locked
Soft recovery Failed with same checkpoint file for more than one database RRS feed

  • Question

  • Hello,

    We have a situation of running soft recovery to mount the databases from backup.

    Here are the details, Exchange 2007 with 4 storage groups and 3 edb files each. For example

    Storage Group1(3 database share single log file location)

    SG1(F:\SG1logs)

    --->DB1.edb(G:\DB1\DB1.edb)

    --->DB2.edb(H:\DB2\DB2.edb),

    --->DB3.edb(H:\DB3\DB3.edb)

    Performed soft recovery for DB1.edb with /s /l /d /i options and it was successful. We tried for DB2.edb with same options and soft recovery succeeds with

    Extensible Storage Engine Utilities for Microsoft(R) Exchange Server
    Version 08.03
    Copyright (C) Microsoft Corporation. All Rights Reserved.

    Initiating RECOVERY mode...
        Logfile base name: E01
                Log files: F:\SG1logs
             System files: F:\SG1logs
       Database Directory: H:\DB2

    Performing soft recovery...

    Operation completed successfully in 0.140 seconds.

    But still the database shows dirty shutdown. Same result for DB3.edb. We tried by taking a backup of *.chk file before starting soft recovery of DB1.edb and replaced the chk file after DB1.edb recovery. Now we performed soft recovery of DB2.edb which succeeds and DB2.edb shows "Clean shutdown". Why is this behaviour ? Do we need to take a backup of chk file before performing soft recovery ?

    Can anyone help in this regard ?

    Thanks

    Inmar



    • Edited by inmar_mine Monday, May 14, 2012 1:19 PM
    Monday, May 14, 2012 1:19 PM

Answers

  • Hi,

    Exchange begins scanning log files, beginning with the oldest file available, instead of starting at the checkpoint log.


    Xiu Zhang

    TechNet Community Support

    • Marked as answer by Xiu Zhang Monday, May 28, 2012 9:13 AM
    Tuesday, May 15, 2012 8:16 AM

All replies

  • At times our .chk point file might get corrupted.

    --

    One other Possible resolution

    --

    At this point we can create a new .chk file by having /s pointing to a new location.

    Even that should get the database to clean Shutdown.


    Satheshwaran Manoharan | Exchange 2003/2007/2010 | Blog:http://www.careexchange.in | Please mark it as an answer if it really helps you

    Monday, May 14, 2012 3:03 PM
  • Hi Satesh,

    Thanks for the reply.

    "At this point we can create a new .chk file by having /s pointing to a new location.

    Even that should get the database to clean Shutdown."

    In this case will the soft recovery process start replaying all the logs or only the logs not replayed? We don't want to replay all the logs as log count is very high.

    Thanks

    Inmar

    Tuesday, May 15, 2012 4:00 AM
  • Hi,

    Exchange begins scanning log files, beginning with the oldest file available, instead of starting at the checkpoint log.


    Xiu Zhang

    TechNet Community Support

    • Marked as answer by Xiu Zhang Monday, May 28, 2012 9:13 AM
    Tuesday, May 15, 2012 8:16 AM
  • Thanks for the reply.

    Tuesday, May 15, 2012 9:50 AM