UNSOLVED

XRay1978

updated

16 years ago

X

XRay1978

1 Rookie

•

61 Posts

0

1937

October 6th, 2010 00:00

Avamar Data Transport sizing

I am looking into implementing ADT.

Some things I am not clear about:

1. how many 2 TB AVE Transport nodes do I need to cover say 6TB Avamar grid? 6:2=3? This means you need 9.6TB usable on your ESX (plus control node)?

2. by looking at setup doc you install Networker client on each Transport node, so from Networker point it is just a linux client really? Need extra licences for Networker I presume?

3. Restore process is just a matter of stopping gsan on relevant Transport node, restoring of whole datastore and then restoring particular file/dataset?

4. From what I see any backup product which can backup RedHat Linux should work right?

  • kishore4

    1 Rookie

    •

    21 Posts

    414

    0

    Posted October 13th, 2010 01:00

    Hi Raimonds,

    There is a sizing tool for ADT. Its not generally available on Powerlink so you will have to ask your EMC representative for it.

    ADT does roughly 1TB/24hrs per node. This is dedupe data. If you have only one transport node, obiously it will take longer as data will move in 2TB chunks but if you have say 4 transport node you can move 4 x 2TB in parallel provided you have dedicated tape drive to each transport node.

    Even if you just one transport node and you have more than 2TB of dedupe data to tape out, ADT will pull 2TB from Avamar server, pauses the replication, dumps the data to tape, cleans up transport node and resumes replication for next chunk for 2TB data. The process continues until all selected data is sent to tape.

    Sizing tool will also calculate and recommend number of Avamar domains required, no. of ESX hosts etc based on total dedupe data, frequecy of ADT job schedule, backup selection and time allocated for ADT.

    Hope this helps.

  • TimFeldmann1

    1 Rookie

    •

    12 Posts

    414

    0

    Posted October 27th, 2010 12:00

    The way you explain the process is not the way that it was explained or has been my experience.   The way you explain it would lead me to believe I can have a client/domain with 6tb of data and send it to ADT, where the process will handle this in 2TB increments.   My understanding is that each ADT job has to be limited to 2tb in size.  If the amount of data within that job exceeds this limit the job fails.    For example, I have a client that has 2.5tb of file system data.  Because ADT's smallest qualifier for a job is by domain and therefore this client, I can not effectively send this data to tape.  The job fails when the transport node fills up.