Repository for Replica metadata.

[cag]

What happens if I need to change or have lost the location of the respository metadata. Can I restart a replication by mapping source to target. Will this rebuild the data ??
Is this location important enough to backup ?

[Sethbartlett]

The data should be rebuilt, it is not THAT important to replication. It weighs about nothing and the VM disk digests would just need re-calculated if it was to disappear. Just make sure you put the metadata on the source side, not target side.

[cmdrriker]

Does this mean if I am using a proxy on the source side, that I should add the source proxy as a backup repository and store the metadata there?

[Vitaliy S.]

Any backup repository located in the source side will work.

[davidb1234]

[merged]

Prod Site
-Physical Veeam Server direct SAN attached to ESXi Cluster. Also direct attached storage for backups
This server runs backups at the Prod Site for all VMs and uses the physical Veeam Server as the proxy because of direct SAN mode
-Virtual Veeam Proxy used for Source Replication at DR Site

WAN LINK 100MB

DR Site
-Physical Veeam Server, Direct San attached to ESXi Cluster, used for Replica Seeding and DR backups if necessary
This server runs the replications jobs and PULLs them from the Prod site.
-Virtual Veeam Proxy used for Target Replication at DR Site

Would it be best to create a repository at the DR Site(target) for replication metadata or to create a repository that links to a storage location in the Prod Site(Source)?

I read conflicting things on where to store this metadata for performance. If the DR site is running the jobs and pulling the data should it have the metadata closest to DR site or should it be at the Prod site proxy closest to the actual data being pulled?

If creating a repository for replication metadata from the DR Site Veeam server pointing to a storage location at the Prod Site across the WAN link is the best way what happens if the Prod Site is completely unavailable and you need to fail over? Will I be able to Fail over and ReIP properly without access to this metadata?

[Gostev]

Actually, the replication wizard itself tells you what site to put metadata in - just read the labels on that step carefully no conflicting statements there...
Metadata is not used for failover or Re-IP, so don't worry about that part.

[davidb1234]

I am not sure what you mean Gostev by replication wizard tells you. I created a replication job and there is nothing that would indicate where to put the repository it just has a dropdown list of repositories to choose from on the JOB SETTINGS step of creating the replication job.

So my question remains. Should I create a repository in the DR site veeam installation pointed to a storage location across the WAN link on the PROD site or leave the metadata repository in the DR site?

my confusion is based on what is source and what is destination. Is the DR site considered source since that is where the replication job is running from or is it talking about source of where the replicated data is in which case I should create a repository across the WAN link from my DR site veeam installation.

Currently the replication metadata is stored on the DR Site Physical Veeam server where the replication job runs but I was wondering if this was a mistake and I should create a new metadata repository across the WAN link at the Prod Site and then change the setting in the job. I am also wondering what impact this will have on the replication job by changing the metadata location.

[Gostev]

I am not sure what you mean Gostev by replication wizard tells you. I created a replication job and there is nothing that would indicate where to put the repository it just has a dropdown list of repositories to choose from on the JOB SETTINGS step of creating the replication job.

That's right. Now, read the two lines description label on top of this dropdown list, and you will see the repository placement recommendation right there
Source site is obviously the one you are replicating VMs from (in other words, production site where virtual machines you are protecting with replication reside).

[davidb1234]

I must be blind. I see it now.. haha. Thanks again Gostev. I will create a new repository and just switch it in the job for the next run. It sounds like it should not impact anything by changing the location.

I created a metadata repository on the C:\ of the virtual source proxy in the Prod Site.

[Gostev]

Yep, don't worry about that. If this data is ever lost, corrupted or not found in the specified location, the replication job will re-create it automatically by having target proxy scan through the replica VM disks in the DR site, create the new metadata file, and send it the specified backup repository.

Essentially, replica metadata files represent actual state of all replica disks blocks, and this information is used by the source proxy for incremental replication - thus the placement recommendation.