- This topic has 5 replies, 4 voices, and was last updated May 7, 2024 by
Julian M.
Zerto VRAH behavior during live failover
-
I recently conducted a Live Failover for a customer with reverse replication enabled. This customer did not want to use production VMs, so we temporarily added VMs to Zerto in order to go through live failover process. These VMs were failed over with reverse protection enabled.
A week later we failed the VMs back to the protected site. Reverse protection was NOT enabled, since these VMs no longer needed to be protected by Zerto. Days after the test, VRAH vms were still on the protected site.
My questions are:
- Do VRAH vms ever get created on a protected site?
- If not, shouldn’t those VRAH vms get cleaned up after the live failover operation back to the protected site?
- Did reverse replication not being enabled have anything to do with this?
Hi Alex,
- A VRA-H is created whenever a VRA detects that it needs one to overcome the VMware limitation of 60 disks per VM. They are just disk-boxes that hardly use any resources (CPU/MEM).
- If a VRA-H is not needed anymore, it will be deleted.
- VRA-H creation and deletion is at the discretion of the VRA. Its presence or absence is fully transparent to any operations that are performed on VMs that are being protected by that VRA.
Basically, just ignore any VRA-Hs you see as they don’t use resources and are transparent.
Is it normal to see VRA-H’s on the protected site?
Julian MMarch 22, 2024 06:01:52 AMNo, VRA-H VMs are only present in the recovery site where replication is received and needs to be stored somewhere (on a VRA duh!).
In production, your data is stored as vmdks on the production VMs. The VRAs in the protected site are only used for replicating data to the recovery site and thus there should be one VRA per host and each one will have 4 OS disks.
If you find a VRA-H in the protected site then at some time the protected site was used as a recovery site maybe in a Failover Test. Or you may have one or more VMs replicating from your recovery site to your protected site that you are not aware of!
Spritney BMay 7, 2024 07:30:19 AMHi duck life
1. Generally, VRAH VMs are deployed on both the protected and recovery sites to facilitate replication and failover operations. These VMs are crucial components of the Zerto infrastructure and play a key role in managing data replication and failover processes.
Julian MMay 7, 2024 08:22:56 AMVRA-H VMs are ONLY found at the recovery site in a normal Zerto setup. Fact!
Check your protected site and you will not find a VRA-H. The protected site does not need a VRA-H as the VRAs in the protected site do not hold replicated data- they just monitor writes to protected VMs, and copy the writes to the VRA at the recovery site so no extra vmdks are required in production. All protection VRAs will thus have 4 vmdks about 12.5GB in total.
The only way you will find a VRA-H at the protected site is after a failover (Protected -> Recovery sites) where the recovery site is now the protected site. Now, if reverse replication is enabled, replication flows from the recovery site to the protected site so yes in this case you will find VRA-H appliances in the protected site however not in the recovery site.
The same is true if you have 2-way replication between sites, where you have VMs replicated from both sites, then you will find VRA-H appliances in both sites.
So to answer OP’s questions, VRA-H appliances should only be located in the replicated site. If you find them in the protected site, then you either have a cleanup issue or you still have replication from the other site coming in. To check, note down the VRA-H name which should include the host name, then open the Zerto GUI, go to the Setup tab, find the VRA with the same name as the VRA-H and click on it to open the VRA GUI. Open the Settings tab at the right-hand-side and it will list the shadow VRAs which are the VRA-H appliances. If there are no shadow VRAs yet you see the VRA-H in your environment then you can be pretty sure to delete the VRA-H.
