Multiple streaming for One client

[RajkumarBaniwal]

Hi Experts, I am having a query, please provide solution. In Netbackup 7.5, one policy Windows is having 100 clients( without multiple streams enabled in policy). I want one client backup to run with multiple streams on daily basis(remaining client should run without multiple streams)Is there any way out for this( except for creating new policy)? Master server OS,Solaris 10

[Nicolai]

No - you have to create a new policy

[Marianne · Accepted solution]

I agree - create new policy.

Settings in Attributes tab will apply to all clients.

Or maybe there is a way....

Set Max Jobs Per Client in Host Properties -> Global Settings to 1.

In Host Properties -> Client Attributes -> Add/select the one client.
Change Maximum data streams to required number of simultaneous streams.

Select 'Allow multiple data streams' in policy and set MPX in schedules.

The one client will activate number of jobs set in Client Attributes, rest will activate 1 job at a time and queue the rest.

But this seems like a lot more work than just moving the 1 client to a new policy... 

[sdo]

You could do it, by setting the master server 'client attributes' for 99 clients to max_jobs of 1, and then the one client max_jobs=N, and then enable multiple streaming on the policy - but this will very likely result in many more jobs in the activity monitor - i.e. many more individual stream jobs - which, for 99 clients, would still only run one at a time (per client, of the 99 clients).  But you would still only have the one policy.

Personally, this is my preference.  I.e. I always configure any client to always have a 'master server client attributes' entry.  This way I get to fine tune/control job instance/concurrency, and also control the VSS attributes, and also fin tune/control the preference for client side de-dupe. 

It also means that if a client does accidentallyget added to a policy which does allow multi-streaming, then I still only ever get one stream per client.

However, one side effect of such a config, is that you get many more jobs reported by OpsCenter.  I dare say that the master server also has to do more work with job-scheduling and resource brokerage.  But, for me, another added benefit is that if one part/drive/file-system/mount-point of a client fails, then I still at least get the rest of the parts/drives/file-systems/mount-points.