Start a Conversation

Unsolved

This post is more than 5 years old

2621

September 24th, 2014 07:00

EMC RecoverPoint 4.0.2 and VMware SRM 5.5

I am hoping someone can help me with an issue. (I don't know if it is really an issue, but a step I am missing)

I have two data center sites with SRM and EMC RPAs/VNX/etc.  I finally got a successful failover and failback of protected virtual machines after weeks of troubleshooting network port issues.

Normally on the EMC RPAs, you can select an RPO time frame, say 15 minutes for your consistency group.  Under the group policy for this consistency group, you tell it that you want VMware SRM to manage you(RPA's) during a failover/failback event.  When you do this, it takes the ability away from the RPA to define the RPO.  Basically grays out the link policy so you can not make changes to it.  So I started looking at SRM to see how you can do this with no luck.

I know with VMware SRM, if you use vSphere Replication you can setup RTO/RPO for your replication.  However, because I am using ARRAY based replication, I do not get the option to tell SRM what my RTO/RPO is.  Is there a config file that I need to edit on SRM to have it define the RTO/RPO?  Since SRM is controling the failover and failback of the RPA functionality.

I hope this makes sense.  Any help will be appreciated.

1.1K Posts

September 24th, 2014 08:00

Hi there,

SRM cannot be configured to control the underlying RPO. Firstly, it's important to understand what the RPO setting actually does in RecoverPoint. When you set RPO to 15 minutes the purpose of the setting in conjunction with the Lag Optimizer group priroity and/or the CLI 'config_link_policy' command using the regulate_application setting is to ensure the lag does not surpass the configured requirement, i.e. 15 minutes. This does not mean that the snapshots are sent every 15 minutes. There is the new snap-based replication feature for VNX in 4.1 which has this type of capability.

When you initiate a recovery plan in SRM it will generate it's own bookmarked snapshot and this will be replicated and distributed to the journal/replica in conjunction with any Lag Optimizer setting. The VSI RecoverPoint Management plugin faciliates the capability to choose a PiT from the journal and tell SRM to use this.

I hope that helps.

Regards,

Rich

11 Posts

September 24th, 2014 11:00

dhattar,

To shed some more light on the RPO settings that Rich( forshr)  mentioned, you may find this presentation helpful.

We explain what exactly is meant by RPO and RTO with RecoverPoint and how those settings apply.

The presentation builds up the required definitions to understand the settings, so it is best to study it cover-to-cover using the slide notes, but you could try to zoom on these slides :

  • #9 to understand how much data you at risk of loosing (RPO)
  • #29-33 for Lag Optimization
  • #48 for the Cheat Sheet

Best,

Boaz

2 Posts

September 26th, 2014 12:00

Hi,

I believe I understand now after reading that document, but it still doesn't help me understand why the RPA loses the ability to define the RPO. 

Lets say we have an RTO of 24 hours to be backup at a DR site.  Well the RPO is what keeps the data synced right?  Like even though the RTO 24 hours and we have the RPO set to 5 hours, that means the system at both sites should not be more the 5 hours apart in regards to data synced?

My issue is that before you take the RPA that an external application will control you, you can select an RPO.  Once you tell the RPA that it is controlled by an external application, such as VMware SRM, the part in the GUI that allows you to put in a RPO is grayed out.  So when you have it setup this way, there is no way to define a RPO right? 

Does this make sense?

1.1K Posts

October 2nd, 2014 08:00

There's nothing stopping you from taking a CG copy out of SRM external management control and changing the policy, then returning it back to SRM external management control.

Also, note that in RecoverPoint the term RPO is used to refer to the required lag, as opposed to the actual lag, which is a system state and still referred to as lag.

6 Posts

May 27th, 2015 04:00

Once a CG is set for external management it can not be altered by RP, only monitored, however any link settings made prior to managed externally are retained. If you want to alter link settings, switch the CG to RP management, make your changes, then set the CG to external management again when you are done.

July 19th, 2016 04:00

RecoverPoint Appliance cluster (RPA Cluster) gives you the ability to define a CG and add Replication sets

It is while defining the CG that you can decide and define RPO/ RTO for the CG.

Now when you give the control to an external management tool (SRM in this case)

The ability to only failover and failback (via a Recovery Plan) is enabled by the SRM and "not" the ability to granularly go back to an image is provided to SRM.

Hence as and when needed, remove the CG control from SRM (bring the CG back to the RPA management) and change the RPO/RTO.

675 Posts

July 19th, 2016 05:00

One really important point to make; SRM, with any SRA, assumes that the underlying replication technology has only 1 PiT. As RP as an inherent Any-PiT capability, one can leverage the free VSI for RP management which is a vCenter plugin that enables to select Any point-in-time with SRM w/ RP SRA.

Hope that helps,

Idan Kentor

RecoverPoint Corporate Systems Engineering

No Events found!

Top