After upgrading to 7.5 through 18.104.22.168, multiple bpdbm processes appear to be using excessive CPU time. The number of bpdbm processes will build until it eventually interrupts the regular operations of NetBackup, triggering various application faults.
When this occurs, various processes may fault:
Wed Jun 20 2012 12:22:34 Application Error E1000 Faulting application HPWbemDump.exe, version 22.214.171.124, faulting module msvcr80.dll, version 8.0.50727.6195, fault address 0x0000000000006a68.
Wed Jun 20 2012 10:34:44 Application Error I1004 Reporting queued error: faulting application bpdbm.exe, version 7.500.312.605, faulting module msvcr100.dll, version 10.0.40219.1, fault address 0x00000000000761c9.
Following one of the hung bpdbm pids, the log will show the following:
11:25:48.970 [5240.1228] <16> change_expiration_date: (-) Translating EMM_ERROR_MediaAllocated(2001049) to 97 in the NetBackup context
11:25:48.970 [5240.1228] <16> change_expiration_date: emmlib_MediaDBOp() failed: 97
11:25:48.970 [5240.1228] <16> change_expiration_date: unable to change expiration date for media ID ABC123
11:25:48.970 [5240.1228] <2> DbmOdbcConnect::beginTransaction: (11353) transaction started
11:25:48.970 [5240.1228] <4> db_error_add_to_file: unable to change expiration date for media ID ABC123; EMM failure: requested media id is in use, cannot process request (97)
11:25:48.970 [5240.1228] <32> expdate_image: unable to change expiration date for media ID ABC123; EMM failure: requested media id is in use, cannot process request (97)
11:25:48.970 [5240.1228] <2> DbmOdbcView::~DbmOdbcView: (.5) Exiting
This issue is caused by an addition to function calls in the 7.5 release of NetBackup in regard to media change expiration calls.
The formal resolution for this issue (Etrack 2810912) is included in the following release:
- NetBackup 7.5 Maintenance Release 4 (126.96.36.199)
Information on NetBackup 188.8.131.52 is available in the Related Article linked below.
When running NetBackup 184.108.40.206, please access the Related Article linked below to obtain a hotfix containing a fix for this issue.
Disable OpsCenter polling for the master server experiencing the issue.