DB2 backup using NMDA 1.2

Is it that you get this message when

a) you try to create NSR client with virtual name, or

b) when you try backup using virtual name

What you said seems to indicate a), but error seems to indicate b).  To make it more complicated, I believe SuSE cluster might not be supported which might require some manual work to make it work (eg. setting up lcmap file).  With lcmap file in place, NW will know which name (virtual name) is cluster and which paths belong to it.  With that in place, create client called in a way to reflect virtual host (do not forget to add physical nodes to remote access list) and off you go.

Hi,

I created the NSR client resource with the virtual name and the error message comes when the backup is fired. Also, I tried adding both the physical nodes hostname in the remote access list but still the backup fails with the below error message:

“72580:nsrdasv: The NSR Client resource for the < > could not be located.”

One more thing I wanted to ask his I have selected the backup schedule as Weekly Full & Daily Incremental….still the backup is happening FULL only daily even though showing the level as incr.

Kindly suggest on this at the earliest.

Regards,

Pratik

I think the cause here is that suse cluster is not officially supported (at least it was not in the past) so NW client is not cluster aware.  Long time ago, there was easy way out as you could just set lcmap file - you can see manual for lcmap for correct format. And for NW to check that file, you will need to instruct it that it is running in cluster (I believe on Linux you will need to touch /usr/sbin/NetWorker.clustersvr).  Once this is done, assuming your NSR client configuration is correct, it should work just fine.  I would suggest to test first client initiated backup before you move on to server initiated one.

Hi Hrvoje,

Do you mean to say that the NW is not supported on SUSE Cluster Environment ?

Also, can you plz tell me where is this lcmap file located or how can that be created ?

I'm quite sure it is supported, but not tested, but I could be wrong.  You should rely on official statements and those you can find in software compatibility guide.  lcmap is always in directory where NW binaries are (eg. /usr/sbin in your case I believe).

Yes...there is a file in /usr/sbin by the name lcmap.ncs.

But unfortunately i am not able to understand any of the parameters under it..:(

So not understanding what parameters should be set for the cluster to work..

Also, as per the NW compatibility guide, the nmda is supporting db2 active passive cluster environment.

Could you please help me out with what parameters should be set in lcmap.ncs ? Attaching the lcmap.ncs file also FYR.

You should create lcmap file and inner structure should be the same as outlined in manual (see man lcmap).  Unless you are sure in what you are doing at this point, I would suggest to engage official support as you don't wish to end up with unsupported or wrongly configured environment.

I have already logged a S2 severity call with EMC support but unfortunately none of them is updating on the same.

Regards,

Pratik Vengurlekar

holyboy1989 wrote: I have already logged a S2 severity call with EMC support but unfortunately none of them is updating on the same.

They are waiting for this thread to give solution

Well, if it is not on compatibility list (check clusters first), then you know this may not be supported, but it only mean that it is not tested 99% of the time.  I would say that quickly verifying this with lcmap file would do the trick - it' easy and inexpensive way of configuring things.  Usually, NW will use script based autodiscovery where it probe cluster software to see which cluster controlled volumes are there - this is helpful when you add new things to equation so you don't have to worry about those being manually added (think of save set All).  So, with manually set lcmap file you have to also manually maintain it.

Now, if we consider worst case scenario, that there is big and fat no from EMC for whatever reason, you can still do it in a what you did it before - via physical node.  And by setting NSR_CLIENT to whatever you want, you can rule out where index will be written (though this is not elegant and certainly not workable in certain conditions). 

Bottom line is - test it. If you can't do it, let support guy do webex and do it for you - it takes couple of minutes to set it up and do quick test on something small.  If you verify both backup and restore to work with lcmap as I suggested, do not forget to ask EMC qualifier - this will qualify solution at your site and make it officially supported by support (this is to avoid those 1st line support folks who check compatibility first and ask all sorts of logs on the system just to avoid real technical work).