SRM 5.8 Failover
January 5, 2015A terrible thing has happened and it’s time to failover your datacenter to your disaster recovery site. Well, maybe you’re just migrating your datacenter to a new one, but this is always a bit of a tense situation. Luckily we’ve had the opportunity to test the failovers many, many times before so we can be confident in our process.
Go to the Recovery Plan and click the monitor tab. Click the “BIG RED BUTTON” (yeah, it’s not that big, but it has big consequences).
Before the failover actually happens, you’ll be given a warning and you actually have to click a check box stating that you understand the consequences of performing this operation. After that you’ll be given the opportunity to do a Planned Migration which will try to replicate the most recent changes and will stop if an error is encountered, or a Disaster Recovery migration which will just failover anything it can and as fast as it can. Pick your recovery type and click Next.
Review the process that is about to happen and click Finish.
While the recovery is running, you’ll be able to monitor the process on the recovery steps screen. Notice that this is slightly different from a test recovery in a few places, such as not creating a writable snapshot but rather making the existing storage writable in the new datacenter. Hopefully everything is working well for you after the failover.
Now it’s time to go back to our original datacenter. Click the “Re-Protect” button which looks like a shield with a lightning bolt on it. This Re-Protect will reverse the direction of the replication and setup a failover in the opposite direction. You can consider the DR site to be the protected site and the original production site to be the recovery site, until you fail back.
When you run the Re-Protect, you’ll need to once again confirm that you understand the ramifications of this operation.
Now that everything is reversed, you can run another failover, but this time a “Planned Migration” is probably more reasonable since you’re likely planning to do a failback and it’s not a second disaster, this time at your disaster recovery site. (That would be awful)
Review the failover and click Finish. When the failover is done, be sure to Re-Protect it again to get your disaster recovery site back in working order!
Summary
Failovers can be stressful but thankfully we’ve tested all of our plans before, so that should take some of the pressure off.
Hi, I’ve carried out a successful migration of VM using Abr. I ran the re-protect and that completed. The VMFS volume now has the additional characters in the name “snap-41698806-VMAX-PRD-01* on the recovery site. The re-protect has not cleaned that part. Should it have reverted to the original VMFS name (VMAX-PRD-01), or it it just something i’ll have to live with. Would manually changing it effect future recovery?
Thanks
Kevin
You should be able to rename the datastores without issue.
In the future, if you would like to prevent the snap-datastore names, go to the advanced settings of the site and look for the storageProvider.fixRecoveredDatastoreNames setting. Enable this and it should prevent those snap names from being an issue for you.
As always, test it out first to make sure it works for you.
Hi Eric, I make the recommended change and tested. It worked perfectly. Its nice just to keep the Datastore naming consistant after a migration or recovery. Your series on srm has been very helpfull. Many Thanks!
Great! It’s always nice to hear that the content is useful for someone. That’s why we blog.
Hello Eric,
First at all, Congratz for you Blog, it´s Really Really awesome!!!
I Have a stupid question,
I´m doing a lab of SRM 5.8 and you know if the main Site Fails it will automatic do a failover to the DR Site?
Or if my main Site cames down, it will be need to do this manually ?
If its automatic, how do I turn it manual.
Thanks in advance !!!
SRM is a manual failover only.
THere are ways you could make it automatic, but a datacenter failover is considered such a big deal that some manual action is expected to take place.
Hi Eric,
Many thanks for your great job. I have a case that need your suggestion.
1- if Protected Site failed/Disconnected of fired then i will manually run Recovery Plan to wake up Vms on Recovery Site. After all the VMs run i will run Reprotect to revesed Recovey site become Protected Site. Right?
2- After few days the Original Protected Site is rebuilt included SAN Storage and Vcenter…
So my question is:
1- How is the Direction of Recovery Plan on both Original Protected Site and New Protected Site?
2- Can SRM aware the new direction of Replication?
3- If SRM can not auto awrw new direction of replication, Must I recreate the new Recovery Plans and then Planned Migration all VMs back the Original Protected Site?
Thanks in advance
Truong Nguyen
Hopefully I understand your questions correctly here.
If you’ve failed over to the new site and want to fail back, a reprotect will reverse the replication for you. However, if you have to rebuild your original protected site from scratch, you’ll then have to re-setup SRM as well with the original site as the new recovery site.
As long as the original SAN, vCenter and SRM servers are still there, reprotect will setup SRM in the reverse direction and reverse the replication for you.