Error code FFFEFFF7 The requested tsa is not loaded LEGATO tape backup

After a few days the tape backup starts failing, and we can see an error message in the Save.nlm's main thread:

 

"9cf5f3a0[SAVE.NLM's main thread   ]0001aab9= fffefff7"

 

 in the smdr debug log. After that, the backup doesn't make any progress. The  backup will just sit  there and never time  out. Reloading Networkr in conjunction with TSA/SMDR doesn't help. The Netware server is still working but we have to restart the cluster  node to get the backup to reconnect and start working again. The local volumes on the Netware server could still be backed, but not the cluster resources. This problem was intermittent. Sometimes several days would  go by before the problem  would happen. Coredump showed that there was nothing wrong with the SMS files  such as smdr.nlm and tsafs.nlm. They were not  hung or in an error condition. They were in fact waiting for work to do. There were no tsafs read threads running. It was as though a backup had never happened, however, this problem would happen right in the middle of the backup. The cluster resources had not failed over to another node but were running fine even after the  problem had occurred.