Replication job can't find dirty block information for CBT

[crond]

Hi All,

Scenario:
* VBR11
* VM on serverA(esxi 7.0.3) is replicated to ServerB(esxi 7.0.2)/replica_VM ; each server has a backup proxy deployed in a VM
* Replication job is executed several times to confirm that "CBT" works as expected
* A backup job for replica_VM is added and executed successfully (manual trigger).
* After that all runs of replication job give a warning "Cannot find dirty block information in the previous restore point, CBT will not be used" and job proceed to read entire VM

Both VM and Replica_VM have no snapshots.
Is it expected behavior ?
How do I get CBT "back on track" for replication jobs ?

[Mildur]

Hi Dmitry

Both VM and Replica_VM have no snapshots.

Sounds wrong. The Replica VM should have snapshots as restore points.
No Snapshots on the Source VM is ok for Replica Jobs.

I run the same test in my lab and I couldn't see the same behavior.
Please open a case with our customer support and let them analyze the logs to find the root cause of this behavior.
Don't forget to share the case number with me.

How do I get CBT "back on track" for replication jobs ?

Starting a new replication of this VM should solve it.

Best,
Fabian

[crond]

Both VM and Replica_VM have no snapshots.

Sounds wrong. The Replica VM should have snapshots as restore points.
No Snapshots on the Source VM is ok for Replica Jobs.

Let me clarify this ... Both VM and replica_VM do not have any manually created snapshots or snapshots created by 3d party apps.
Replica_VM does have all "Restore point XYZ" snapshots created by VBR

[crond]

Please open a case with our customer support and let them analyze the logs to find the root cause of this behavior.

Don't forget to share the case number with me.

I'm on free tier, so if you you have any ideas / recommendation for experiments I can run, that would be great.

Starting a new replication of this VM should solve it.

Best,
Fabian

It doesn't solve it. After the is triggered, all replication jobs unable to detect CBT information.

The only solution I found so far that solves the issue is to manually delete all snapshots on replicated VM. and create a new replication job / map existing replica VM.

[Mildur]

Maybe I missunderstood you.
Are you getting the dirty cbt info on your Replica Job or on your Backup Job which protects the replica vm?

The later is expected. CBT cannot be used for replica VMs in a Backup Job.

Best,
Fabian

[crond]

Sorry for confusion.

* I don't expect CBT to work for Backup job for Replica_VM
* I expect CBT to work for Replication job.

To recap
#1 Once new Replication jobs is created and "Replicated VM" is mapped, it works as expect and able to obtain CBT information and transfer changed blocks
#2 After a new backup job for "Replica VM" is created and executed (successfully), all future runs of #1 "Replication job" are unable to use CBT. In Action windows I see following message "Cannot find dirty block information in the previous restore point, CBT will not be used"

[crond]

Ok I think I found the problem.

It appears that
1) By default replication job (CBT=on) sets ctkEnabled=FALSE for replica_VM
2) everything works as expected including CBT tracking on source VM for replication jobs.
3) when backup job is created (with CBT=on) for replica_VM, it removes ctkEnabled=FALSE from replica_VM config (?)
4) replication job is unable to find / use CBT.
5) Manually setting ctkEnabled=FASLSE on a replica_VM, resolves the issue. I.E replication job is able to find/use CBT information.

it's a bit of a surprise that replica_VM config can affect VBR ability to use CBT for source VM, taking into account that replication is done based on snapshots.