Solved!
Go to SolutionHH
Hamilton.Hanco
14 Posts
0
1060
December 2nd, 2016 13:00
NW9.0.1.4 - NDMP CIFS GOT HANG
Hi everyone,
We have a customer backing up CIFS from a VNX that only support 4 streams(the VNX only have 4 GB of memory).
The customer has 6 CIFS which are configured in a single group and single policy.
The policy GRP_FILESERVER starts at 5:00 PM
So, obviously only 4 CIFS starts its backup and 2 CIFS (random) stays on queue.
But the next day, even when the first 4 CIFS backups have ended... the 2 pending CIFS got HANG and never start.
So, the customer has to cancel the job and run these 2 CIFS manually
I already set the timeout to 0 but still doesnt work.
Am I missing something?? why the 2 pending CIFS never start?
Thanks!
719
1
Posted December 2nd, 2016 13:00
If you can run only 4 sessions, then limit your client parallelism to 4.
719
0
Posted December 2nd, 2016 13:00
But, shouldn't the 2 pending CIFS be able to start as soon as the 1 or 4 sessions are available?
I'm not looking to get the 6 CIFS starts at the same time. I want to know why 2 CIFS that are on queue never start?
thanks
720
0
Posted December 13th, 2016 06:00
Thank you very much
That's the explanation I expected.
the backups are fine now
ble1
6 Operator
•
14354 Posts
•
56186 Points
719
0
Posted December 3rd, 2016 05:00
I can't say what happens from filer point, but from NW point if you didn't restrict sessions, it will try to run them more. If such case, on filer side they might be placed asleep (2 of them) and you might have filer having issues waking up those. To rule that out, simply restrict this on initiator side (NetWorker) to 4 sessions (or even 3; I suspect they will share bandwidth so in overall run for whole group you won't see much difference).