How to go about implementing Per VM Backups

[ricenhower]

I'm really excited about using some of the new features in v9 , especially the per vm backup, but I'm not sure what impact it would have in my environment. Below are some of the concerns;

Environment is pretty simple, I have 2 proxy servers for hot-adds, 1 proxy for direct SAN and tape transport. I have 3 repositories, 1 is the primary landing zone and the other 2 or backup copy repositories for GFS and a longer retention set.

1. If I want to enable per vm backups I have to do this on the repository, so this would change every backup chain after the 1st full backup on this repository?
2. If for whatever reason I don't like this option how hard is it to revert back?
3. If the primary repository has per vm backups enabled will this effect the backup copy chains stored on repositories dedicated for backup copies? Is there any additional configuration I'd have to do for the backup copy repositories?
4. Does the concurrent connections on the proxies need to changed in order to use per vm backups?

Lastly, I'm having a hard time understanding what needs to done to get the parallel data processing to flow properly to really optimize this new feature...Could someone help explain this? I was also wondering what others are using to help determine the proper throttling for network traffic if any?

Respectfully

[foggy]

1. Correct.
2. The procedure is the same, you need to disable the option on the repository and perform active full for each job pointing to it.
3. Since it is a repository setting, there will be no effect on the secondary repository.
4. Depending on the actual storage, you might want to adjust the max number of tasks assigned to proxy and repository servers after enabling the per-VM backup chains setting, based on the bottleneck stats for the first job runs.