Netbackup 7.6 VMware backup fails with error 6

[BTLOMS]

Netbackup VMware Snapshot Backup fails with Error 6.

Other error codes in the logs are 90, 23

BPBKAR logs show this:

00:47:56.146 [4956.10512] <16> dtcp_read: TCP - failure: recv socket (776) (TCP 10053: Software caused connection abort)
00:47:57.160 [4956.10512] <16> dtcp_read: TCP - failure: recv socket (776) (TCP 10053: Software caused connection abort)
00:47:58.174 [4956.10512] <16> dtcp_read: TCP - failure: recv socket (776) (TCP 10053: Software caused connection abort)
00:47:59.188 [4956.10512] <16> dtcp_read: TCP - failure: recv socket (776) (TCP 10053: Software caused connection abort)

01:21:32.779 [968.8012] <2> tar_base::V_vTarMsgW: INF - EXIT STATUS 6: the backup failed to back up the requested files
01:21:32.779 [968.8012] <4> tar_backup_tfi::backup_done_state: INF - Not waiting for server status
01:21:32.795 [968.8012] <4> dos_backup::fscp_fini: INF - backup status:<6>
01:21:32.795 [968.8012] <4> dos_backup::fscp_fini: INF - backup failed, deleting current track log
01:21:32.795 [968.8012] <16> dos_backup::fscp_fini: ERR - ct_fini() failed, error (15)
01:21:32.795 [968.8012] <16> dos_backup::V_PostCleanup: ERR - fscp_fini() failed, error : (15)

The job details ends like this:

01/27/2015 01:16:21 - Error bpbrm (pid=10400) from client XXXXXXX: ERR - Error opening the snapshot disks using given transport mode: Status 23
01/27/2015 01:16:22 - Info bpbkar32 (pid=9484) accelerator sent 0 bytes out of 0 bytes to server, optimization 0.0%
01/27/2015 01:16:23 - Info bpbkar32 (pid=9484) bpbkar waited 0 times for empty buffer, delayed 0 times.
01/27/2015 01:16:33 - Info bptm (pid=11104) waited for full buffer 0 times, delayed 0 times
01/27/2015 01:16:33 - Info bptm (pid=11104) EXITING with status 90 <----------

01/27/2015 01:16:43 - Info bpbkar32 (pid=0) done. status: 6: the backup failed to back up the requested files
01/27/2015 01:16:43 - end writing; write time: 0:00:10
the backup failed to back up the requested files  (6)

Browsing through the forums had many threads and solutions but none pointed in the right direction.

Eventually we found that a week ago on the VCenter server there was a network disconnect and few other errors that caused backups to fail. While all other servers recovered and finished backups, the bpbkar process for this server was hung and the process was sitting there on the media server for a whole week. The backups completed fine after we killed the hung process.