After resolving the cause of a restore failure, the next restore does not go active.

Article:TECH33379  |  Created: 2004-01-22  |  Updated: 2012-08-15  |  Article URL http://www.symantec.com/docs/TECH33379
Article Type
Technical Solution



Issue



If a restore fails in an abnormal fashion, it may orphan the bprd mpx-main process that manages the media server processes.  This will prevent subsequent restore jobs requiring the same media from progressing.  How to clear the condition so that a restore will complete?


Error



If no other restores are occurring, this message in the bprd debug log confirms that a bprd mpx-main process has been orphaned.

17:42:45.515 [7756] <2> run_mpx_main_bprd: Another bprd is already mpx main

These messages will be logged by the orphaned bprd mpx-main when it receives a new restore request.  In this case the restore requested by PID 2707 failed after it started the mpx-main (PID 2782) and this affected the restore requests from PIDs 7640 and 26204.

17:37:57.181 [2782] <2> process_msgQ_msg: Got is mpx main up request from pid 2707
17:37:57.281 [2782] <2> process_msgQ_msg: Got new restore request from pid 2707
17:38:27.753 [2782] <2> process_msgQ_msg: Setting portnum = 0 for myclient_1334572283
17:42:45.656 [2782] <2> process_msgQ_msg: Got is mpx main up request from pid 7640
17:42:45.756 [2782] <2> process_msgQ_msg: Got new restore request from pid 7640
18:00:31.386 [2782] <2> process_msgQ_msg: Got new terminate request from pid 7640
18:00:57.292 [2782] <2> process_msgQ_msg: Got is mpx main up request from pid 26204
18:00:57.392 [2782] <2> process_msgQ_msg: Got new restore request from pid 26204
18:05:03.664 [2782] <2> process_msgQ_msg: Got new terminate request from pid 26204
 


Environment



NetBackup 6.x and 7.x


Solution



The bprd mpx-main (PID 2782 in this example) must be terminated to allow successful restores from the media that it is has reserved.
 
If the bprd main which started the orphaned mpx-main is still running, canceling the restore job will terminate both the bprd main and the bprd mpx-main.
 
If the bprd main is no longer running, either
  • stop all bprd processes on the master server, confirm that they have exited, and restart bprd.
  • identify the bprd mpx-main process and terminate the process manually.
  • in extreme cases a reboot of the server may be necessary.
Alternatively, some versions of NetBackup can temporarily bypass the problem by creating the following touch file:

  /usr/openv/netbackup/NON_MPX_RESTORE

This will allow the restore to skip the MPX processing and perform the restore serially. This workaround should be considered temporary, as it will affect the performance of conncurrent multiple stream restores from the same multiplexed media.
 


Legacy ID



269675


Article URL http://www.symantec.com/docs/TECH33379


Terms of use for this information are found in Legal Notices