nsrclone waiting / nsr server busy

[ukraine12]

Hi,

we have a tape robot with 8 drives(2 lto4's,3 lto3's and 3 lto 2 drives) and a vtl library. we make our backups on the vtl and when the savegroups are finished we clone our data from the vtl to real tapes. we extract a list of ssid's which we feed to the nsrclone command. Now we have the problem that when doing this cloning we very often get the message: NSR server busy, nsrclone waiting. Now this would make sense if all the drives were bus doing something(backups/recovers) but it is not. right now there is just 1 drive reading and 1 drive writing (1 vtl and 1 in the robot). legato version 7.3.4.

What is causing this???? thx in advance.

 

[605]

Are you sure that you have labeled your destination media to a CLONE pool ?

 

[ukraine12]

yes, i am sure. what we do is this: when our savegroup finishes we start a script in which we select the ssid's which were created and feed them to mminfo. so for every group that is ready a clone process is created. we have 4 clone pools (incr. and full for lto3 and lto2) so in the most optimistic view , we should have 4 drives reading data and 4 drives writing data. But now most of the time there is only 1 reading and 1 writing while the others are just saying NSR server busy....

 

[605]

Without knowing further configuration nobody can really help you. However there are general issues worth to check:
- Did you select devices for pools?
I am pretty sure because you use at least different tape drives (with different media?) - and of course there are incompatible cobinations
- Does the system wait for media? - Is there no media NW can recycle automatically?
- Do you use Auto Media Management at all?
- Do you use the options "Recycle from/to other pools"?
- Do you use Recover/Clone storage lists at all?

Looking at your daemon.log files should at least show you more details.

 

[a060463xyz]

605 is right, you have to check your /nsr/logs/daemon.log or daemon.raw file.
Have a look on your Server paralelism. Older Nwetworker installations had a default of only four paralelle datastreams.You should set your Server paralelism to 4 times the number of devices with following limits :
Maximum 12 with workgroup or business Edition.
32 with Networkedition with additional 32 per storagenode license.
64 with Power Edition with additional 32 per storagenode license .