how to fix replica invalid snapshot configuration

[frigomiam]

Hi all,

veeam B&R v7,
1 replication job returns "Detected an invalid snapshot configuration"
Another job works fine with same production VM, so it's not issue on the prod VMs.

I found there were invalid CIDs in the vmdk files on the replicas which I corrected.
But after, the job failed again with same error message
Now the replication snapshot manager points to 1 restore point as usual. But the virtual hard driver points directly to my original vmdks, not to the snapshots files. There is no more snapshot files within the VM replica folder.

What's the best direction to fix this ?
should I delete everything and recreate the jobs ?

I have to say I also powered on manually the replica VMs at 1 stage ( without taking a new snapshot as required ).

thanks

[foggy]

I have to say I also powered on manually the replica VMs at 1 stage ( without taking a new snapshot as required ).

This could confuse the job, since starting replica manually creates a snapshot that is not controlled by the job.

I would simply create another job to start from scratch, leaving the current restore points temporarily, until the new job reaches its retention.

[frigomiam]

I don't have storage to create new replication job and new replica VMs. Can I just not delete or rename the replica VM, and start again the job so this will recreate the replica VM ?

[Vitaliy S.]

Can I just not delete or rename the replica VM, and start again the job so this will recreate the replica VM ?

You can force creation of active full run by removing VM replica from inventory and registering it again with a different name. Not sure that just renaming will change anything, since replication jobs track VMs via moref ID.

[frigomiam]

If I remove the VM replica from veeam B&R inventory list ( remove from disk ), will the current replication job recreate a new replica VM on the next job occurence ?

[foggy]

Yes, the replica will be re-created.

[frigomiam]

I deleted from disk the replica from veeam B&R interface,
that didn't work,

all VM for which I deleted the replica VMs, had their replication job failing with
with "task failed error: Object reference not set to an instance of an object"

will the following fix the issue ?

- edit replication job
- seeding / replica mapping / detect

( in seeding some of VMs had replica VM = No mapping, some others had mapping to 1 replica VM that don't exist because I deleted them )

or should I recreate everything from scratch ?

[kjc3303]

If I understand correctly I think turning off seeding altogether is the best option. The VM's that are already there will contunue as normal and the ones you deleted will be created from stratch.

Kev

[Vitaliy S.]

It's hard to say what is going wrong based on this log snippet, try to re-add source VMs to the job and trigger full replica cycle. If it doesn't help, then please contact our technical team for further troubleshooting.

[Vitaliy S.]

If I understand correctly I think turning off seeding altogether is the best option.

Yes, seeding is not needed, since this is going to be fresh VMs created on the target host.

[frigomiam]

all right, thanks for the advises,
I will follow http://www.veeam.com/kb1850 to disable replica mapping ( I'm not using replica seeding ..)

I understand that the replica VMs that were deleted will be recreated. And for the one that exist already, I understand the whole content of prod VM will be copied accross and overwrite the existing replica VMs.
If I'm wrong let me know !

I will test the result tomorrow morning, thanks again.

[Vitaliy S.]

And for the one that exist already, I understand the whole content of prod VM will be copied accross and overwrite the existing replica VMs.

Correct, assuming that you do not target existing replica jobs to these VM replicas.

[kjc3303]

I understand the whole content of prod VM will be copied accross and overwrite the existing replica VMs

.

Just the changes since last replication will be copied across as a restore point (snapshot).

*edit I assume you will be using the same job.

[frigomiam]

.

Just the changes since last replication will be copied across as a restore point (snapshot).

*edit I assume you will be using the same job.

even if seeding and replica mapping are disabled as per kb1850 ?
veeam can automatically check if there is an existing replica VM that matches the prod VM ?

yes it's the same job

[kjc3303]

Yes, it will just continue as per your retention policy ( assuming the Job has\is already written to that target VM). I have turned off replica seeding\mapping after the initial seed\map and the job continues as normal ie doesn't create a new VM.

I have just tried it on one of my replica jobs - unticked the seeding check box - ran the job - result was it added a restore point to the exsisting VM (the job must hold the UUID's of the source and Target VM's after a sucessful Rep)

[foggy]

Replica seeding and mapping settings take effect during the initial job run only. All subsequent job runs are incremental and do not take these settings into account (except the side issue described in the mentioned KB).

[frigomiam]

all right, got to say, I finally had to copy the VMs from 1 datastore to another on DRSite and used replica mapping, so I haven't run the whole copy ( it was too long to run ).
Thank you, this was very helpful to understand