none
Using eseutil in a DAG - Which databases do you scan?

    Question

  • I have a simple 2 node DAG set up, with 5 databases being replicated. To check database integrity using eseutil /g I know the database must be dismounted. When I unmount the database only the active shows as "unmounted" and the passive still simply shows as "healthy". So, which database file do I actually scan with eseutil? The active, passive, or both? 

    Also, can anyone help me decipher the results of eseutil /g? If it detects any errors will it inform me on the screen with the progress bar? Or do I need to parse the entire INTEG.RAW file looking for something other than "error 0 (0x0)"?

    Wednesday, May 16, 2012 11:27 AM

Answers

  • That;s used against a DB, so if it;s active or passive copy it should be tested against both if you want to use the /g.  They both are DB, independent of each other.


    Sukh

    • Marked as answer by zoltank Thursday, May 17, 2012 6:50 PM
    Wednesday, May 16, 2012 12:47 PM
  • None of the databases are having an issue per se, and they're all currently mounted, healthy, and accessible. However, we had a 15 minute SAN outage (it spontaneously reset) last Thursday which affected both DAG nodes and required incident support from MS to fix the resultant issues. I just want to make sure the database are actually healthy, and there isn't any corruption which could cause serious issues down the road. 

    - So it sounds like I need to go back and scan the passive copy as well. I'm assuming it's safe to do so even though it's status still shows as healthy, so long as the active database is unmounted?

    - Regarding the ESEUTIL /G results, so I'll need to parse the entire INTEG.RAW file looking for an error other than 0? That's going to take a while...

    FYI, I've read through all those articles, and none of them seemed to really address the questions I have.


    well, I wouldnt bother actually. If there were lingering issues that would affect client access or store integrity, you would know it by now.

    If you are really worried, create a new database and move mailboxes to it. Thats time better well spent.

    • Marked as answer by zoltank Thursday, May 17, 2012 6:49 PM
    Wednesday, May 16, 2012 1:40 PM
    Moderator
  • I've already brought up the backup I created last night in our test environment, so it's pretty easy for me to run tests on the databases currently.

    On a tangent, can the database be damaged in a why which would allow it to work correctly for a while, and then cause serious problems down the road?

    That would be very unusual. Typically, damaged stores will throw alerts in the event logs and users will experience issues accessing mailboxes and items in the mailboxes rather quickly, and backups my fail as well or throw errrors.


    Wednesday, May 16, 2012 2:03 PM
    Moderator
  • For mailboxes it may happen, however if this is the case you can always create a repair request http://technet.microsoft.com/en-us/library/ff625221.aspx


    Sukh

    • Marked as answer by zoltank Thursday, May 17, 2012 6:49 PM
    Wednesday, May 16, 2012 2:09 PM

All replies

  • That;s used against a DB, so if it;s active or passive copy it should be tested against both if you want to use the /g.  They both are DB, independent of each other.


    Sukh

    • Marked as answer by zoltank Thursday, May 17, 2012 6:50 PM
    Wednesday, May 16, 2012 12:47 PM
  • You will run /g on the databsae which has issues
    Here is the thread with the same question
    http://social.technet.microsoft.com/Forums/eu/exchangesvradmin/thread/be607d42-b581-4fc8-a98c-982c90cca666 
    http://technet.microsoft.com/en-us/library/aa998737(v=exchg.65).aspx

    Gulab Prasad,
    MCITP: Exchange Server 2010 | MCITP: Exchange Server 2007
    MCITP: Lync Server 2010 | MCITP: Windows Server 2008
    My Blog | Z-Hire Employee Provisioning App

    Wednesday, May 16, 2012 12:52 PM
  • Why the need to do this?

    Wednesday, May 16, 2012 12:53 PM
    Moderator
  • None of the databases are having an issue per se, and they're all currently mounted, healthy, and accessible. However, we had a 15 minute SAN outage (it spontaneously reset) last Thursday which affected both DAG nodes and required incident support from MS to fix the resultant issues. I just want to make sure the database are actually healthy, and there isn't any corruption which could cause serious issues down the road. 

    - So it sounds like I need to go back and scan the passive copy as well. I'm assuming it's safe to do so even though it's status still shows as healthy, so long as the active database is unmounted?

    - Regarding the ESEUTIL /G results, so I'll need to parse the entire INTEG.RAW file looking for an error other than 0? That's going to take a while...

    FYI, I've read through all those articles, and none of them seemed to really address the questions I have.


    • Edited by zoltank Wednesday, May 16, 2012 1:35 PM
    Wednesday, May 16, 2012 1:30 PM
  • None of the databases are having an issue per se, and they're all currently mounted, healthy, and accessible. However, we had a 15 minute SAN outage (it spontaneously reset) last Thursday which affected both DAG nodes and required incident support from MS to fix the resultant issues. I just want to make sure the database are actually healthy, and there isn't any corruption which could cause serious issues down the road. 

    - So it sounds like I need to go back and scan the passive copy as well. I'm assuming it's safe to do so even though it's status still shows as healthy, so long as the active database is unmounted?

    - Regarding the ESEUTIL /G results, so I'll need to parse the entire INTEG.RAW file looking for an error other than 0? That's going to take a while...

    FYI, I've read through all those articles, and none of them seemed to really address the questions I have.


    well, I wouldnt bother actually. If there were lingering issues that would affect client access or store integrity, you would know it by now.

    If you are really worried, create a new database and move mailboxes to it. Thats time better well spent.

    • Marked as answer by zoltank Thursday, May 17, 2012 6:49 PM
    Wednesday, May 16, 2012 1:40 PM
    Moderator
  • I've already brought up the backup I created last night in our test environment, so it's pretty easy for me to run tests on the databases currently.

    On a tangent, can the database be damaged in a why which would allow it to work correctly for a while, and then cause serious problems down the road?

    Wednesday, May 16, 2012 1:44 PM
  • I've already brought up the backup I created last night in our test environment, so it's pretty easy for me to run tests on the databases currently.

    On a tangent, can the database be damaged in a why which would allow it to work correctly for a while, and then cause serious problems down the road?

    That would be very unusual. Typically, damaged stores will throw alerts in the event logs and users will experience issues accessing mailboxes and items in the mailboxes rather quickly, and backups my fail as well or throw errrors.


    Wednesday, May 16, 2012 2:03 PM
    Moderator
  • For mailboxes it may happen, however if this is the case you can always create a repair request http://technet.microsoft.com/en-us/library/ff625221.aspx


    Sukh

    • Marked as answer by zoltank Thursday, May 17, 2012 6:49 PM
    Wednesday, May 16, 2012 2:09 PM