Exchange 2003 Emergency

dynamik

A client of ours lost the drive their Exchange DB was on (system drive is ok). It was configured for RAID, but for some reason it's not usable. The drives are being sent in for data recovery, but I need to get them up and running ASAP. I assume I would start by creating a new storage group and then a new a mailbox store. Edit: I guess I can only have a single mailbox store in the standard edition. Do I have to delete the existing one or can I do something as simple as pointing it at a new location?

I'm going to research this a lot more myself, but if anyone has any tips, I'd certainly appreciate it. They of course want this up by tomorrow AM, so I'm a bit scatter-brained at the moment.

TIA.

NetAdmin2436

Ouch!

....Let me guess, they don't have any backups.

dynamik

Not really

Claymoore

You can only have one storage group in the standard edition, but you can have multiple message stores. That's not really important here though.

What you want to do is create a dial-tone database to restore service. If you can re-create the storage path with other storage, you can try to re-mount the database from system manager. Since there is no message database there, it will warn you that it will create an empty database. As messages are received, the mailboxes will be created. The employees will be able to send and receive mail, but they won't have access to old mail.

When the drives and original database are recovered, you will have to move the databases around and merge data using the Recovery Storage Group. You'll dismount and copy the dial-tone database to a safe location, then copy the recovered database back to the original location (using the 'replace this database with a backup option'). Then you will mount the dial-tone database in the recovery storage group and merge the new mail with old database.

It's high level, but I hope that gets you started. Good luck, and keep us updated.

Edit: Dial Tone Links
http://technet.microsoft.com/en-us/library/aa998947(EXCHG.65).aspx
http://www.msexchange.org/tutorials/Exchange-Dial-tone-Restore-Method-Part1.html

rjbarlow

What's not usable, the entire survivor drive in the RAID or just the Information Store? If it was a mirror and the databases are available You could try the fortune running Eseutil with the /r parameter and trying then to mount the Information Store, next the /p parameter if the /r does not produce results.
The /r parameter is intended for re-doing the databases in a "consistence" state, the /p parameter tries to recover corruptions.

The Eseutil /r needs to run against the "Transaction Log" files (ex. E00.log), while the Eseutiil /p needs to run against the databases (file "edb" and ".stm")
Eseutil /r
Eseutil /p

Of course is not like eating an ice cream.

P.s. Prior to try any operation on databases and Log files remember to make a whole copy of all the Storage Group files.

---

If nothing of that, then the solution explained by Claymoore is a very well work around.

dynamik

Ok, everything was gone, but now we've got it back. We moved the three good drives over to the next four ports (these are SAS drives, so there's eight ports which are split into two groups of four for each SCSI port), and Windows booted right up. Bad cable or port I guess. So now I find myself in an interesting situation because I followed Claymoore's suggestion (which worked like a charm), except now I need to put the old server back into production and move the new items they've received since this morning into that

Oh well, it was quite a relief for them because they would have lost about 70 days worth of data and all of their Exchange.

wedge1988

Maybe now you can suggest a backup media to them and say you'll install it for a small fee

I bought symantec backup exec for where i work, and i am amazed that its a good product. Ive tried tons of backup software and it seems to be the best of the bunch for me (BTW, i hate symantec AV, PF etc, they suck)

Claymoore

+1 to rjbarlow's recommendation to copy the files (and work on copies instead of the originals). Copying the databases probably saved my job once when I was performing a recovery.

You will need to move the databases around and use the Recovery Storage Group to mount the dialtone database and merge the messages from the RSG to the recovered original database. Since the dialtone database is much smaller, merging data from it to the recovered database is faster. Technet explains it better than I can.

Configuring the Recovery Storage Group
Swapping Databases between the Recovery Storage Group and the Original Storage Group
Merging Data between Swapped Databases
Benefits of the Messaging Dial Tone Strategy

Once everything is back to normal, work to develop a backup, archive and recovery plan. Just make sure whatever backup software you recommend works with the RSG or supports single mailbox recovery. Streaming backups won't always work with the RSG so check to see if the backup software uses the Exchange VSS writer.