IT42178: DSMC CRASH WITH MESSAGE LIKE "*** ERROR IN DSMC :.." ON MULTI-TARGET REPLICATION NODE

APAR status

• Closed as program error.

Error description

• On a multi-target replication node, the Backup/Archive Command
  line client (dsmc) can  crash or hang with no error in
  dsmerror.log, and with the following type of output seen in the
  terminal/console:
  
   IBM Spectrum Protect
   Command Line Backup-Archive Client Interface
     Client Version 8, Release 1, Level 14.0
     Client date/time: 24-05-2022 10:58:16
   (c) Copyright by IBM Corporation and other(s) 1990, 2022. All
  Rights Reserved.
  
  
   Node Name: NODEXX
   *** Error in `dsmc': double free or corruption (out):
  0x0000000001604ce0 ***
   ======= Backtrace: =========
   /lib64/libc.so.6(+0x81329)[0x7fb393e75329]
   dsmc[0x6e2363]
   dsmc[0x6e1c7f]
   dsmc[0x6dc941]
   dsmc[0x6705dc]
   dsmc[0x668faf]
   dsmc[0x669733]
   dsmc[0x652eb2]
   dsmc[0x654220]
   dsmc[0x65625d]
   dsmc[0x46c90f]
   dsmc[0x43be1a]
   dsmc[0x43dc26]
   dsmc[0x43ecdc]
   dsmc(__gxx_personality_v0+0x332)[0x43b3aa]
   /lib64/libc.so.6(__libc_start_main+0xf5)[0x7fb393e16555]
   dsmc(__gxx_personality_v0+0x281)[0x43b2f9]
   ======= Memory map: ========
   00400000-0112c000 r-xp 00000000 fd:04 4239410
           /opt/tivoli/tsm/client/ba/bin/dsmc
   0122c000-01280000 rw-p 00d2c000 fd:04 4239410
           /opt/tivoli/tsm/client/ba/bin/dsmc
   01280000-01374000 rw-p 00000000 00:00 0
   013a3000-017da000 rw-p 00000000 00:00 0
           [heap]
   7fb384000000-7fb384021000 rw-p 00000000 00:00 0
   7fb384021000-7fb388000000 ---p 00000000 00:00 0
   7fb389e45000-7fb389ea3000 r-xp 00000000 fd:00 6677520
           /usr/local/ibm/gsk8_64/lib64/libgsk8valn_64.so
   ....  Removed lines for better reading
  
   7fb393653000-7fb393752000 ---p 003d7000 fd:04 6291689
           /opt/tivoli/tsm/client/api/bin64/libtsmxerces-c.so.28.0
   7fb393752000-7fb393796000 rw-p 003d6000 fd:04 6291689
           /opt/tivoli/tsm/client/api/bin64/libtsmxerces-c.so.28.0
    7fb393796000-7fb393797000 rw-p 00000000 00:00 0
   7fb393797000-7fb393805000 r-xp 00000000 fd:04 6291690
           /opt/tivoli/tsm/client/api/bin64/libtsmxerces-depdom.s
  o.28.0Aborted
   root@nodexx:/opt/tivoli/tsm/client/ba/bin#
  
  The error text at the console may be different.
  Another example which has been seen is:
  
  *** Error in `dsmc': corrupted size vs. prev_size:
  0x0000000002405010 ***
  
  If a core dump is created, the stack trace of the dump may also
  vary from  one occurrence to the other.
  The symptoms are quite similar to IT41452, but it is not the
  same root cause.
  
  IBM Spectrum Protect Versions Affected:  Client versions
    8.1.13.0  to 8.1.15.1 on all supported platforms
  
  Additional Keywords: TS009375536 crash hang multi replication
  

Local fix

• Do not use multi-target replication until a fix is available.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * Backup-archive client and API versions 8.1.13.0 - 8.1.15.2   *
  * on all supported platforms                                   *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * Support for multi-target automatic replication failover,     *
  * added as of version 8.1.13.0, inadvertently introduced this  *
  * defect.                                                      *
  ****************************************************************
  * RECOMMENDATION:                                              *
  * Apply fixing level when available. This problem is currently *
  * projected to be fixed in level 8.1.17. Note that until the   *
  * fixing level is available, this information is subject to    *
  * change at the discretion of IBM.                             *
  ****************************************************************
  

Problem conclusion

• The code is corrected. The crash no longer occurs.
  

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:

Modules/Macros

• dsmagent dsmc     dsmcsvc  tsmapi64 libapits
  

Fix information

• Fixed component name

  TSM CLIENT

• Fixed component ID

  5698ISMCL

Applicable component levels