Seeding a Repl job with an excluded disk

[nixon]

Hi all,

We have a server that has several (separate) vDisks. One of the disks is used for a nightly file dump from SQL. We want to exclude that disk (SCSI 0:5) from the replication. I can see pretty clearly where this needs to be selected in the the job properties, and am aware of the issue of the DR server missing the drive - all good.

My question is with regards to offline seeding. With the temp backup job to build the files for the offline seed, do I:
1. Include the SCSCi 0:5 disk
2. Exclude the SCSI 0:5 disk; or
3. It does not matter

Are there any other issues that others may have experienced doing this kind of replication?

Thanks very much.

[Vitaliy S.]

I do not believe that matters. You may want to try to seed this VM with excluded disk from the backup job, however we need to double-check that.

Here is a detailed description of replica seeding procedure: v6 Offsite Replication : Seeding How To (if required)

[renpen]

[merged]

Hi, I am using VBR 6.0 (patch 3) and setup a backup job of a VM which has 7 disks on it. I have excluded the last 4 disks as these are not required and have selected the option to exclude disks from VM configuration. I have then transported this backup across to a DR site and setup a replication job with seeding and pointing to the repository where the backup file exists. The error I get when running the replication job is

Error: Backup file does not contain all VM disks and cannot be used as a seed.

In the replication job I have again excluded the disks that were excluded in the backup job. Either removing these disks or leaving all disks in the job produce the same error message as above.

Anyone else expoerienced this?? Habe logged a call with support and initial reponse has been quite poor.

[foggy]

I've confirmed with QC, currently you cannot use the backup job with excluded disks as a seed (irrespective of whether they are also excluded in the corresponding replication job). Moreover, if you have a backup job containing all VM disks and use it as a seed for the replication job where some disks are excluded, in case of restore you will get those disks on a replica VM at the time when that VM was backed up (they will not be updated by the replication job). Maybe this behavior will change somehow in future updates.

[renpen]

Thanks Foggy, figured it may be a bug or shortcoming in ver 6.0. It would be nice if first level Veeam support knew about this issue and could quickly answer - save us all a heap of time!!

[tsightler]

You should be able to perform a manual restore of the VM without the excluded disk and then simply map the replica. That's basically all the "seeding" process does anyway.

[renpen]

yes, I tried this but it seems to scan the entire VM as CBT changes have not been tracked. I only have a 2MB link so when I tried this the calculating disk digests took hours just to get over 1%

Thanks

[tsightler]

That would indicate that you don't have the proxies configured correctly. There needs to be a proxy on both the source and the target side of the link. The proxy on the target will perform the scan so that the speed is not impacted by the link bandwidth. If you do have two proxies and you're still seeing this issue then you may need to manually select the proxies in the job configuration.

[renpen]

pretty sure I checked this but will try again..

Thanks

[tsightler]

The entire concept of replica mapping is that you can use it for seeding in cases where a copy of the VM is already on the target side, perhaps from a previous replication job, an alternative replication technology, or even a restore of a VM. If it's using significant bandwidth during the "calculate disks digests" step then something is not configured correctly.

[nixon]

So it sounds as if it is best to "1. Include the SCSCi 0:5 disk" in the backup job for seeding.
Thanks for your help.

[foggy]

Nixon, you can try what Tom has suggested: backup the VM with this disk excluded, restore it manually, and then map the replica job to the restored VM. Should work. Otherwise you'll have this disk outdated on the replica VM.

[dbarroco]

[merged]

Hello,

I know this might be odd but I'd like to check for a procedure on the following scenario:

One of the VM's on the original cluster has a RDM disk. Backups are made with Veeam, disk is excluded, it goes well.

We're also replicating vm's to a DR site. This is where it gets more complex:

When seeding this particular VM from the backup it errors due to a missing disk. (this RDM disk contents are also replicated between 2 servers with MS DFSR so no backup to consider here for the purpose).

My questions:
Can I manage replica snapshots and future failover + failback with different vm settings on each host? (original with RDM disk, failover host without this disk).

If the above is not doable (because the vm is different on each site), how could I make this replication and failover happen?

Thank you
David

[dbarroco]

Hello,

My post was merged here. This is what I had in mind, using a restore and a mapping procedure.

However I was not too sure on the following:

When failing over and/or back, the disk information in the only info that crosses between hosts? Won't the vm configs be rewritten?

Thank you
David

[foggy]

David, if you exclude the disk in replication job, VM config will not be rewritten.

[andrewpetre]

Possible this could be submitted as a feature request? Allow replication to check disks selected vs backup disks selected and perform successful seed based on matching omitted disks?

[Vitaliy S.]

Yes, thanks for the feedback, meanwhile Tom's workaround is still the best way to address this scenario.

[fll]

Hello, we need to know if a backup job of a virtual machine with excluded disks can be used for a replication job with the replica seeding option enabled. It seems that it is not so ('Backup file does not contain all VM disks and can not be used as a seed').
Is it necessary to perform a full backup job including all drives to perform a partial replication of the virtual machine or is it possible to add the missing drive to the backup job in order to do partial replication?

Thank you.

[DGrinev]

Hi,

You can do that if these disks will be excluded in the replication job settings for the particular VM too. Thanks!

[fll]

Hi,
I have repeated the process:
1.- Generate a backup job that includes only the system unit of the virtual machine, but not eliminating it from the configuration if it can be included later in the replication.
2.- Generate a replication job with the replica seeding option activated, selecting the system unit (as in backup job) and the option "Source Repositories" configured in "From backup files".

When you launch the replication job it fails and shows the following error:
"Backup file does not contain all VM disks and can not be used as a seed".

Thank you.

[DGrinev]

Hi,

Did you exclude the same disks in the replication job on the Virtual Machines step -> Exclusions -> Disks?
The error you've mentioned tells you that you are trying to replicate more disks than the backup seed contains. Thanks!

[fll]

In both cases I choose the option "System disk only"

Thank you.

[DGrinev]

Is there any chance you have another backup with the mentioned VM without disk exclusion in the repository that was picked for seeding? Thanks!

[fll]

No, in the repository there is only the backup that we are talking about. There is nothing else.
Thanks!

[foggy]

Hello, there was some misunderstanding between Dmitry and QA, it is not actually possible to use backup with excluded disks for seeding. However, what you could do to seed is change the source of the replication job to the corresponding backup file (replication from backup), make sure exclusions are configured similarly to the source backup job, and perform initial replication job run, then switch replication job source to the original VM back (preserving exclusions). In this case replication job will be seeded and continue normally.

[fll]

Hello, in fact, using the corresponding backup file as source of the replication job, with the same exclusions in the backup job and the replication job, the replication ends correctly.
Will a faillback from a replica with exclusions eliminate the excluded units in the original machine?
Thanks!

[foggy]

No, original VM disks will be left as they are.

[fll]

Thanks!