Backup the replicated VMs

[veremin]

For other issue related to CBT

Actually, there are two existing topics regarding the issues you experience; might be worth reviewing:

Cannot use CBT
OIB already exists in point

Hope this helps.
Thanks.

[kamalkant]

thanks v. ...I have been gone through above both of the above links, my issue is "OIB already exists in point" ... this links talk about some custom patch from Veeam , I ma not sure if I need that.
as I have already worked out on all other recommendation. so i am going to open a case

[veremin]

So I am going to open a case.

This is probably the best way to address technical issue.

Furthermore, kindly update the corresponding topic with a found solution for the convenience of further readers.

Thanks.

[rmessier1975]

I have 3 hosts ... One with live production servers, one that has replicas and some live servers, and a test environment. IS it NOT good practice to backup the entire host if replicas reside on it? I mean, if you are already backing up the live server, why backup the replica? Or can Veeam Backup even backup replicas?

[Vitaliy S.]

Hello Keeney,

Veeam can backup replica VMs, no issues with that. Please see this thread for more details on this process. Using VM containers such as folders, datastores, hosts is actually a recommended way of backing up your infrastructure. However if you want to exclude replica VMs from the backup job, then this is also possible via Exclude options of the backup job settings.

Thank you!

[veremin]

Hi, Keeney.

However, please be aware that it won’t be possible in this case to utilize CBT and it’s a Veeam built-in mechanism that will be used, instead. So, the jobs are likely to take more time to finish.

Thanks.

[tfloor]

How about the initial question for v7

If you need backup and replica jobs on the DR site.
Do we need 1 copy job to DR, and 1 replica job to DR

Or choosing to backup the replica vm's. Is this working the same as on the primary site.
Creating a backup Job with the vm selection from the DR esxi hosts, and give the offsite storage as destination?

[Vitaliy S.]

Tristan,

If you need backup and replica jobs on the DR site.
Do we need 1 copy job to DR, and 1 replica job to DR

Yes, if you want to have offsite backups and VM replicas, then you should have 1 backup copy job and 1 replication job.

Or choosing to backup the replica vm's. Is this working the same as on the primary site.
Creating a backup Job with the vm selection from the DR esxi hosts, and give the offsite storage as destination?

Nothing has changed in regards to backing up VM replicas - CBT option will be disabled.

Thanks!

[cmdrriker]

two questions:

1. can a replica safely be backed up?

2. will the replica backup include all of the replica restore points?

we will need to recover soon from a cryptolocker infection and I am trying to retain all of the restore points in case the infection comes back after a replica failover and failback is done.

[foggy]

Will, please review this topic for considerations regarding backing up replica VMs.

2. will the replica backup include all of the replica restore points?

No, since existing snapshots get consolidated into a single flat file during backup job.

[cmdrriker]

Foggy,

If the restore points of a backed up replica are consolidated, which version is preserved? The last (most current) one?

[foggy]

Yes, the current state of the replica VM is backed up.

[EddyChong]

Hi Foggy,

I named replicated VM as "replica" and replicated replica as "replica_replica"
And I assume the restore point snapshots consolidation applies to replication job as well. Please correct me if i am wrong.

If restore point snapshots of a replica is being consolidated during the replication/backup job, that means the replica's restore points will disappear and only most current state available? Or it simply means the replica's snapshots is preserved but the replica_replica will have the most current state of the replica.

Thanks
Eddy

[dellock6]

The second replica job only sees the actual state of that VM, ignoring its previous snapshots.
The several snapshots of each replica and their retention is managed independently but the two corresponding jobs, and removing a restore point from the first replica will have no impact on the second replica.

Luca.

[foggy]

Or it simply means the replica's snapshots is preserved but the replica_replica will have the most current state of the replica.

This one.

[EddyChong]

Thanks Luca and foggy.

Eddy

[gavinmclean]

Hi All,

I know this question has been asked before... http://forums.veeam.com/vmware-vsphere- ... 03-15.html, however, is it possible to backup replicated VMs with CBT enabled (using Veeam v7)?

Currently we have VMs in a hosted DC, being replicated with Veeam to our DR site (what was originally production), which is working really well. We'd like to backup the replicated VMs for the purposes of long-term retention, and we've got the backup jobs configured, unfortunately, CBT doesn't appear to work the backup jobs, which means they are taking far too long (they aren't even close to finishing before the next set of replication jobs start - replication jobs run daily). Long term, we will also have backup storage in our hosted environment, but this would be (very) short term retention only. I am aware that we could backup the source VM to the backup server in the DR environment, however I'd like to avoid shoving what is effectively the same data through our link twice, given that we already have a limited link. It would also mean, once we've got the short term retention backups working, our production VMs would be getting hit 3 times...

So far I have tried removing the snapshots on the replicated VMs (we only have 1 per VM), resetting CBT, and running a backup job on the replicated VM (twice), which indicated that CBT was working. However, after running the replication again, and subsequently the backup job - CBT is again not working

I have read this should work, particularly given that Veeam v6 (and presumably v7) replicate the VMs differently to previous versions of Veeam. I have also read that there can be issues with CBT on powered off VMs. Can anyone shed some light on this situation?

Cheers,
Gavin

[foggy]

Gavin, actually, nothing has changed since this reply from Vitaliy - CBT is still cannot be used when backing up replica VMs.

What could probably make you feel slightly better is that we're planing to add "replica from backup files" functionality down the road. This will allow to avoid pushing the same data over the slow link twice, as well as hitting production VMs twice.

[kevin.mcdonnell]

When attempting to backup a VM replicated by VBR7 I noticed an error that "VMware CBT could not be enabled due to an existing snapshot".

When a VM is replicated VBR sets snapshots to give the ability to recover previous versions - the minimum being just 1.

Even if the source VM has CBT enabled, VBR does not replicate this setting to the target.

It is not possible to enable CBT with a snapshot in place and any attempt to modify the VM "under the covers" results in VBR doing a full sync and overwriting the changes.

The result of all this is that any backups of the replicated VM can not take advantage of CBT.

Has anyone else seen this or have a work-around.

Cheers,

Kev

[foggy]

Kevin, this is actually an expected behavior, please review the thread for details. Also, please be aware that the just released v8 has the remote replica from backup functionality that allows to have both remote backup and replica of the same VM without the need to transfer the same data twice. Thanks.

[romwarrior]

So in v8 what are the ramifications of backing up a replicated VM for an offsite backup copy? Is it only that CBT can't be enabled? I know the preferred way is backup copy first but I have one VM that the backup copy did not succeed while the replication did.

[Vitaliy S.]

Actually, in v8 there are no changes in how replicated VMs are backed up (CBT can't be used due to VMware limitation). Alexander was referring to a possibility to use backup files as source for replication jobs. Thanks!

[severson_bnc]

Just wanted to chime in since it's been awhile for this thread.

I have successfully setup an ongoing Backup from Replica configuration using CBT on VMware ESXi 5.5 and Veeam 8. CBT remains enabled in the job logs, and the performance is there when backing up from the replicas. I set my replica's to only have a single restore point at the moment, so I cannot vouch for a setup that has more than that per replica.

Hope this helps someone, this thread was the top forum hit for backing up from replicas.

[Vitaliy S.]

Was it a first job run or an incremental pass? What do processed and read counters show?

[severson_bnc]

Incremental jobs have maintained the CBT in the logs. Processed and read counters show the transferred data to be at a significant savings. Here's a screenshot: http://i.imgur.com/nzLu2Er.png.

I've had some speed issues as the replication and backup jobs interfere with each other. The window is approximately 40 minutes, so I'm now setting it up so that runs every other hour, instead of hourly.

[Gostev]

From the statistics on your screenshot, it does not look like CBT is being leveraged. Can you select any VM to the left for per-VM statistics?

[Vitaliy S.]

Processed and read counters show the transferred data to be at a significant savings.

These counters will always be larger than transferred value, in the full and without change block tracking runs.

[severson_bnc]

I'm still working out some kinks in this environment- for example, some VMs were using network backup mode due to the lack of a proxy. I'll have that taken care of in a few days.

Any logs in particular that I could pull that would give you guys a more direct answer if CBT is working or not?

[foggy]

Please select one of the VMs in the list on the left on your screenshot and look for the [CBT] tag in the hard disk processing line.

[Vitaliy S.]

Please note that you will [CBT] tag only on the first full job run, all subsequent runs should report as CBT disabled.