IBM Support

IT17379: JOURNAL BACKUP FOR A WINDOWS CLUSTERED MOUNT POINT APPEARS TO BE VERY SLOW OR HUNG.

Subscribe

You can track all active APARs for this component.

APAR status

  • Closed as program error.

Error description

  • A journal backup for a Windows mount point in a clustered
    environment may appears to hang.
    
    The backup output shows that the backup is processing journal
    entries like this below:
    
    Querying Journal for '\\clusname\z$\mnt\point'
    Processing 25,315 Journal entries for '\\clusname\z$\mnt\point'
    
     Then there is no more activity during hours.
     That may suggest a hang, but a "service" trace shows that the
    client is looking for each file using the UNC name of the
    cluster, that the file scan is long for each file and that it
    finally fails with the message "file not found".
    
    
    Tivoli Storage Manager Versions Affected:  V6 and V7 on Windows
      platform.
    
    Customer/L2 Diagnostics:
    
    A "service" trace shows that the client is trying to access the
    changed files using the UNC name of the cluster.
    That file scan is very long and it finally fails until another
    file scan is attempted
    
    Example:
    
    17:03:44.085 [005316] [5252] :
     ..\..\common\ba\incrdrv.cpp(3636): BaJournaled(): Backup fs
     differs from journal fs:
       fs being backed up     = '\\clusname\z$\mnt\point'
       fs from journal entry  = '\\host-node\z$\mnt\point'
       vol being backed up    = 'z$\mnt\point'
       vol from journal entry = 'z$\mnt\point'
       hl being backed up     = ''
       hl from journal entry  = '\dirname\path'
       ll being backed up     = ''
       ll from journal entry  = '\filename.pdf'
       fsIsVMP                =  Yes
       SpecUsingVolAlias      =  No
    
    17:03:44.085 [005316] [5252] :
     ..\..\common\ba\incrdrv.cpp(3702): BaJournaled():
     '\\clusname\z$\mnt\point\dirname\path\filename.pdf' will be
     processed as a mountpoint dir:
       fs = \\clusname\z$\mnt\point
       hl = \dirname\path
       ll = \filename.pdf
    
    17:03:44.085 [005316] [5252] :
     ..\..\common\ba\incrdrv.cpp(3707): BaJournaled(): Received Jnl
     Query Response.
     Journal response
     \\clusname\z$\mnt\point\dirname\path\filename.pdf
    17:03:44.085 [005316] [5252] :
     ..\..\common\ba\incrdrv.cpp(3774): BaJournaled(): Checking for
     dir exclusion for path '\\clusname\z$\mnt\point\dirname\path'.
    17:03:44.085 [005316] [5252] :
     ..\..\common\fio\fileio.cpp(1193): fioCheckDirEclude() calls
     fioFsIsCaseSensitive()
    17:03:44.085 [005316] [5252] :
     ..\..\common\ba\incrdrv.cpp(3842): baJournaled(): calling
     fioCheckFileAccess:
      object     =
    '\\clusname\z$\mnt\point\dirname\path\filename.pdf'
      obj type   = dir
      entry type = backup
    
    17:03:44.085 [005316] [5252] :
     ..\..\common\fio\ntfileio.cpp(6279):
    fioCheckFileAccess(\\clusname\z$\mnt\point\dirname\path\filename
    .pdf): Entry.
    17:03:44.085 [005316] [5252] :
     ..\..\common\fio\ntfileio.cpp(6199):
    fioCheckObjAccess(\\clusname\z$\mnt\point\dirname\path\filename.
    pdf): Entry.
    17:03:44.085 [005316] [5252] :
     ..\..\common\fio\ntfileio.cpp(10338):
     NTObjAccess(\\clusname\z$\mnt\point\dirname\path\filename.pdf):
     Entry.
    17:03:44.085 [005316] [5252] :
     ..\..\common\fio\ntfileio.cpp(10341): NTObjAccess(): entry:
     file
     '\\clusname\z$\mnt\point\dirname\path\filename.pdf', requested
     access mode FIO_ACCESS_EXIST
    17:03:44.085 [005316] [5252] :
     ..\..\common\fio\ntfileio.cpp(6152):
    fioFileExists(\\clusname\z$\mnt\point\dirname\path\filename.pdf)
     : Entry.
     ...
    1 minute later ..
    ...
                    B/A Txn Producer thread
     2================>
    17:04:48.384 [005316] [5252] :
     ..\..\common\fio\ntfileio.cpp(10370): NTObjAccess(): Object
     '\\clusname\z$\mnt\point\dirname\path\filename.pdf' doesn't
     exist.
    17:04:48.384 [005316] [5252] :
     ..\..\common\fio\ntfileio.cpp(6212): fioCheckObjAccess():
     Access to Object
     '\\clusname\z$\mnt\point\dirname\path\filename.pdf'
     failed: Object doesn't exist .
    17:04:48.384 [005316] [5252] :
     ..\..\common\fio\ntfileio.cpp(6251):
    fioCheckObjAccess(\\clusname\z$\mnt\point\dirname\path\filename.
    pdf): returning 104.
    17:04:48.384 [005316] [5252] :
     ..\..\common\ba\incrdrv.cpp(3846): baJournaled(): fioAccess()
     returned 104 .
    17:04:48.384 [005316] [5252] :
     ..\..\common\ba\incrdrv.cpp(3854): BaJournaled():
    fioCheckFileAccess('\\clusname\z$\mnt\point\dirname\path\filenam
    e.pdf'): returns File not found.
    
     And it continues with next file from the journal.
    
    Initial Impact: Medium
    
    Additional Keywords:   TSM CLUSTERNODE
    

Local fix

  • If possible, use a drive letter instead of a mountpoint.  This
    problem does not exist when the volume has a drive letter
    assigned.
    

Problem summary

  • ****************************************************************
    * USERS AFFECTED:                                              *
    * Backup-archive client versions 6.4, 7.1 and 8.1.0 running on *
    * all Microsoft Windows platforms                              *
    ****************************************************************
    * PROBLEM DESCRIPTION:                                         *
    * See ERROR DESCRIPTION                                        *
    ****************************************************************
    * RECOMMENDATION:                                              *
    * This issue is projected to be fixed in the Backup-Archive    *
    * Client version 7.1.8 and 8.1.2 on all Microsoft Windows      *
    * platforms                                                    *
    * Note 1: This is subject to change at the discretion of IBM.  *
    ****************************************************************
    

Problem conclusion

  • So as the access via the UNC cluster name is disabled by
    Microsoft and cause the 'hang',
    the UNC cluster name access has been replaces with UNC machine
    name access.
    

Temporary fix

Comments

APAR Information

  • APAR number

    IT17379

  • Reported component name

    TSM CLIENT

  • Reported component ID

    5698ISMCL

  • Reported release

    71W

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2016-10-06

  • Closed date

    2017-03-22

  • Last modified date

    2017-06-13

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

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

Modules/Macros

  • TSMJBBD  DSMC
    

Fix information

  • Fixed component name

    TSM CLIENT

  • Fixed component ID

    5698ISMCL

Applicable component levels

  • R71W PSY

       UP



Document information

More support for: Tivoli Storage Manager

Software version: 71W

Reference #: IT17379

Modified date: 13 June 2017