Highlighted
RobClendenin
1 Nickel

RecoverPoint/Oracle ASM Question

We are just about ready to go live with RecoverPoint and have been testing a bunch of different scenarios and we are working through the process for expanding an ASM diskgroup.  I understand that online volume expansions are not supported, that we would have to either remove the pair as a Replication Set or remove the CG completely in order to expand the volumes, or we can add a new volumes at the primary and remote sites then create the Replication Set in RecoverPoint.  I am trying to figure out which would be better/more efficient with the smallest unprotected window, or find any best practices.  My current thought, as our configuration is XtremIO+Vplex on the primary side and XtremIO on the DR side, is to just oversize the volumes to reduce the chances of needing to increase the ASM diskgroup size.  Any suggestions or recommendations will be appreciated.

Labels (1)
0 Kudos
5 Replies
forshr
3 Zinc

Re: RecoverPoint/Oracle ASM Question

Hi Rob,

Online volume/LUN expansion is not supported with the VPLEX splitter or XtremIO in conjunction with RP as a source or replica array, However, from RP version 4.4 adding a new volume/LUN to a CG (adding a replication set) does not cause loss of journal history.

Regards,

Rich

RobClendenin
1 Nickel

Re: RecoverPoint/Oracle ASM Question

That is what I thought.  I am preparing documentation for the team on all of this, so to make sure I am clear these would be the steps to expand an ASM disk group via adding a volume:

1. Create a RecoverPoint bookmark as a "previous configuration rollback point"

2. Create new volumes in primary and secondary/DR sites - adding them to the proper storage consistency groups

3. Create a new Replication Set for the RecoverPoint CG that was made up of the newly created volumes.

This would allow us to save the journal history, and once the ASM rebalance was completed on the primary server and delta's replicated, we would be back in a fully protected state.  Until that time we would still have the ability to roll back to a point up to the bookmark with a potential host side modification.

Basically, the reason I am asking, is that we will use our DR site as a UAT/DEV location as well.  Due to the effectiveness of XIO snapshots we are making snapshots of the target volumes to use for UAT and DEV instances, adding volumes to the ASM disk group affects any refreshes of those volumes (not pertinent for this forum obviously), so I just want to be sure I have the process nailed down to minimize any chances of breaking UAT/DEV copies.

0 Kudos
forshr
3 Zinc

Re: RecoverPoint/Oracle ASM Question

The procedure is fine. There may be one exception to this. Are you using any Distributed CGs?

0 Kudos
Quinc
1 Nickel

Re: RecoverPoint/Oracle ASM Question

How did you get RP to work with ASM?  I have been working with multiple EMC support techs and they cannot figure out why RP is removing my VOL disks and not replicating the disks from production.

0 Kudos
forshr
3 Zinc

Re: RecoverPoint/Oracle ASM Question

It's important to state that RP cannot remove any disk as it simply replicates source volumes to replica volumes as part of the rsets  added to the CG. I would move to image access mode and then compare the v$asm_disk_view as each side.

Regards,

Rich Forshaw

RecoverPoint Corporate Systems Enngineering

0 Kudos