv is restarted, freshclam isn't.
I don't see any valid reason to have them joint, this is a very critical point of failure, mostly for unattended stop/start operations.
I have changed zmclamdctl to take in count of frashclam delay:
Code:
90- >> ${zimbra_log_directory}/freshclam.log 2>&1 &
91: echo "sleeping 300 secs"
92: sleep 300
93- echo -n "Starting clamd..." and loop:
Code:
98: for ((i = 0; i < 120; i++)); do
99- checkrunning
but imho it would better to have another solution, not this ugly patch.
If all this makes sense, i could file an enhancement bug on bugzilla
-m