No announcement yet.

Corrupted Database Exchange 2003

  • Filter
  • Time
  • Show
Clear All
new posts

  • Corrupted Database Exchange 2003

    Hi helpful people,
    I have a question I’m hoping somebody will be kind to explain please. In a disaster recovery situation, when the SG database got corrupted (hardware related) and backups run with errors, would it make sense to create a new Storage Group and move mailboxes into it? Would this be a possible "quick fix", a workaround for clearing database corruption? The reason I’m asking is for the sake of reducing downtime for the users affected. Eseutil /P on this database took about 4 hours for 36Gb, after which we decided to cancel it as failed. We’ll be trying to repair db again, but in the meantime we would like to give users back e-mail service. Can anybody please advice if by moving mailboxes into blank storage group we might “eliminate” the corruption issue?
    *Exchange 2003 SP2 on a W2K3 SP1 domain member server, Windows2000 AD.
    Many many thanks in advance.
    Experience is something you don't get until just after you need it

  • #2
    Re: Corrupted Database Exchange 2003

    Just wanted to clarify, do you mean the mailstore that got corrupted? I've never come across a situ where the storage group structure got corrupted...


    • #3
      Re: Corrupted Database Exchange 2003

      Assuming that the database is corrupted... can the mailstore still be mounted?

      With a hard disk or controller failure, the corruption will depend on where the heads were on the disk. But it's likely to have damaged data at the mailbox level.
      Not all mailboxes will have been affected. Messages that were being written to the databases from the logs might likely be affected.

      Of course, if it's a head dive into the drive then the whole database structure is likely to have been affected, leading to the mailstore not being mountable. Especially as you say that ESEUTIL has failed.

      OK. If the mailstore is still mountable, then it would indeed make sense to try to move mailboxes to a new database (could be in the same storage group if desired). You'll rescue perhaps most mailboxes. You won't be able to move the mailboxes that are out of shape from the corruption - although there is an option in the Move Mailbox where you can ignore errors - set this to maybe 30.

      If the mailstore is not mountable, and ESEUTIL /p won't work, then I'm afraid you are staring at a restore. You could use the dialtone method to restore empty mailboxes to your users. It'll keep them working whilst you recover from a backup.

      good luck and regards,



      • #4
        Re: Corrupted Database Exchange 2003

        Thank you very much, Robbyb!

        Yes, I was talking of the mailstore. we have three here, two have been corrupted, one of which we were able to re-mount, but still unable to back up properly, so I'll move the mailboxes out of it, while still can.
        the other one - still no luck in mounting. So, the dialtone would be my best bet, I think. I'm thinking of using something like Ontrack powercontrol tools to try to recover the mailboxes from .edb

        when you said: "could be in the same storage group if desired" - do you mean that new database will not be corrupted even if it shares transaction logs with the corrupted one if it is in the same SG? what would happenned to the new database if I'll run the restore of the corrupted one that is in the same SG? I thought, it has to be in separate storage group for this reason?

        thanks a lot for the dialtone tip - I didn't know of it before. So going to try, at least to keep the mailflow. By the way, can you tell from your experience, what happenns to the messages, if they come to the server and there is no mailbox availiable? do they returned to sender with NDR? or are they lost without notice? what is the best way to keep the message loss to the minimum, while restoring database?
        thanks again for your help!
        Experience is something you don't get until just after you need it


        • #5
          Re: Corrupted Database Exchange 2003

          Hi Lana,

          Well, I was thinking of a new database completely there. It'll be unrelated to the old one. So, it could go into the same storage group. But I understand more of your problem now. As long as you are happy that the drive that contains the logs for those databases hadn't crashed, you can leave the new one in the same SG.

          But if that transaction log drive had crashed, or you are otherwise not confident in that drive (perhaps it is on the same physical disk as the crashed database was) then you are right to be concerned. You'll then have two choices:
          1. Use the existing SG but move the logs to a new drive.
          2. Create a new SG with logs on a new drive.

          I'd go for No. 2 because you'll otherwise be moving potentially corrupt logs. You'll also want to leave the old logs where they are to continue to try to recover any databases you can.

          On the mailflow question, the messages will just get queued on the server. Messages that are in the queues for long enough (usually 2 days if no one has changed the settings) will time out and then return a NDR.

          The best way to keep the message loss to the minimum might be to take the following action:

          1. Set the time on your Default SMTP to more than 2 days. If you change it now, unfortunately, those in the queue won't be affected by the change - they'll still be counting to two days.

          2. Implement the dialtone, if this is the way you'd like to go. Doing the dial tone will give your users back an empty mailbox in no time. Once you do that, any messages in the queue should be delivered. Any future mail will be delivered. This'll take any pressure you get from your users off quite a bit, because they can at least work; they'll then be patient while you recover their mailboxes!