Pre-seeding data in Avamar

Copying the data from the client system(s) to an external device, shipping it to the destination site, and backing it up to the Avamar server at the destination site from any client will write the chunks to the storage so they don't have to be sent across the wire. You should see a decent performance boost from doing that but you won't be able to take advantage of the cache files because the file paths would be different.

There would likely be a benefit to doing this in a pure Avamar configuration, especially if bandwidth is limited. However, if the data is being written to a Data Domain back-end, you probably wouldn't see much benefit from this approach because of the way Data Domain handles the data that defines the structure of the files.

Thanks Ian,

Were running data domain over this way, with regards to the seeding just to confirm we can ingest the data from any client as a one time backup and then once I hook up the client agent it will definitely know the data is related to this client?

I don't know how much benefit you'll get by pre-seeding if you're using Data Domain storage. It might not be much.

As long as a chunk of file content data has the same hash, we don't care what client it came from. If the atomic chunk (Avamar) or L0 segment (Data Domain) is already present on the storage, the client won't send it again.

The reason you might not get as big a benefit from pre-seeding if you're using DD is that DD intentionally salts higher level hashes (L1 segments, etc.) which forces the client to run fingerprint lookups for all the L0 segments. This is done to improve the spatial locality of the data which is important to DD for performance reasons.

Still, pre-seeding can't really hurt so you don't really have anything to lose.