topic Re: P4300 Multi Site Cluster reconfig to normal Cluster in StoreVirtual Storage

P4300 Multi Site Cluster reconfig to normal Cluster

Martin Lichthardt — Fri, 12 Jul 2013 16:05:01 GMT

Hi, I have the Problem that my Multisite Cluster Running Full and we now want to do with the waiver Redundancy again a normal cluster.
The load factor for time to over 90%, the data should be retained.
I have read that one should remove the individual P4000 from the site and thus the multi-site cluster is a normal cluster Migrates.
Is there a description of that?

Thangs

Martin

Re: P4300 Multi Site Cluster reconfig to normal Cluster

Emilo — Sat, 13 Jul 2013 17:09:13 GMT

All you have to do is go into the management group settings and change from muliti-site to single site.

Make sure your VIP is reachable from all the nodes.

Just to make sure are you going to break up the cluster?

Or exactly what is it that you are trying to accomplish?

Re: P4300 Multi Site Cluster reconfig to normal Cluster

Martin Lichthardt — Wed, 24 Jul 2013 09:00:26 GMT

I can do that on the fly??

there are some ESX volume and file server data LUNs in the cluster

best regards

Martin

Re: P4300 Multi Site Cluster reconfig to normal Cluster

HPstorageTom — Wed, 24 Jul 2013 10:36:18 GMT

Martin, changing the cluster back to a standard cluster doesn't safe any capacity.

You would need to change the Network RAID level of volumes in order to safe capacity. But changing the NRAID level while your StoreVirtual nodes are spread across 2 sites isn't really something that is recommended.

And you need to be aware what changing the NRAID level from 10 to a different level really means:

NR 10 -> NR0: no access to the data if one node fails; i.e. no online firmware upgrade capability.

NR 10 -> NR5/6: no multi-site support, can only be used for volumes that mainly have read access, and only safes capacity if the volumes are thin provisioned.

This brings me to question: are your volumes thin or full provisioned?

Re: P4300 Multi Site Cluster reconfig to normal Cluster

Martin Lichthardt — Fri, 26 Jul 2013 14:05:40 GMT

the volume are fully provisioned.
if I'm building a NRAID 5 volume in this multi-site cluster requires exactly as much memory as a NRAID 1

So I have to return the cluster to a cluster Singel site to build.

the question is more towards someone has done this in operation??

how to do it I have already read. Unfortunately, there was nothing going on in full operation

regards

Martin

Re: P4300 Multi Site Cluster reconfig to normal Cluster

oikjn — Fri, 26 Jul 2013 21:09:43 GMT

I can't say I totally understood what you were saying there. Not to be rude, but is english your 2nd language?

If you have NR10 volumes and they are all full provisioned then you should DEFINITELY convert them to thin right away and you will get your space back. This is as simple as right clicking on a LUN in CMC and selecting edit volume. In the adbvanced tab you will see "provisioniong" select thin and then click ok. this will get you any unwritten space back on your san. One caution there is I think this can be a write intensive action so I would either turn down your management group bandwidth so it doesn't affect production or do this during off hours.

Re: P4300 Multi Site Cluster reconfig to normal Cluster

Jitun — Mon, 29 Jul 2013 13:07:30 GMT

It is recommended to have Volume as Thin Provisioned when one is taks Snapshots for the Volume.
With Network RAID 5, there are hidden Snapshots taken which gets converted to Network RAID 5, but the top layer which is 10GB or 10% the size of the volume is always at Network RAID 10.

So if you plan to create a Network RAID 5 Volume, please only use Thin Provisioning.

One creates a Multisite assuming one has some kind of physical or logical separation of the Sites.
Space Utilization has nothing to do with Site configuration.