This is my situation, I'm really hoping somebody can help me.

Our Exchange 2000 server had 2 HDD fail.  The raid couldn't be rebuilt.  We had the last full backup and subsequent incrementals so we replaced the drives and rebuilt the server and reinstalled Exchange.

Windows 2000 SP4, Exchange 2000 SP3 and Post sp3 patch

This was all fine. So then restored the full backup followed by the incrementals with the logs going to a temporary folder.

This was all fine until it started to replay the transaction log files.

It does the first 10 - 20 in a minute or so but then starts slowing down.  At the moment it is doing 1 every 40 minutes and it has 1500 to go.  I've res
tarted the whole process but no change.  The first few log files get processed very quickly but then it gets slower and slower.

Errors appearing the app log are ID 9175, MAPI error, because the SA is trying to contact the IS.  These errors appear more and more while the log proce
ssing events get less and less.  

Microsoft say these errors are normal (which I understand) but they do seem to be slowing the whole process down.

I've been at this for almost 48 hours, people are coming back to work tomorrow.  Please can anybody help me?

I have the last full backup from a week ago and all 4 incrementals since then.  I must be able to get this back somehow???

I have run eseutil /ml on some of the log files and all is ok.

Stopped all services and processes that I can.  There is no AV on the server or anything else running on it.

I have been on the phone to MS for 4 hours over this weekend, I am in the UK.  I've paid $245 but they cannot tell me what is happening.

P.S.  If I restore the full backup only and play the logs it works fine - whizzes through them and mounts the stores.  It then starts processing the incremental logs and does a few of these very quickly but then it slows down and down until it just isn't practical to carry on.

Please can somebody help me - my users are not very forgiving at the best of times and if I have to roll back a week there's going to be hell to pay!!

My email address is my username at hotmail.

Who is Participating?
OneHumpConnect With a Mentor Commented:
That's a tough one.  You certainly know what you're doing and have gone down the right support path.  I'm not sure there is much any of us can do here without sitting at your server and, even then, it would be tough.

That doesnt mean we can't try.

Here is a quick and dirty option:

1.  Give your users the data you can get them right now and tell then the rest is on the way.
2.  Restore your server to another server offline and exmerge the data out.
3.  Import that data into the production store.

This will give your users most of their data and keep their performane real-time.

If there is an alternative to that not-so-great answer, we'll have to find it by tinkering.

Fire up perfmon and load every single object.  This is going to be a pain, but it can work.  Look through your counters for excessively high values.  Try to identify the bottleneck.

One think that comes to mind is your disk performance.  I have a couple of questions about that.  Is there a chance that you have a RAID rebuilding with a high rebuild priority?  Is write-back caching enabled?  What is your storage architecture like?

Let's start there.

sarahb_workAuthor Commented:
No,  I've been through all those kinds of things and nothing is helping.  I just moved all the log files into the mdbdata file and started the eseutil /cc again.  It says not to do this but I am running out of time.  Anyway, it did more quicker, but it's still only done about 200 and the errors are popping up again and everything is slowing down....I've shut down every service I can get away with.  Maybe it's something to do with the disks, I don't know.
Problems using Powershell and Active Directory?

Managing Active Directory does not always have to be complicated.  If you are spending more time trying instead of doing, then it's time to look at something else. For nearly 20 years, AD admins around the world have used one tool for day-to-day AD management: Hyena. Discover why

sarahb_workAuthor Commented:
I ran eseutil /cc on all incremental log files (2000 of them).  This is when it slowed down to be unusable.

I ran eseutil /cc on the log files from the full backup only - this works fine.

I brought the full backup online, mounted the stores.  Dismounted them and then ran eseutil /r E00.  This only processed the E00 file.  It didn't process the incrementals which start at E000CB11.log.

I understand this is because of the break in the sequence.  

I did try to run eseutil /r E000CB11 but it didn't work.  How can I make it play the log files from this log file given that it doesn't start from E00??

Please can someone help?  There has to be a way to do this.  I'm getting evil glares from my users...:(
sarahb_workAuthor Commented:
Thanks for this, I shall look into all these things tomorrow.  I have some sleep to catch up on!  In the meantime there server is restored to a week ago, and the users have been told we will try to get the rest back asap!

Thank you again, it's great to know there's people out there willing to try and help.
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.