Video Screencast Help

Dedupe pool scale out to 5020 appliances.

Created: 20 Apr 2013 | 4 comments
AB Chaudhry's picture

We have six appliances (5020) with 32 TB each for total dedupe capacity of 192TB without requirement of replication. We can combine these appliances in following proposed patterns

A) Single 192TB global dedupe pool combining all six appliances
B) Two dedupe pools with 96TB each with combining 3 appliances for each pool.

Which one is better in terms of:

  • Availabiliy (1+1)
  • MTBF
  • Backup Strategy (Planning, administration and management)
  • Impact on backup environment if an appliance goes offline.
Operating Systems:

Comments 4 CommentsJump to latest comment

Andrew Madsen's picture

How much data do you have to protect and how many medis servers do you have? WHat is your backup strategy? Are you looking for more resilience? Are you taking a test?

The above comments are not to be construed as an official stance of the company I work for; hell half the time they are not even an official stance for me.

AB Chaudhry's picture

In traditional model we have more than 40 Media Servers and roughly 60TB of data backup daily. Now we have induction of four 5220 and six 5020 appliances. Each 5220 will perform as independent FT Media server with its own Advanced Disk pool. We are combining six 5020 for one dedupe pool of 192TB and concern is:
What if an appliance 5020 will go down?
What is MTBF and MTTR of 5020? How often they go down and how long it takes them to be back operational?
Will it be better to split Single dedupe pool to two smaller dedupe pools for resilience?

Nicolai's picture

If you loose a 5020 and all 6 are in one pool you will have no access to data at all. Don't calculate in MTBF, the change of loosing a 5020 is one - It will happen at some time (Murphy law).

When it come to de-dupe data is really think having data in two copy's is paramount -I think the choice is easy. It's  B.

You will need SLP toperform the replication.

Assumption is the mother of all mess ups.

If this post answered your'e qustion -  Please mark as a soloution.

Mark_Solutions's picture

I am with Nicolai - loose one 5020 and the whole pool goes down - loose it forever and you loose everything!

For resiliency use two pools and when ever possible always have two copies of the data, either duplicate between them or to tape (just in case!)

For planning and administration one pool is easier, especially if you will be using AIR as you cannot put tow pools into a storage unit group when using air so would need to target half of your backups to one pool and half to the other.

So summary:

1) two pools is better for availablity - you have a second pool to re-point backups to if one goes down

2) n/a

3) One pool is much easier to manage , especially if using AIR

4) one down - all down!

So - I would accept the extra management cost and go for two pools

Of course you will also be able to store less data if you have 2 pools due to loss of global de-dupe - and that will make quite a big difference in the capacity of the system

Hope this helps

Authorised Symantec Consultant

Don't forget to "Mark as Solution" if someones advice has solved your issue - and please bring back the Thumbs Up!!.