none
the logs are not truncated RRS feed

  • Question

  • Hi every supporter
    It is my first time to post here. Maybe it is not MS problem. 
    I set up new Exchange 2010 DAG with 5 servers. 3 with active DBs and 2 passive DBs.
    Our backup software is HP Data Protector.
    The backups are running through fine without error but the logs are not truncated. No error messages in Event log. 
    I find these events. 
    Index:    6496
    ID:       2046
    Level:    Information
    Provider: MSExchangeRepl
    Message:  The Microsoft Exchange Replication Service VSS writer (has successfully completed the backup of storage group 'MBX-DB01'.  
    Database log truncation has been requested for this storage group. Log truncation will occur on the active copy after the next log generation is created. Log truncation will occur automatically on the passive copy after that log file is copied.


    Index:    6498
    ID:       2046
    Level:    Information
    Provider: MSExchangeRepl
    Message:  The Microsoft Exchange Replication Service VSS writer has successfully completed the backup of storage group 'MBX-DB02'.  
    Database log truncation has been requested for this storage group. Log truncation will occur on the active copy after the next log generation is created. Log truncation will occur automatically on the passive copy after that log file is copied.


    Saturday, February 25, 2012 9:47 AM

Answers

  • Hi
       please check whether not more than 100 MB of log files are kept, which would be by design.Since Exchange 2010 SP1 has a static checkpoint depth, this means that up to 100 logs would be left over at any given time.  Although this logic worked in the majority of cases, it was determined that in some cases logs could be truncated off a source server that would be necessary to perform an incremental resynchronization of a passive copy.  In an attempt to correct this condition we introduced a new concept known a CACL (Checkpoint At Log Creation).  When a passive copy replays a log it reports an updated GTP (global truncation point) to the active copy.  The active copy then looks at the CACL.  If the passive copy is behind, the CACL will also be behind the checkpoint file on the active node.  Therefore, when we set the GTP on the active copy, we take the CACL value and add +100 to it (checkpoint depth), in theory that would mean that if all copies were up to date with advanced checkpoints we’d have at least 200 logs retained on the source.It is by design.

    Terence Yu

    TechNet Community Support

    • Marked as answer by ydy1c Friday, March 2, 2012 12:18 PM
    Monday, February 27, 2012 2:49 AM

All replies

  • Hey fellow,

    You need to check following things to progress towards resolution of this issue:

    1. Check VSS Writers status on all the Exchange Mailbox Servers by running command "vssadmin list writers" on command console with administrative privileges. Check if all the writers are stable. Most of the time this issue is related to VSS Writers are not in stable state. If any of the writer is not in stable state, try restarting Volume Shadow Copy service on all the servers

    2. Try to take backup using Windows Server Backup (VSS Full Backup). You can take help from post http://exchangeserverpro.com/exchange-2010-mailbox-database-backup-restore-windows-server-backup

    If backup goes successfull and logs are truncated you should also engage HP vendor support.

    3. Try to take backup from passive database copies from HP DP.

    4. Verify if Exchange Management Console (Powershell) is working perfectly on all the mailbox servers. Because HP DP uses Powershell API to truncate the logs once backup is done.

    Thanks.


    Junaid Ahmed

    Saturday, February 25, 2012 10:22 AM
  • VSS writer is fine because backup is completing. Yes, try with Windows Server Backup and check what happens.
    You always run backup on Passive copy of the database.

    Post the update..


    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

    Saturday, February 25, 2012 3:20 PM
  • Hi
       please check whether not more than 100 MB of log files are kept, which would be by design.Since Exchange 2010 SP1 has a static checkpoint depth, this means that up to 100 logs would be left over at any given time.  Although this logic worked in the majority of cases, it was determined that in some cases logs could be truncated off a source server that would be necessary to perform an incremental resynchronization of a passive copy.  In an attempt to correct this condition we introduced a new concept known a CACL (Checkpoint At Log Creation).  When a passive copy replays a log it reports an updated GTP (global truncation point) to the active copy.  The active copy then looks at the CACL.  If the passive copy is behind, the CACL will also be behind the checkpoint file on the active node.  Therefore, when we set the GTP on the active copy, we take the CACL value and add +100 to it (checkpoint depth), in theory that would mean that if all copies were up to date with advanced checkpoints we’d have at least 200 logs retained on the source.It is by design.

    Terence Yu

    TechNet Community Support

    • Marked as answer by ydy1c Friday, March 2, 2012 12:18 PM
    Monday, February 27, 2012 2:49 AM
  • Hi 
       I backup it again when logs are over 200 MB. It works now. 
    Friday, March 2, 2012 12:20 PM