FEATURE REQUEST- Control over shutdown of VM during Failback

[smunawar]

Greetings Veeam Community!

I have a feature request that I feel is not only critical but also relatively easy to implement. This feature request would drastically reduce downtime and maintenance windows for when performing a Failback. Let me elaborate:

-When a Failback is initiated, a snapshot is created then the base disk is replicated to production. This process can take fairly long depending on how long the replica was in an Active Fail Over state and how many changes were made. In some cases it may need to replicate the entire base disk back to production!
-During this entire process, the VM is still running and users may still be making changes on the replica VM.
-Once the base disk is replicated, Veeam proceeds to AUTOMATICALLY SHUT DOWN THE ACTIVE REPLICA VM. Users may still be logged in and making changes and can risk losing data or data getting corrupted due to the shutdown. Veeam leverages VMware Tools to perform a shutdown of the guest OS for this procedure.
-After the VM is fully shut down, the remaining changes are replicated to production.

Due to the nature of the Failback, the best practice is to provide a maintenance window for the entire Failback procedure, even though the initial replication of the base disk could take a very long time!

MY FEATURE REQUEST:
During a Failback, allow the Veeam administrator the option to wait for approval before shutting down VM OR to automatically shut down the VM after the base disk is replicated.

This simple feature would drastically shorten maintenance windows for Failback. Unless I'm missing something, I believe it would be relatively simple to implement.

Thank you for reading and considering this request!

[veremin]

Wait untill Veeam Availability Suite v8 is released and check whether planned failover functionality is something you're looking for (it seems so, from my perspective). Thanks.

[smunawar]

Thanks for the reply v.Eremin. I do not think v8 is going to address my concern and request. This v8 "Planned Failover" feature does look good, however, there is no reference to the Failback feature which is my concern. Failback's are always something that are planned and there is no mention of enhancing the Failback functionality in v8 (only the Failover).

It's actually very easy even now to perform a "Planned Failover" manually. Simply do a regular replication job to replicate the major changes, then shut down the VM's, perform another replication job for the remaining changes, and then Failover. It's cool that Veeam is automating this in v8 though.

In reality, I don't see why everyone else isn't clamoring for the same thing I'm asking about- this would benefit every Veeam administrator. Doesn't everyone want to minimize maintenance windows when performing Failback procedures?? That's essentially what my proposal would accomplish. The ability to perform a Failback, replicate the major changes to the base disk (which could take a LONG time), then give the option to either (1) wait for approval from the Admin before shutting down the VM instead or (2) shut down the VM automatically, and then replicate the remaining changes.

v.Eremin- please correct me if I'm mistaken, but I am quite positive the Planned Failover feature is not adding this FAILBACK functionality in v8.

[veremin]

Yep, the planned failover has nothing to do with failback. However, it still might answer your original requirements. Perform permanent failover to DR site, set replication jobs in other direction (from DR to production), run replication job. Once the it's finished, perform planned failover (in the allowed time window) that will finalize the state of production VMs and switch the load to them.

This way, the major data transfer will go unnoticed by users, and remaining changes will be copied only when you decide to do so.

Or you want something different to what I've described above?

Thanks.

[smunawar]

Greetings,

v.Eremin, unfortunately, that is not an ideal solution. Performing a permanent failover to DR site then changing replication jobs has a few undesirable effects. I have spoken to Veeam tech support about this "workaround" and even they didn't recommend it as the best approach and recommended to do the standard failover and failback when in a DR scenario to minimize the risk of issues.

Issues with the workaround such as the following:

1) If one is to permanently Failover to the DR site simply for the purposes of Failing back, additional licenses are needed when doing a replication from DR to Production. Generally, most companies do not have the target replica hosts as licensed hosts. There are typically no VM's being backed up and replicated on a target DR host. So if one is to do this, additional licenses will be needed simply to Failback VM's.

2) I have found that it's safest to simply failover and failback rather than doing a permanent failover, modifying existing replication jobs, replicate, failover again, and modify replication jobs again. This causes undesirable and unforeseen consequences. Some of them include the replication job replicating the entire base disk instead of just changes, the VM being registered with the name "_replica" on the production host after all is done, and strange error messages when modifying the replication job.

Again, my feature request is something I believe to be simple to implement by the Veeam software engineers, please let me know if I'm mistaken. It's simply a matter of giving the option to the Admin to wait before shutting down the VM or automatically shut down the VM after the base disk is replicated. This would reduce Failback maintenance windows manifold.

I really hope this message gets to the software engineering team as something to consider.

Thank you,
-Saud

[wa15]

While we were testing the replication feature of Veeam for DR purposes, this was one of the drawbacks that made us look elsewhere for DR. SRM allows this. It would be GREAT if Veeam can add this option as well.

[smunawar]

wa15 - thanks for the reply. Your post perfectly illustrates the magnitude of this issue. The Veeam leadership team should be sounding the alarms, have an emergency executive board meeting, and convene all department heads to implement this feature!!!!

[Gostev]

I've logged this FR, thank you for both your feedback. You are right, most features are easy to add, but there are too many pending features, so we have to prioritize them based on the amount of requests. And to be honest, this is the first time when this particular feature is requested for all the time since we have added Failback feature to our product. Thanks again!