last 4-5 months i got experience with multi-serer installation of Zimbra, and I think that my current production system is working really good. To share what I've done, i'll start that i had to migrate from Zimbra 6 something, to newly hardware cluster, and the Zimbra installation contains over 12 000 accounts. Using some knowledge over the great Zimbra Forums and Wiki, i was able to migrate the whole cluster to Zimbra 8.0 without loosing even one email
At the moment, my issue is not critical, but at some point i'll need to stop experimenting with the production server, but it's bothering me way too much. OK, i have a cluster system with 3 mailbox machines, 1 LDAP, 1 MTA and of course 1 Proxy. Everything on the admin panel indicates that server is running really good, and since 2 months now, it's in production. All of the servers pass their logs to the server which is setup to take this role (only 1 logger server), users, their mailbox and preferences are up-to-date and i didn't had any problems at all.
To the point, the proxy server is configured to redirect to HTTPS connection, but once i make a change from the admin panel (running on 9071 on the proxy), the change is saved, but for at least 3-4 minutes after this action - the proxy server on port 9071 is returning error 502 Bad Gateway.
Tried to locate the problem withing nginx, but i couldn't find anything more than just a report for failed connection. Have in mind that the user web panel is working perfectly during the time when admin panel is returning 502 Bad Gateway error.
I've read some info that this might be a nginx problem due to low timeout value, but it wasn't.
If anyone faced such issue, or probably can help me further understand why can this happen, please, report something
Now the info:
CentOS 6.5 (all servers are using the same software)
No iptables settings at all, no router envolved.
Anything security related (like selinux or other software) is disabled
Thank you in advance!