6.0.1 to 6.0.2 went badly, on a test machine

[Vladimir]

I was playing on my test server and ran an upgrade. When the server started up, the user side of the web interface seems broken. The Login screen and standard mode work, the advanced mode never starts, the client gets stuck on loading screen.

mailbox.log shows:

2009-10-22 16:11:29,650 INFO [main] [] StatsImageServlet - Servlet StatsImageServlet starting up
2009-10-22 16:11:29,715 WARN [main] [] log - No value for env-entry-name trustedIPs
2009-10-22 16:11:29,734 INFO [main] [] log - No Transaction manager found - if your webapp requires one, please configure one.
2009-10-22 16:11:30,078 WARN [main] [] log - No value for env-entry-name trustedIPs
2009-10-22 16:11:30,080 INFO [main] [] log - No Transaction manager found - if your webapp requires one, please configure one.
2009-10-22 16:11:30,423 INFO [main] [] log - No Transaction manager found - if your webapp requires one, please configure one.


I'm still looking around, and being a test machine not really critical, but if anyone has any ideas let me know

[Vladimir]

No luck so far. Tried all sorts of thing, running the update again, etc, etc,.

[Vladimir]

nginx complaining:

2009/10/22 17:45:44 [crit] 2396#0: *72 open() "/opt/zimbra/nginx-0.5.37-zimbra/proxy_temp/1/00/0000000001" failed (13: Permission denied) while reading upstream
, client: 192.168.216.191, server: zimtest.rutgers.edu, request: "GET /zimbra/res/I18nMsg,AjxMsg,ZMsg,ZmMsg,AjxKeys,ZmKeys,ZdMsg,Ajx TemplateMsg.js.zgz?v=0910201
50430&skin=bare HTTP/1.1", upstream: "http://192.168.170.50:8080/zimbra/res/I18nMsg,AjxMsg,ZMsg,ZmMsg,AjxKeys,ZmKeys,ZdMsg,Ajx TemplateMsg.js.zgz?v=091020150430&
skin=bare", host: "zimtest.rutgers.edu", referrer: "https://zimtest.rutgers.edu/zimbra/"

[iway]

Do you need nginx (proxy). If not, try disabling it.

[Vladimir]

My production servers do, so I need to make sure it works. In any case, I've disabled it to narrow the problem but that wasn't it.

[uxbod]

Anything additional in /var/log/zimbra.log ?

[Vladimir]

Not really. But I've changed the owner of /opt/zimbra/nginx-0.5.37-zimbra/proxy_temp* to zimbra, which made the nginx errors go away. Then I had to clear my browser before it started working.

So .. I guess it works now, but I have no idea why this went so poorly compared with previous upgrades.

[Dirk]

So far, the upgrades I've done are:
5.16 to 6.0.1,
5.18 to 6.0 and
6.0 to 6.0.2

None of which match your situation exactly. I did have to redeploy the commercial certificates to make the first one (516>601) work properly, but I dont think that's your problem.

Check your install log, grep for any error during install?

[Vladimir]

Saddly I can't look at them now, I bounced the machine at some point. To be fair, this install started with a RC1 updated to 6.0.1 and then to 6.0.2. Its entierly possible this had some RC1 history that wasn't happy. I'll worry more about this when I do a the test I really care about, namely 5.0.16 to 6.0.2 which is what I'm going to have to do to my production server.

[borngunners]

Can Someone clarify to me the steps to take in upgrading to another version. I have never done an upgrade, and I am scared to do an upgrade. I have version 6.0.1 and will like to move to the latest version. I need tips on what to keep in mind and what changes I should be expecting after the upgrade. Thanks.