NDMP Backups are failing with a Warning : NDMP backup/restore streams reach the configured concurrent streams 4, but less than the maximum supported concurrent streams 8.
Problem
During an NDMP backup job operation of an EMC Celerra device, the following message is reported:
"Active NDMP backup/restore streams reach the maximum concurrent streams supported 8. No more NDMP sessions can be started. Wait current backup/restore sessions finish before start new ones."
The active job will immediately fail with Exit Status Codes 25 or 99.
When backing up a EMC Celerra device that has been configured with a vast number of filesystems that have been defined within one single policy, and the 'Allow Multiple Streams'
option is enabled, this will initiate a number of streams (one stream per filesystem)
Error Message
NetBackup Job detailed status and may fail with Exit Status Codes 25 or 99.
From Activity Monitor > Jobs > Detailed Status :
Exit Status Codes 25 Failure
25/08/2019 12:00:02 - begin writing
25/08/2019 12:00:02 - Error ndmpagent (pid=7020) filername: Active NDMP backup/restore streams reach the maximum concurrent streams supported 8. No more NDMP sessions can be started. Wait current backup/restore sessions finish before start new ones.
25/08/2019 12:00:03 - Error ndmpagent (pid=7020) ndmp_data_connect_v3 failed, status = -1 (-1)
25/08/2019 12:00:03 - Error ndmpagent (pid=7020) NDMP backup failed, path = /volume/dir/
25/08/2019 12:00:04 - Info bpbrm (pid=6636) child done, status 25
25/08/2019 12:00:04 - Info bpbrm (pid=6636) sending message to media manager: STOP BACKUP filername_1413000000
25/08/2019 12:00:05 - Info bpbrm (pid=6636) media manager for backup id filername_1413000000 exited with status 0: the requested operation was successfully completed
25/08/2019 12:00:05 - end writing; write time: 0:00:03
cannot connect on socket (25)
NDMP backup failure(99) Failure
25/08/2019 12:51:51 - Info bptm(pid=9316) start backup
25/08/2019 12:51:52 - Error ndmpagent(pid=8772) NDMP_LOG_ERROR 0 Active NDMP backup/restore streams reach the configured concurrent streams 4, but less than the maximum supported concurrent streams 8. No more backup/restore sessions can be started. Wait current backup/restore sessions finish, then increase concurrent DataStreams parameter.
25/08/2019 12:51:53 - begin writing
25/08/2019 12:51:53 - Error ndmpagent(pid=8772) ndmp_data_connect_v3 failed, status = -1 (-1)
25/08/2019 12:51:53 - Error ndmpagent(pid=8772) NDMP backup failed, path = /volume/dir/
25/08/2019 12:51:59 - end writing; write time: 00:00:06
NDMP backup failure(99)
Cause
This reported failures are caused by the concurrent streams limit of 4 that is defined by EMC Celerra Host.
And by default each a Data Mover (individual data movers map network and storage connectivity in the Celerra) which supports a total of 4 NDMP sessions
Solution
Adjust the "Limit Jobs Per Policy
" by setting the count to 4 from the 'Policy Attributes Tab'
By lowering the concurrency to 4 jobs should allow for "NDMP sessions on the Celerra " to become active, and the remaining jobs will queue until such time any inflight NDMP session are marked complete.
If an EMC Celerra device has a minimum of 8GB memory, then it is possible to adjust the NDMP sessions count to 8
Celerra command example:
# server_param server_xx –facility NDMP –modify concurrentDataStreams –value 8
Applies To:
EMC Celerra
NOTE: Please Consult the most current EMC documentation for more information.