Link to home
Start Free TrialLog in
Avatar of bdillingham
bdillingham

asked on

OWA on front-end server periodically stops functioning

We have OWA running on a front-end server in our DMZ, connecting to one back-end server and both running Exchange 2003 SP2. Periodically, OWA stops functioning but no services stop on either server, and all other mail flow is unaffected. The only way to get OWA working again is to reboot the back-end server. The server logs show no issues, and the IIS logs only show 500 errors when users attempt to login after OWA stops working. This is a random occurence, happening twice in the past 3 months, but only once before that in the past year. Any help would be greatly appreciated.
ASKER CERTIFIED SOLUTION
Avatar of Exchange_Geek
Exchange_Geek
Flag of India image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of bdillingham
bdillingham

ASKER

IIS works fine on the front end server, but when the issue occurs OWA straight to the back-end server does not work, nor from the front-end server (which is to be expected if the back-end OWA is not responding). The back-end server already has the /3GB and the userva=3030 set per the EXBPA tool. The front-end server has 1GB of RAM so setting the /3GB switch is not an option. The back-end server is very high-end and has plenty of resources, which makes this issue even more confusing. It was built for Exchange 2007, but it was decided to continue running 2003 for a while, so the OS was "told" via the MS recommended registry setting that the server only has 4 GB RAM even though it physically has 12 GB RAM.
It's difficult to diagnose exact what could be the problem since its taking place with such less frequency.

Ideally, I would place perfmon on the server to understand what is going wrong "at the particular time of the incident"

Apart from that it is pretty difficult to guess, unless some astrologer is consulted on this issue.
And now you see my delima! :) I'll give the monitoring a shot and see what we come up with.
Wow - that would be a wait for another 2 mths :)

Njoi.
Update: The issue has occurred a couple more times since my last post. After using your suggestions as a good starting point, I dug into the system and it turns out the non-paged pool memory was being exhausted on the back-end server, and that was causing the system, specifically OWA, to stop responding. I ended up removing the /3GB and /userva=3030 switches from boot.ini to give more resources back to the OS. This seems to have resolved the issue, at least for now. This is one of those issues that doesn't materialize every week, even though it was happening with more frequency lately. Sometimes the "best practices" aren't applicable in every situation. :) Thanks for your help.