- This topic has 1 reply, 1 voice, and was last updated August 31, 2017 by Lukasz Z.
ZVM protection without recovery site
You must be logged in to create new topics.
Click here to login
When there are sites or hosts dedicated seperately for production and recovery there is no problem with keeping the ZVM safe, it resides on recovery site/host and in case of production failure it works as it always did.
But what when we have ZVM in an environment where there are only two hosts and two storage arrays (each connected with both hosts) and each of the hosts works at the same time as production and recovery, meaning there are vms on the first host in production being replicated to the second host and the other way around (vms in production from second host replicated to the first host).
If ZVM is placed in the first host, and the power goes down turning off the first host and first array, then there is no ZVM on the second host and there is no possible protection or recovery.
My question: Is it possible to replicate ZVM to the second host and second array, and in case of power failure of host bring it up on second host using only vSphere replication or I should be doing or using something else? Won’t the embedded database of ZMV be corrupted on the replicated vm?
I am working in such environment and would like to keep it safe. I look forward to your reply.
Regards
Lukasz
It occurs that without two vCenters it is not advised to replicate the ZVM in a “self site” environment because of possible database corruption.
I am posting reply from the Zerto support:
“The scenario you are describing is a common scenario which we call ‘Self Replication’.
Under this replication we are replicating between two hosts under the same vCenter, therefore, only one ZVM exists.
When replicating under this condition your protection is ‘weaker’ then having two vCenters and two ZVMs – one at each site.
Regarding the state you’ve described, if there is a power outage on the host where you have the ZVM installed, Zerto operations can’t be performed.
Your insight regarding replicating the ZVM and causing the DB to be corrupted due to a migration of the ZVM to the second host is true and we don’t advise on doing so.
To sum up, in order to avoid ZVR failure due to ZVM being down the best solution is to have the two hosts on different vCenters with ZVM installed on each of them and have them paired through the Sites tab.
Otherwise, using vSphere HA and moving the ZVM to the second host using simple vMotion is possible if a host down-time is expected.”