Adding new node and additional LUN

If you are adding another content rounter, then just mount the new LUN up there and re-route.  the load will spread accross both the All-in-one and the new node.  This assumes you want to keep both systems as CR's.

You can extend the volume, IF you used LVM or VxVS to create the original volume.  Then you just add the LUN to the existing disk group and expand the file system.

If however you just used the LUN without a volume manager, then you are stuck.  You can migrate to a new volume created with a Volume manager, but you can't extend the current volume.   Unless you are comfortable with volume managers, I'd contact Symtantec support and have them help you through it (like I did the first time when I set up wrong.)

See Page 157 of the Puredisk 6.5 AdminGuide (Storage Pool Management) for detailed instructions on adding nodes. 

It all depends on the reason you decided to add a node, I guess.

Well you mentioned "unclustered" right - unless you add the new machine as a spare node, to make it clustered all-in-all one virtual server, there will be a risk of single pont of failure, always.

That being said, if you add the machine as a second content router, that by itself actually means that you are going to allocate some storage (another 2TB you mentioned, most likely) on it and share the load together with the existing all-in-all node.

So it will end up look like,

Node1 = SPA, MB, CR, 2TB storage

Node2 = CR, 2TB storage

Like the PureDisk Admin Guide pdf page 172-on shows, you are going to do "parallel" rerouting of the contents because you are adding one CR to one CR.

That means this.

1. You add the second content router with 2TB and activate it's content router service.

2. Now your SPA sees two content routers.

3. SPA will half the current load on the existing storage, dump it to the second node's storage.

4. Now both node's storages (LUNs) have approximately equal amount of load.

5. All backups thereafter will devide the storage on to the two nodes, equally (based on FP, yes, but usually it gets equally distributed)

Anyway, unless you add it to be a spare node and convert existing all-in-all into a clustered server, you will not need to worry about gluing the new 2TB LUN on to the existing 2TB LUN, I think.

Best practice, from my view, is yes adding it as second CR, to share the load, to lower network traffic, to avoid reconfiguring volumes.

Cheers,

Abe

ok, so i have setup a test environment in my lab and when i reroute the data on the CRs and activate the additional netbackup export engine, the exports fail. I updated my weekly export to be two seperate policies (the first policy having half of the clients and the second policy having the second half), the first policy has the first puredisk node being used and the second policy has the new puredisk node being used.

However, when I run the policy, it fails, expect for 2 or 3 clients. Why is this?

Is there another way to configure to work properly?

ok so i found the problem but i do not know how to resolve it.....both the export policies should be using the master media server. However, one of the policies will use the regular media server and not the master media server. How do I change it so that both export policies will use only the master media server????