bugsaif
asked on
OWA Crashes
Aright folks, like the title says... our OWA crashes... here are some details, please ask for any information that you need.
The setup
* Exchange default setup on one box.
* Exchange 2003 SP1 is installed on a box running Windows Server 2003 SP1.
* Box is a Xeon 2.4 Ghz, 2.5GB Ram, RAID 1 (Exchange Data), Raid 1 (Exchange Logs), Gigabit NIC.
* Information store ~ 45GB
* McAfee Groupsheild 6.2
* GFI FAXMaker connector
* Clients use Outlook 2003 / Firefox / MSIE 6.0
The symptoms
* There seems to be no rhyme or reason to when it will crash. Appears to be completely at random.
* Has crashed from a frequent of twice in a week to once in 3 weeks.
* Restarting all Exchange services will not bring OWA back.
* Restarting IIS service will not bring OWA back.
* Restarting Groupshield service will not bring OWA back.
* Only restarting the box will bring OWA back.
* While OWA is down, all other exchange related services, IMAP, SMTP and POP are fully funtional and running.
* While OWA is down, port 80 on the exchange box appears to be open but instantly disconnects. Appears to be refusing connections. Checked using "Telnet ExchangeIP 80"
* No clues appear in the event logs. There are a few errors and warnings but seem unrelated.
I have System Information captures for both while owa is running and when it crashed. Nothing suspicious there...
So ask me what you want/need to know...
The setup
* Exchange default setup on one box.
* Exchange 2003 SP1 is installed on a box running Windows Server 2003 SP1.
* Box is a Xeon 2.4 Ghz, 2.5GB Ram, RAID 1 (Exchange Data), Raid 1 (Exchange Logs), Gigabit NIC.
* Information store ~ 45GB
* McAfee Groupsheild 6.2
* GFI FAXMaker connector
* Clients use Outlook 2003 / Firefox / MSIE 6.0
The symptoms
* There seems to be no rhyme or reason to when it will crash. Appears to be completely at random.
* Has crashed from a frequent of twice in a week to once in 3 weeks.
* Restarting all Exchange services will not bring OWA back.
* Restarting IIS service will not bring OWA back.
* Restarting Groupshield service will not bring OWA back.
* Only restarting the box will bring OWA back.
* While OWA is down, all other exchange related services, IMAP, SMTP and POP are fully funtional and running.
* While OWA is down, port 80 on the exchange box appears to be open but instantly disconnects. Appears to be refusing connections. Checked using "Telnet ExchangeIP 80"
* No clues appear in the event logs. There are a few errors and warnings but seem unrelated.
I have System Information captures for both while owa is running and when it crashed. Nothing suspicious there...
So ask me what you want/need to know...
Any reason you haven't installed Service Pack 2 for Exchange 2003?
Simon.
Simon.
ASKER
Hmmm... this is interesting, 3 months of looking in to the problem and no results. The day I put on EE we solve the problem.
To: Sembee
Call it careful treading... Exchange 2k3 SP2 contains significant enough changes to Exchange internals for us to wait for a while before installing. Even more so because we really don't have a test setup and just the production server.
To: Everyone else who might also be experiencing similar symptoms...
We recently deployed Quest's Spotlight on Exchange... it revealed a significant number of NDRs being generated everytime OWA went down. Exchange had been configured to forward a copy of the NDRs to an administraive mailbox, which we completely forgotten about. Another not so good thing, we don't have a limit on mailbox sizes and the said mailbox was over 2GBs in size. Purging the mailbox resulted in somthing that had never happened before... OWA sprung back to life. Turned out it wasn't really an Exchange issue. Next we tracked down the cause of the NDRs which turned out to be from one of our GFI FaxMaker servers that was misconfigured among other SPAM that had worked its way in. So now that tighter filters are in place on Exchange and our mail gateway, and the Fax server being configured properly... we are monitoring Exchange and OWA with Quest... it looks pretty stable, for now.
I'll post back in a couple of days to confirm if this really was the cause and solution to this problem. In the mean time, if you experts out there thing it ain't it... let me know...
To: Sembee
Call it careful treading... Exchange 2k3 SP2 contains significant enough changes to Exchange internals for us to wait for a while before installing. Even more so because we really don't have a test setup and just the production server.
To: Everyone else who might also be experiencing similar symptoms...
We recently deployed Quest's Spotlight on Exchange... it revealed a significant number of NDRs being generated everytime OWA went down. Exchange had been configured to forward a copy of the NDRs to an administraive mailbox, which we completely forgotten about. Another not so good thing, we don't have a limit on mailbox sizes and the said mailbox was over 2GBs in size. Purging the mailbox resulted in somthing that had never happened before... OWA sprung back to life. Turned out it wasn't really an Exchange issue. Next we tracked down the cause of the NDRs which turned out to be from one of our GFI FaxMaker servers that was misconfigured among other SPAM that had worked its way in. So now that tighter filters are in place on Exchange and our mail gateway, and the Fax server being configured properly... we are monitoring Exchange and OWA with Quest... it looks pretty stable, for now.
I'll post back in a couple of days to confirm if this really was the cause and solution to this problem. In the mean time, if you experts out there thing it ain't it... let me know...
How do you call waiting for a while? Exchange 2003 SP2 has been out for almost 18 months and is considered rock solid.
In the current security environment waiting over 12 months before deploying a service pack is not really a good idea. If there was a problem with the service pack it would be well documented and well known. The only issue is ensuring that you have removed the old version of IMF before installing the new version. Other than that, unless you have a third party vendor who doesn't support that service pack (who would also get the same question asked of why) there is no reason not to deploy it.
Simon.
In the current security environment waiting over 12 months before deploying a service pack is not really a good idea. If there was a problem with the service pack it would be well documented and well known. The only issue is ensuring that you have removed the old version of IMF before installing the new version. Other than that, unless you have a third party vendor who doesn't support that service pack (who would also get the same question asked of why) there is no reason not to deploy it.
Simon.
ASKER
Um, it crashed again... so the above explanation wasn't it...
To: Sembee
Point taken... that excuse for not upgrading was given to me by the current Exchange Admin... I'll have a word with him. So you think this problem somehow may be related to not having SP2?
To: Sembee
Point taken... that excuse for not upgrading was given to me by the current Exchange Admin... I'll have a word with him. So you think this problem somehow may be related to not having SP2?
ASKER
Quick question... will a 2gb mailbox being accessed via OWA crash it?
Considering that there were extensive changes made, plus there is an update for OWA for the change in behaviour in Internet Explorer with plugin support, I wouldn't like to go much further forwards until it has been updated.
I would service pack Exchange and then ensure that you have all the relevant updates for Exchange. There was one re-released on Tuesday night.
Simon.
I would service pack Exchange and then ensure that you have all the relevant updates for Exchange. There was one re-released on Tuesday night.
Simon.
I have users with larger mailboxes accessing the mailbox over OWA. I have sites with 4 or 5gb mailboxes that are accessed by OWA, but not to bang on about it too much... they are all on SP2. I would actually struggle to know what an SP1 server behaves like.
Simon.
Simon.
ASKER
Well the SP2 update will be installed but not right away. Its not my decision. So, Mr. Simon or anyone else that is able to... I'd appreciate if you could help me tackle this problem OR some how prove that it will be resolved by updating to SP2 in which case I can force the powers that be to update now.
I little more information regarding the issue...
We think we have narrowed down the OWA crash and can even reproduce the crash most of the time. It turns out that we also have Veritas Backup Exec 10 for Servers that takes care of backing up our Exchange server. When the backup isn't running, everything is dandy. However while the backup is running, if we access one of our 2gb mailboxes via OWA, it'll crash OWA, it will also bring the Exchange server down to a sluggish pace. The event log is registering event 510's. If we manage to login to the Exchange server and kill the backup agent, OWA and Exchange recover. Now it may sound like a Veritas issue but we also have WinNT Backup (I think thats what its called) and even though we haven't yet tried to test this because we ran out of time on our maintanence window... we have reports of OWA being unavailable while the WinNT backup was running too. So now we think its either Group Shield or Exchange itself... any ideas??
I little more information regarding the issue...
We think we have narrowed down the OWA crash and can even reproduce the crash most of the time. It turns out that we also have Veritas Backup Exec 10 for Servers that takes care of backing up our Exchange server. When the backup isn't running, everything is dandy. However while the backup is running, if we access one of our 2gb mailboxes via OWA, it'll crash OWA, it will also bring the Exchange server down to a sluggish pace. The event log is registering event 510's. If we manage to login to the Exchange server and kill the backup agent, OWA and Exchange recover. Now it may sound like a Veritas issue but we also have WinNT Backup (I think thats what its called) and even though we haven't yet tried to test this because we ran out of time on our maintanence window... we have reports of OWA being unavailable while the WinNT backup was running too. So now we think its either Group Shield or Exchange itself... any ideas??
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER