1

Every 3 to 6 months we start losing connection to exchange from outlook (i.e Outlook has lost connection to exchange -> Outlook has restored connection to exchange). After this happens the mailbox database inevitably becomes dismounted and won't mount again without me doing something with eseutil to repair it, sort the logs out etc. Sometimes, I need to restore the backup from the night before. In all cases eseutil /mh reports the database to be in a Dirty Shutdown State.

We are small company (20 employees) running SBS 2011 running a variety of windows versions from WinXP to Windows 7 (all running Outlook 2003 - 2010 or Thunderbird with IMAP connection).

Edit: Something else that may be worth noting is that store.exe uses half of the servers memory. I have 16GB in the server and store.exe takes about 9GB with the server running on 90-95% memory usage mostly.

My questions are:

  1. How do I go about debugging this situation? - This is my main query.
  2. Would moving to a separate server / exchange instance help (if this can be answered without knowing the cause of the problem)?
  3. The mailbox is about 20GB, is this too large?
Improve this question
6
  • What is your storage configuration? Is disk write-caching enabled?
    – Greg Askew
    Jan 24, 2013 at 16:47
  • When you say storage configuration, do you mean where is the data stored? The Mailbox DB is located on the same HD exchange is installed on. Write caching is enabled, yes.
    – webnoob
    Jan 24, 2013 at 16:57
  • 1
    You're shooting yourself in the foot with disk write caching enabled.
    – Greg Askew
    Jan 24, 2013 at 17:21
  • I didn't even think of looking at it until you mentioned it in your comment. I will be disabling it tomorrow morning. Could this explain the issues even without any power outage?
    – webnoob
    Jan 24, 2013 at 18:38
  • 1
    Disk write caching is specifically identified by Microsoft as not supported. It also known for trashing Exchange databases. Combined with inexpensive consumer-grade hard drives, that's a recipe for an outage. You may also want to check the drive smart statistics.
    – Greg Askew
    Jan 24, 2013 at 18:59

2 Answers 2

Reset to default
1

In line with the comments:

It turns out that write caching was enabled on the hard drives. Since turning this off the problems have gone away.

Improve this answer
1
  • Although your problems have gone away, caching is supported if you look at this page technet.microsoft.com/en-gb/library/ee832792(v=exchg.141).aspx - Caveat, use a UPS. Why wouldn't disk caching be supported? It's completely transparent to the EDB. Also Exchange is chewing your RAM up for the exact same reason; caching. Caching increases performance as you don't have to go to your spindle(s) for the data, it's quickly available via RAM chips, or RAID controller memory chips etc. I dont think caching is the root cause.
    – Snellgrove
    Apr 8, 2013 at 19:55
0

There can be many reason for frequent database corruption:

  1. transaction logs file size: You can enable circular logging to reduce the transaction logs.
  2. Exchange Server storage configuration options : Microsoft's recommendation is: Best practice: • 200 gigabytes (GB) or less. https://docs.microsoft.com/en-us/exchange/plan-and-deploy/deployment-ref/storage-configuration?view=
  3. Update Windows with the latest patches.
  4. Update Exchange to the supported CU.
  5. Also analyze application event viewer

You can check more Reasons and Solutions of Exchange Database Corruption from here.

For repairing Exchange database, Don't try it your own. Call Microsoft support!

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .