Hi all,

We set up apache with mod_proxy based off this article Multiple SSL Virtual Hosts - Zimbra :: Wiki so that we can provide multiple SSL virtual hosts, however we found it far too slow.

So we have switched to nginx and it seems a lot better, but looking at the nginx error.log I see this going on all the time:

Code:
2011/06/09 09:29:42 [error] 15016#0: *63490 upstream timed out (110: Connection timed out) while reading response header from upstream, client: xxx.xxx.xxx.xxx, server: zimbra.mydomain.example, request: "POST /service/soap/WaitSetRequest HTTP/1.1", upstream: "http://10.150.1.61:8888/service/soap/WaitSetRequest", host: "zimbra.mydomain.example"
2011/06/09 09:29:44 [error] 15016#0: *63550 upstream timed out (110: Connection timed out) while reading response header from upstream, client: xxx.xxx.xxx.xxx, server: zimbra.mydomain.example, request: "POST /Microsoft-Server-ActiveSync?User=user@mydomain.example&DeviceId=Appl8793158xxxx&DeviceType=iPhone&Cmd=Ping HTTP/1.1", upstream: "http://10.150.1.61:8888/Microsoft-Server-ActiveSync?User=user@mydomain.example&DeviceId=Appl8793158xxxx&DeviceType=iPhone&Cmd=Ping", host: "zimbra.mydomain.example"
I'm guessing this is the keepalive mechanism/polling for new messages. Is this a bad error? If not, can I suppress it somehow?