I get a blank page when trying to access the admin web UI

[dwmtractor]

I'm going out on a limb here, but I just reread your original post and have another idea. In the past, I have seen Firefox do weird things (like, blank screen weird things) when I tried to access a secure site (such as a webmin page) for some site where the hostname and IP were the same, but the certificate had changed due to a restore or upgrade of software. This has happened on my firewall as well as Zimbra, I think. You might try going into your certificate stores (and root certs path) on your client machine, and deleting all certs for your Zimbra box and/or its IP, so that next time you try to access the UI you have to re-load the cert. It's a long shot but it might do the trick.

Dan

[bubarooni]

woo hoo!

i'm taking any victories i can get here...

[bubarooni]

victory in that case being able to upload that file not a non-blank admin page.

i'll look into dwmtractor's suggestion too. i don't know if that is the same thing as clearing the cache but will try anything at this point. the problem occurs on other machines and in ie 6 too.

also, hopefully this will make it easier to diagnose and not harder, trying to login as one of the test user accounts results in the infamous 'A network service error has occurred (zclient/io).'

[bubarooni]

oh, i missed this part in the firebug:

uncaught exception: AjxException
before_unload(beforeunload )

[dwmtractor]

Originally Posted by bubarooni

i'll look into dwmtractor's suggestion too. i don't know if that is the same thing as clearing the cache but will try anything at this point. the problem occurs on other machines and in ie 6 too.

I can't say about the rest of your issues, but the certificates and cert. paths are NOT cleared by clearing the cache in either Firefox or IE (I've never used Safari). The reason I suspect certificates to be the problem is your comment (if I understood you correctly) that you are able to log on to the admin UI from other client machines, but not yours.

In Firefox the certificate stores can be found by going to Tools > Options, click on the Advanced tab, and the View Certificates button. Go to the Web Sites tab and find any references to either the hostname or ip address of your server and delete them. This will merely force you to accept certificates again when next you try to access the site.

In IE go to Tools > Internet Options, choose the Content tab and click on the Certificates button. Look through all the tabs, but on my client machine (XP Pro, IE6) it's installed in the "Trusted Root Certification Authorities" store. Again, just delete the certificate.

Hope this helps,

Dan

By way of clarification--if my theory is right, this problem is happening on any machine that ever logged successfully into the old (pre-upgrade) server, but if you go to a "virgin" machine or VM, you'll be able to get in without a hitch. If that virgin machine can't get in either, I'm on the wrong track.

[bubarooni]

well, i get the same error on the server itself and on my workstation. cleared the certs on both same error. went to a machine where i'd never accessed the zimbra machine's ip before and same thing. it was worth a shot, gave me hope and eliminated another variable.

[dwmtractor]

Originally Posted by bubarooni

well, i get the same error on the server itself and on my workstation. cleared the certs on both same error. went to a machine where i'd never accessed the zimbra machine's ip before and same thing. it was worth a shot, gave me hope and eliminated another variable.

OK, so much for my theory.

I'm really confused by this statement in your OP:

I can fire up VMware and access the admin UI fine through IE or FF in the windows session.

Is this still true? Where is the VMware session running? Is there anywhere else you CAN access the UI? If so then what, if anything, do the successful clients have in common that's different from the unsuccessful ones? I ask because it seems tough to think that the UI would work *anywhere* if the problem were a server-side problem. On the other hand I grant that the problem occurring so widespread makes it less likely for it to be client-side. . .

Try this: Log on using your successful VMware login, log back out and try to log in from one of the failing machines. Then tail /var/log/zimbra.log and see if you can find any indication of the two events that might show you the success or failure.

At the risk of going for the painfully obvious, you *have* done a stop and restart of Zimbra services, yes? Any errors show up in zimbra.log during the restart?

I freely acknowledge I'm grasping at straws here, so feel free to laugh maniacally at any of my suggestions

[bubarooni]

OH NO! That wasn't me. I'm just tagging on to this discussion. I'm running Zimbra 5 on CentOS 5 only. Nothing fancy. That would obviously be way beyond my skill level!

[bubarooni]

i'm getting desperate though. i may just have to do a fresh install so i can maybe have it up by Monday morning.

[dwmtractor]

Oh, sorry for my (considerable) confusion. Are you able to access the UI at all? What does a command-line zmcontrol status tell you?

However these diagnostic steps may still be worth doing for you too:

Try to log in from one of the failing machines. Then tail /var/log/zimbra.log and see if you can find any indication of the event that might show you the success or failure.

At the risk of going for the painfully obvious, you *have* done a stop and restart of Zimbra services, yes? Any errors show up in zimbra.log during the restart?

And since you said CentOS 5 (and without trying to bring up a lot more heat) have you checked the perl issue that caused all the headaches a couple days back?