Eugene
Thu Nov 15 10:28:16 PST 2007
hi Phil,
As there're several possible causes, it's helpful to use an approach that
could eliminate possibilities.
Defragging the store does't fix corruption but helps to free up space.
If you fix the corruption and it recurs, this is indicates that something
else
is contributing to the problem.
Consider the possible causes of the problem.
Hardware RAID which is transparent to the OS is good, but is the h/w
fast enough or are there no issues with it? If there are, it may not show
up with files, but it may occur on databases such as Exchange. And the
problem will still recur after repairing.
Since you're using some archiving software, is it possible to temporarily
disable or uninstall?
Clean shutdown means consistent. It doesn't mean the store has no problems.
Try to understand what each process is trying to check and accomplish,
don't just follow. I know there is much to grasp. The KBs provide an
overview and some background.
Do you still have the problem email(s) that can't open? Can it be deleted?
Use Isinteg to check for problems in the database. You may need to run
several passes (use a few times). Again, make a backup of the original
files because you may need to restart alternate repairs with this set.
If this fixes the issues, then you're good to go. You'll know if you
mount the store and the bad files are gone.
Eseutil /p is a 'hard' repair - it dumps data in the logs not yet committed
to the store. But it helps fix the store by ignoring what was not comitted.
You need Isinteg to fix problems ESEutil /p causes.
Use this step if the above doesn't solve the problem emails.
The bad emails may still appear but you shd be able to delete them.
HTH,
Eugene Tan
SBS MVP
================================
"Phil Partridge" <philp@pebble.demon.co.uk> wrote in message
news:qmMDzBACLCPHFwZq@pebble.demon.co.uk...
> Right,
>
> I will try to answer these points in a little more detail..
>
> In article <WsIsxAAV34OHFw2J@pebble.demon.co.uk>, Phil Partridge
> <philp@pebble.demon.co.uk> writes
>>In article <qQGjeDAvLkOHFwyu@pebble.demon.co.uk>, Phil Partridge
>><philp@pebble.demon.co.uk> writes
>>>In article <#kp2ssOJIHA.4592@TK2MSFTNGP02.phx.gbl>, Eugene Tan
>>><TechHelp-at.insights.com.sg@newsgroups.nospam> writes
>>>>hi Phil,
>>>>
>>>>I'm not sure if I understand correctly the reason for the new disks.
>
> The server has a small (10GB) C: partition.
> Mdbdata was on the data partition, and there was not a lot of free
> space..
> Adding the extra disks gave me space to defrag (without moving data,
> USB1.1, slow!) and tapes not big enough to backup data and exchange, so
> mirror better than nothing. - Not much I know.
>
>>>>
>>>>A pair of 36GB SCSIs for Exchange provides good I/O and RAID1
>>>>is helpful for a mission critical function that email is.
>>>>But it is not necessary for ESEutil - the temp drive/folder for this can
>>>>be any other drive, preferably not USB, and a different drive often
>>>>gets some performance benefit though only during this troubleshooting.
>>>>
>>>>Since you're not sure the problem occured before or only after you
>>>>move the mail store to the new drives, you'll probably want to deal
>>>>with all the suspects. It could be another app interfering like GFI.
>>>>I've also found that low-end RAID controllers can cause trouble esp.
>>>>those that don't have their own cpu. What are you using?
>
> Onsite later today, so will get as much detail as I can.. OS doesn't
> know about the RAID, done at a lower level.
>
>>>>
>>>>Exchange2000 is limited to 16GB. The registry tweak to temp
>>>>increase the size is meant for recovery only, it resets back to 16GB
>>>>when you reboot. Don't leave in this state unless you are prepared
>>>>to have an unrecoverable store.
>>>>As a guide, watch how fast the store grows each week, and leave
>>>>at least a 2 month buffer. Generally, alarms bells went off whenever
>>>>the store exceeded 13.5GB.
>
> Sorry, I was not clear on this.. Only used to 'get at' the db, not using
> all the time.
> 13.5GB luxury! After defrag last night;
> priv1.ebd = 10,137,640
> priv1.stm = 4,523,232
> pub1.edb = 1,184,840
> pub1.stm = 8,200
>
>>>>
>>>>Since the store is damaged, it also needs to be fixed. ESEutil -d only
>>>>defrags. ESEutil /? to find out the other options.
>>>>In repairing an Exchange store, there's probably quite a lot to grasp
>>>>so pls read before you use any of the tools. You'll use some of the
>>>>tools to identify where errors may exist, and note that it doesn't
>>>>necessary fix everything in 1 go/pass.
>>>>Before you do any of the following, I'd advise you make a copy of all
>>>>of the files of the Exchange store - a IDE/PATA/SATA drive would
>>>>suffice, which could later also serve as a fast but low reliability
>>>>backup.
>>>>
>>>>I forgot to include references earler.
>>>>
>>>>We're looking mainly at the first half of this KB313184:
>>>>How to recover the information store on Exchange 2000 Server or Exchange
>>>>Server 2003 in a single site
>>>>
http://support.microsoft.com/kb/313184/en-us
>
> Stopped Information Store..
> Eseutil /mh reported 'clean shutdown' on both databases. - To me this is
> not the same as 'consistent' as stated in the KB.
> Is this OK or not?
>
>>>>
>>>>How and when to manipulate Exchange transaction logs in disaster
>>>>recovery
>>>>
http://support.microsoft.com/kb/812592/en-us
>>>>
>>>>XADM: STM Database Files Are Inconsistent After You Run the Eseutil /p
>>>>Command
>>>>
http://support.microsoft.com/kb/327346/en-us
>>>>
>
> What should tonight's task be?
> Eseutil /p ? Eseutil /r ?
> KB also mentions Isinteg. - Should I run this.
>
> This is all a bit scary! ;-)
>
>
> Thanks for any feedback,
> Phil Partridge.
>
>>>>
>>>>And something to read about for 3rd party products working
>>>>with Exchange:
>>>>Microsoft support policy for third-party products that modify or extract
>>>>Exchange database contents
>>>>
http://support.microsoft.com/kb/904845/en-us
>>>>
>>>>HTH,
>>>>Eugene Tan
>>>>SBS MVP
>>>
>>>Thanks for the extra info..
>>>Lots to read up, and need to do things out of hours, so will be a day or
>>>so before I get back to the group..
>>>
>>>Thanks again,
>>>Phil Partridge.
>>>
>>
>>New disks were to give room to run a defrag.. No space on the rest.
>>
>>So far, so good.. Eseutil /mh shows clean shutdown on both pub and priv.
>>
>>>>
>>>>=============================
>>>>"Phil Partridge" <philp@pebble.demon.co.uk> wrote in message
>>>>news:GNK4ZAATn5NHFw+H@pebble.demon.co.uk...
>>>>> In article <ugdeiNIJIHA.484@TK2MSFTNGP06.phx.gbl>, Eugene Tan
>>>>> <TechHelp-
>>>>> at.insights.com.sg@newsgroups.nospam> writes
>>>>>>hi Phil,
>>>>>>
>>>>>>Jumping in here a bit late but I've seen something similar on a SBS2k
>>>>>>box.
>>>>>>It was a problem with the Exchange store and it my case, it was caused
>>>>>>by
>>>>>>the RAID array firmware and BIOS.
>>>>>>
>>>>>>The symptoms are similar - there're some emails which are unable to
>>>>>>open,
>>>>>>and Outlook appears to lockup. In my case, most times the email
>>>>>>subject
>>>>>>appears to be spam, reported as several hundred MB! And Outlook
>>>>>>eventually recovers, saying there's a problem retrieving. I didn't
>>>>>>try
>>>>>>OE/POP
>>>>>>but since the email can't be retrieve I wouldn't be surprised about
>>>>>>the
>>>>>>errors.
>>>>>>That SBS box was also using POP Connector, but am unsure if there's
>>>>>>any
>>>>>>connection.
>>>>>>
>>>>>>I concur with Merv's thinking it's a malformed mail or something
>>>>>>interfering
>>>>>>with the Exchange store.
>>>>>>
>>>>>>Can you provide some specs on the hardware? CPU, RAM, storage size
>>>>>>partitions and RAID (if any)?
>>>>>>
>>>>>>For a solution, use an isolation strategy. You may have to consider
>>>>>>disabling
>>>>>>GFI temporarily. For antivirus, ensure you exclude the Exchange store
>>>>>>folders
>>>>>>and use one that is VAPI compatible.
>>>>>>Use ESEutil over several passes to find out what might be wrong, fix
>>>>>>as
>>>>>>the
>>>>>>clues indicate. In my case, I eventually used the DR recovery method,
>>>>>>and
>>>>>>it was fixed though the problem returned after a while - since it was
>>>>>>due
>>>>>>to
>>>>>>a h/w error.
>>>>>>
>>>>>>IMO, using the PST method won't work because the corrupt file won't be
>>>>>>retrieved to a PST, and when you restore the PST to the rebuilt
>>>>>>Exchange
>>>>>>the mail would occupy more space (since this would defeat
>>>>>>single-instance
>>>>>>store) and your box does sound 16GB space challenged. But no harm to
>>>>>>try to create PSTs if you have the time.
>>>>>>
>>>>>>HTH,
>>>>>>Eugene Tan
>>>>>>SBS MVP
>>>>>
>>>>> Eugene,
>>>>>
>>>>> Thanks for the input..
>>>>> This may be due to a change I had instigated to ease my life..
>>>>> Had a pair of 35GB disks fitted to put the Exchange store on (simple
>>>>> mirror, low level, OS does not know about it). - This seemed the most
>>>>> expedient way to; a) Get enough free disk space for running Eseutil.
>>>>> b)
>>>>> mirrored, as the tape sub-system is not big enough (even with
>>>>> compression) to backup Data + Exchange.
>>>>> I know I am only protected from a disk failure, but better than the
>>>>> nothing of before!
>>>>>
>>>>> I am not sure if the problem started before the new disks were
>>>>> fitted..
>>>>> Eseutil -d, I can do in my sleep now! ;-}
>>>>> I admit the other options (well Eseutil generally) are 'black magic'
>>>>> to
>>>>> me.. so far the -d option has been fine to claw back some extra
>>>>> space. -
>>>>> Also have the registry tweak in there foe if I need the store to go to
>>>>> 17.5GB.. Had to do this to get an important email on one occasion, but
>>>>> I
>>>>> know it is not a good thing to rely on.
>>>>>
>>>>> What switches do you suggest I hit it with, and what do they do /
>>>>> mean?
>>>>>
>>>>> Thanks again,
>>>>> Phil Partridge.
>>>>>
>>>>>>
>>>>>>================================
>>>>>>"Phil Partridge" <philp@pebble.demon.co.uk> wrote in message
>>>>>>news:RM3tJDAjw4MHFwFA@pebble.demon.co.uk...
>>>>>>> In article <#TjFgagIIHA.4228@TK2MSFTNGP02.phx.gbl>, Merv Porter
>>>>>>> [SBS-
>>>>>>> MVP] <mwport@no_spam_hotmail.com> writes
>>>>>>>>For the OL 2000 and OE clients, can you use Shift-Delete (rather
>>>>>>>>than a
>>>>>>>>just
>>>>>>>>Delete) to remove the bad mail from the mailboxes?
>>>>>>>>
>>>>>>> Merv,
>>>>>>>
>>>>>>> There seem to be two differing 'faults'. - Am not on site so bear
>>>>>>> with
>>>>>>> me..
>>>>>>> Outlook 'locking up' on certain messages, is on PC's which the Users
>>>>>>> are
>>>>>>> using.. There are (at least) 2 machines where this is happening. - I
>>>>>>> can
>>>>>>> try Shift-Del when next onsite..
>>>>>>>
>>>>>>> OE grabs mail on the Server (POP) from exchange, these are the un-
>>>>>>> deliverables this gives over 100 errors, where it can't find the
>>>>>>> mail
>>>>>>> item..
>>>>>>>
>>>>>>> Message number 1 could not be retrieved. Account: 'Administrator'.
>>>>>>> Server: '10.0.0.2'. Protocol: POP3. Server Response: '-ERR Access
>>>>>>> denied.', Port: 110, Secure(SSL): No, Server error: 0x800CCC90,
>>>>>>> Error
>>>>>>> Number: 0x8004420CD
>>>>>>>
>>>>>>> Any ideas?
>>>>>>> Philip Partridge
>>>>>>
>>>>>>
>>>>>
>>>>> Philip Partridge
>>>>
>>>>
>>>
>>>Philip Partridge
>>
>>Philip Partridge
>
> Philip Partridge