- This topic has 3 replies, 1 voice, and was last updated November 29, 2019 by Przemyslaw J.
Promotion & Delta sync after DR failover?
Team PMay 30, 2018 10:24:09 AM
Just wondering what is happening in these steps after a DR failover:
Promotion – given that it has powered the vm on already and it is available on the network, what is the promotion process actually doing at this stage as the failover has already completed?
Delta Sync – this can take a while on larger VMs. Given that the source vm was shut down and that all data should have synced between sites, what is it delta syncing when the destination VM has come up? Shouldn’t all the data already be at both sites?
many thanksTeam PJuly 11, 2018 02:07:23 PM
any ideas?Sean MJuly 11, 2018 03:05:35 PM
The promotion process involves “promoting” the journal volume(s) into the replica volume(s). Promoting meaning writing the journal blocks into the replica space. This is done to eliminate the recovery site journal volume(s) as it/they are no longer needed. We don’t journal in the source site and, after a failover, this is now the “new” source site. This is transparent to the VM and the hypervisor/platform it’s on.
Zerto performs a Delta Sync after a Live Failover, assuming you enable Reverse Protection and we have network access back to the source site, due to us not wanting to trust the source data. What we may want to assume is the same data may not be the same at all, especially after a disaster. In addition, you are not failing over to the current disk state as Zerto replicates asynchronously, and so a Live Failover, even using the current checkpoint which is only a few seconds old, is not 100% synchronized with the source. Although you may shut down the source VM as part of failover, we can’t assume the type of disaster that led you to click Failover, so we don’t replicate data coming out of that VM during the Failover process.
Side note: if you do want a perfectly synced VM recovered at the Target site, you want to use the Move operation. We’ll still do a Delta Sync in that case, again assuming you select Reverse Protect, to ensure a 100% accurate Move or Failover. Just as with a Failover, we can’t be making assumptions, for example that the data wasn’t changed by some other process or system after our 100% synced migration, so we Delta Sync.
-Sean M, ZertoPrzemyslaw JNovember 29, 2019 10:53:25 AM
Why after the VPG move the Delta Sync is performing Delta sync on data that was in the VPG ? is this by design ? I though that delta sync will be only data difference between during the move. Example we move a 7TB VPG to another site with reverse protection and autocommit set to 120minutes to allow to run checks on VM’s.
The VM’s checks takes 1 lets say 30 minutes, then we do a commmit and after a commit Delta Sync is synchronising all the Data in the VPG.
Thank in advance for your response