IC78987: RESTORE OPERATIONS MIGHT FAIL IF THE OPERATION INCLUDES A LONG VSS FULL RESTORE AND ONE OR MORE VSS INCREMENTAL RESTORES

APAR status

• Closed as program error.

Error description

• Tivoli Storage Manage for Mail: Data Protection for Exchange
  database restore and mailbox restore might fail when trying to
  restore an incremental VSS backup.
  
  The following tdpexc.log messages are representative of the
  problem:
  
  09/20/2011 15:02:40 Backup(s) to be restored:
  09/20/2011 15:02:40 <Logs : VSS : full : 09/19/2011 21:29:42>
  09/20/2011 15:02:40 <File : VSS : full : 09/19/2011 21:29:42>
  09/20/2011 15:02:48 Preparing for restore of 'MAILDB02' from TSM
  backup.
  09/20/2011 16:17:27 VSS Restore operation completed with rc = 0
  09/20/2011 16:17:27    Files Examined   : 206
  09/20/2011 16:17:27    Files Completed  : 206
  09/20/2011 16:17:27    Files Failed     : 0
  09/20/2011 16:17:27    Total Bytes      : 47335522146
  09/20/2011 16:17:27 Backup(s) to be restored:
  09/20/2011 16:17:27 <Logs : VSS : incr : 09/20/2011 01:59:10>
  09/20/2011 16:17:27 FMV1017E (RC-50)  Session rejected: TCP/IP
  connection failure
  09/20/2011 16:17:27 FMV1017E (RC-50)  Session rejected: TCP/IP
  connection failure
  09/20/2011 16:17:31 ACN5060E A Tivoli Storage Manager API error
  has occurred.
  
  The signature for this problem includes these elements:
  
  - A (relatively) long time between the "Preparing for
  restore..." and "VSS Restore operation completed..." messages.
  In this example, the duration is longer than one hour, but can
  be longer or shorter, depending on the environment.
  
  - The FMV1017E message following the "<Logs: VSS : incr ..."
  message.
  
  The problem is known to occur when all of these conditions are
  true:
  
  1. A configuration setting exists in the computing environment
  that causes idle TCP/IP sessions to close after some period of
  inactivity. Examples include, but are not limited to, the TSM
  server IDLETIMEOUT setting and firewalls that close inactive
  ports after some period of inactivity.
  
  2. The incremental VSS backup requires a full VSS backup to be
  restored first
  
  3. The elapsed time for the full restore exceeds the threshold
  for closing inactive sessions (described in item 1 above).
  

Local fix

• Until a fix is available circumventions for this problem
  include:
  
  * Check the TSM server activity log for any ANR0482W messages
  that occur during the full restore operation. If the TSM server
  has timed out a session for the node being restored, increase
  IDLETIMEOUT.
  
  * In the case of an Exchange database restore, first restore the
  full VSS backup without the /RECOVER option. Next, restore the
  individual incremental backups. Note that this does not work for
  a mailbox restore.
  
  * Check for other elements in the environment, such as a
  firewall, that might close inactive sessions after some period
  of time, and if possible, increase the timeout threshold.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: Users of Tivoli Storage FlashCopy Manager    *
  *                 for Exchange running version 2.1 and later.  *
  ****************************************************************
  * PROBLEM DESCRIPTION: See ERROR DESCRIPTION                   *
  ****************************************************************
  * RECOMMENDATION: Apply fixing level when available. Until a   *
  *                 FlashCopy Manager for Exchange component     *
  *                 fix is available, FlashCopy Manager          *
  *                 customers may install and use a fixed        *
  *                 level of the Tivoli Storage Manager for      *
  *                 Mail: Data Protection for Exchange           *
  *                 component fix in their environment.          *
  *                                                              *
  *                 The Data Protection for Exchange component   *
  *                 provides equivalent functionality as         *
  *                 FlashCopy Manager for Exchange when when     *
  *                 used in FlashCopy Manager environments.      *
  *                                                              *
  *                 The following Data Protection for Exchange   *
  *                 levels are currently projected to include    *
  *                 this fix: 6.1.3.3 and 6.3.0.1.               *
  ****************************************************************
  *
  

Problem conclusion

• When performing a VSS full restore followed by a VSS
  incremental restore, the Data Protection thread may become
  disconnected from the TSM Server following the full restore
  for various reasons (such as firewall or network switch
  policy, or TSM server policy).  Tivoli Storage FlashCopy
  Manager for Exchange has been changed to automatically
  reconnect between restores to prevent this problem.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  TDP EXCHANGE WI

• Fixed component ID

  5698DPXAP

Applicable component levels

• R23W PSN

     UP

• R52W PSN

     UP

• R53W PSN

     UP

• R54W PSN

     UP

• R55W PSN

     UP

• R61W PSY

     UP