locked
ESE 531 - Database Corruption on Exchange 2010 SP1 + RU1 RRS feed

  • Question

  • We've had a significant latency issue with our storage over the past few nights during backups and are now getting the event pasted below for some of our databases.  The servers are DAG members, each database has roughly 1200 mailboxes totalling about 400G per database. The warnings are being generated slower as time goes on, so I suspect it's nothing to worry about.  I'd prefer some confirmation of that suspicion as there's no documentation (official or anecdotal) that I can find on it...  If the warnings continue, I'm tempted to activate another copy and reseed the one generating warnings after wiping the disk.  Any suggestions or information is much appreciated.

     Thanks in advance!

    Andy

    --------------------------------------------------------------------

    Log Name:      Application
    Source:        ESE
    Date:          11/17/2010 3:20:08 PM
    Event ID:      531
    Task Category: Database Corruption
    Level:         Warning
    Keywords:      Classic
    User:          N/A
    Computer:      Server02.suddenlink.cequel3.com
    Description:
    Information Store (4292) The database engine attempted a clean write operation on page 1760936 of database G:\Mbx02-01\Mbx02-01.edb. This acton was performed in an attempt to correct a previous problem reading from the page.

    For more information, click http://www.microsoft.com/contentredirect.asp.
    Event Xml:
    <Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
      <System>
        <Provider Name="ESE" />
        <EventID Qualifiers="0">531</EventID>
        <Level>3</Level>
        <Task>12</Task>
        <Keywords>0x80000000000000</Keywords>
        <TimeCreated SystemTime="2010-11-17T21:20:08.000000000Z" />
        <EventRecordID>94732</EventRecordID>
        <Channel>Application</Channel>
        <Computer>Server02.domain.local</Computer>
        <Security />
      </System>
      <EventData>
        <Data>Information Store</Data>
        <Data>4292</Data>
        <Data>
        </Data>
        <Data>1760936</Data>
        <Data>G:\Mbx02-01\Mbx02-01.edb</Data>
      </EventData>
    </Event>

    Wednesday, November 17, 2010 9:38 PM

Answers

  • Well, obviously that needs to get fixed, so good luck with the storage vendor.
    • Marked as answer by Gen Lin Friday, November 26, 2010 2:32 AM
    Saturday, November 20, 2010 1:07 PM

All replies

  • Have you been able to solve the storage issue?
    Thursday, November 18, 2010 2:42 AM
  • The error is from the Write operations. Any errors in the system log reagdring storage, adapter etc.
    Thursday, November 18, 2010 5:54 AM
  • Unfortunately, no.  I've rescheduled backups and maintenance windows so the load is reduced, but something in the storage stack cannot keep up even with ESEUTIL prior to taking the backup.  I've got our storage vendor poring over logs, but they haven't found anything yet.
    Thursday, November 18, 2010 9:33 PM
  • The system log is clean.  There are other warnings from ESE (510 and 533) and errors from ExchangeStoreDB (118) in the Application Log indicating the storage is taking a long time to respond to writes.
    Thursday, November 18, 2010 9:41 PM
  • The events point to corruption and is from storage, the storage vendor should provide more info. Also Caching etc on staorage controllers is an issue. please check exbpa
    Saturday, November 20, 2010 8:49 AM
  • Well, obviously that needs to get fixed, so good luck with the storage vendor.
    • Marked as answer by Gen Lin Friday, November 26, 2010 2:32 AM
    Saturday, November 20, 2010 1:07 PM
  • I am facing the same problem exactly on few DAG Databases, checked the system log n disk IOP's counters, everything seems ok. Our storage Admin checked from the storage side and everything is OK.

    Any thoughts on this?

    Many thanks

    David

    Tuesday, March 27, 2012 8:37 AM
  • It's a bit too late for you, but the next time you plan on buying hardware to run Exchange server, it help to download and run the corresponding version for Exchange Load Generator for some time to see if the hardwares can run up to your expectation. (Note: Don't do this on production environment, or anywhere with network connection to production)
    Tuesday, November 6, 2012 3:22 AM