Exchange cluster stops OWA after fail over...
So, I have an active/passive cluster, Exchange 2003 SP1, Windows SP2. For the longest of time I have had the problem that everything [messaging, OWA, mobile active sync (push)] processes perfectly running on node 1. But, when I fail over the server to node 2, OWA, and active sync stops working. Messaging processes work fine on node 2 but not the OWA and active synch stuff. When I move the acitive VS back to node 1 'everything' begins to work again.
I have copied the IIS config.xml for Exhange which ensures everything is the same on both servers. Man o man, I can't figure it out.
Any help would be appreciated. Oh, I really need node 2 working fully, so that I could upgrade to SP2. I want it all working before upgrading the passive node and verify all processes continue to work after the SP2 upgrade.
Thanks in advance for any suggestions.
April 17th, 2008 11:51pm
are all the IIS services coming on line when you failover? Has this always happened or has it just started? Anything of interest in either node's application log or cluster.log?
Free Windows Admin Tool Kit Click here and download it now
April 18th, 2008 12:07am
Hey, thanks for replying!
Yes, all services are starting and no errors are found in the event viewer.
I guess the problem has always occurred, but I did not notice during fail over because I was usually occupied with bringing node 1 back online and the 'remote' features were not a priority.
I restarted the node 2 of the cluster this morning at 7 AM (thus the log entries).
I think the below entries are because the IIS focus is on the active server (node 1).
After bouncing the passive node (2), I failed the cluster over this morning and pulled the following from the cluster log:
[INFO] [IISCLUSCFG] Could not delete resource type "IIS Server Instance" because the resource type was not found. Trying the next resource type in the table... (hr = 0x800713d6) 2008-04-18 07:02:35.625 [INFO] [IISCLUSCFG] Could not delete resource type "SMTP Server Instance" because the resource type was not found. Trying the next resource type in the table... (hr = 0x800713d6) 2008-04-18 07:02:35.638 [INFO] [IISCLUSCFG] Could not delete resource type "NNTP Server Instance" because the resource type was not found. Trying the next resource type in the table... (hr = 0x800713d6) 2008-04-18 07:02:35.638 [INFO] [IISCLUSCFG] Could not delete resource type "IIS Virtual Root" because the resource type was not found. Trying the next resource type in the table... (hr = 0x800713d6) 2008-04-18 07:02:35.638 [INFO] [IISCLUSCFG] Could not delete resource type "Time Service" because the resource type was not found. Trying the next resource type in the table... (hr = 0x800713d6) 2008-04-18 07:02:35.638 [INFO] [IISCLUSCFG] Leaving startup notify... (hr = 00000000)
This is a tough one!
Thanks for 'thinking' about this with me.
April 18th, 2008 7:22pm
THat looks like the service isn't shutting down cleanly on failover; that's just an educated guess though. Is there anything in the SP2 release notes that show a fix for that?
Free Windows Admin Tool Kit Click here and download it now
April 18th, 2008 7:48pm
I discovered the solution to the problem and thought I should post here. As it turns out the certificate apparently had not been applied correctley or was not updated when the cert was renewed. There were no errors in the logs indicating a cert problem but just to try something, I exported the cert fromnode 1 (the working node), imported to node 2. On fail over the active sync service still did not work. So, after failing back to node 1, I then restarted node 2. I failed over again, and this time all services including OWA, active sync, etc worked.
Since all services are running on both nodes I moved forward on the upgrade to Exchange Service Pack 2. All went well with the upgrade and active sync stayed intact and continued to work on both nodes. But (there is always a but) I now have a post SP2 upgrade issue. I will open a new thread for the new issue.
Thanks for your help.
May 13th, 2008 3:48pm