Replication Meta Data

[lobo519]

What is best practice for the replication meta data repo? How much space can we expect this will use?

[Gostev]

You should be using a backup repository in the source site (as explained in the wizard).
Disk space required is about 40MB for every 2TB of data replicated (that is, with the default block size).

[mrt]

what happens when it is lost?

[Gostev]

This is a good question that I forgot to put in FAQ.

In this case, replica metada will be automatically re-created during the next job run (which will obviously make it run longer). To do that, target backup proxy will read and hash the entire replica VM disks contents, and send the metadata back to source site, where it will be stored in the selected backup repository again.

[averylarry]

I messed that up originally by having the metadata stored at the destination site. Does it really make a difference?

[Vitaliy S.]

It does, especially when you perform a failback operation (it will be significantly faster), so it's better when you place metadata repository on the source site.

[Gostev]

Having meta data in DR site will slow down every incremental run, because the source agent would have to pull the metadata over from DR site over WAN. Depending on the metadata size and WAN speed, delay may vary from unnoticeable to a few minutes.

[averylarry]

Excellent. I originally messed up and didn't have a backup proxy (by the way -- I think it should be called "backup and replication proxy" or just "Veeam proxy" or something because it took me awhile before I figured out that I needed a backup proxy for replication, but I'm not always that smart and I tend to learn by doing instead of by reading user guides -- but I digress). The metadata was taking forever because it was being read from a source-side proxy. I thought that the repository was supposed to be in the destination, and when I did that (which happened to include creating a backup proxy) it worked fast. It was a mistake with the proxy instead of the metadata repository.

[mrt]

does this mean i could "move" a replication job from one backup server to another, if I move the metadata folder with it? that would be an enormous time saver

[Vitaliy S.]

Neither backup nor replication job can be shared between two different backup consoles, however these jobs can use different proxies if you happen to change them.

[mwant]

[merged]

I have a DR site that I intend to replicate my VM's to and have a source side backup repository that I am going to use for the metadata. The repository is the same one that I will be using for backups and currently runs at around 360GB for my backup.

I can't find from the documentation how much space is required on the repository for replica metadata? Can I assume it is roughly the same amount of space as a backup would take of the equivalent VM's?

[NightBird]

Is it possible to move metadata from one backup repository to another ?

Regards,
Boris

[Vitaliy S.]

Hi Boris, you can choose another backup repository in the replication job settings, and metadata files will be automatically re-created there on the next job run.