Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

713

March 24th, 2017 05:00

Grid or redundancy possible with what I was sold?

As a new comer to Avamar, I'm still stumbling through configuration of the equipment sold to us.

We had existing DD2500 units at Site A and Site B, replicating CIFS.

We were sold the following:

Two Avamar G4S M1200 STORAGE Nodes

BRM for Avamar

The Avamar nodes were to leverage their internal storage as well as the DD2500 as a target. The sales pitch was that M1200 at Site A will replicate all local storage backups AND meta data to M1200 at Site B. Site A DD2500 data would be replicated to Site B DD2500. If Site A is destroyed, Site B is fully functional since jobs and meta data were replicated, along with the DD2500 data replicated.

At this point, it may be important to note that the original EMC sales team left EMC shortly after purchase. The reseller has no working knowledge of Avamar systems, so they were of no help.

As of today:

I have successful data protection of servers in Site A utilizing both Ava storage and DD storage.

I have successful replication of Site A DD backups to Site B DD backups, managed by Site A Storage node.

I've no need to replicate Site A Storage node backups, so I have not set up Site B Storage node as target.

Site B Storage node sits idle.

My question is this: Can I make the M1200 Storage Nodes perform as promised? I cannot find any documentation on this scenario. Only that if I wish to do a multi node grid, I must have Utility nodes plus a minimum number of Storage nodes.

Any real world advice or experience is appreciated.

2K Posts

March 31st, 2017 07:00

Using native Data Domain replication with Avamar is not supported. In order to protect the metadata stored on the Avamar node, Avamar / Data Domain configurations need to be set up to replicate through Avamar.

If the Avamar on Site B has the Avamar Server software installed and ready to go, all you need to do to get Avamar replication up and running is integrate the Site B Data Domain (if it's not already) and configure a replication group on Avamar A that replicates the clients you want replicated. The Avamar Administrator Server data is replicated automatically as part of every replication group. If Site A went down, you could then restore affected clients from the Avamar on Site B if necessary. As I mentioned above, the groups, schedules, etc. will be protected but will not be usable on Site B.

If you don't have checkpoint backups enabled and Avamar B isn't ready to go, I'd strongly recommend that you enable checkpoint backup immediately while you get Site B sorted out. If you have neither checkpoint backups nor Avamar / DD managed file replication configured, a loss of the Avamar server on Site A will result in total loss of all backup data.

Edit: When I say "the clients you want replicated", I'm referring to the clients that are backing up to Data Domain since you mentioned that you don't want to replicate any of the clients that are backing up to Avamar storage.

2K Posts

March 28th, 2017 06:00

In Avamar / Data Domain configurations with single node Avamar systems, one of two protection mechanisms is normally used to cover the DR case. Either the systems will be configured as a replication pair (as you've described above) or the Avamar will be configured to perform checkpoint backups to the DD which will protect the Avamar metadata.

If Site A is destroyed, Site B is fully functional since jobs and meta data were replicated, along with the DD2500 data replicated.

If Site A is destroyed, its data will be protected by replication. However, job configuration information (datasets, schedules, groups, etc.) is not in an immediately usable form and would have to be restored to the repaired or replaced replication source in order to be useful. To fail backups over to the DR site, you would need to recreate the datasets, schedules, and groups on the DR Avamar system.

I have successful data protection of servers in Site A utilizing both Ava storage and DD storage.

I have successful replication of Site A DD backups to Site B DD backups, managed by Site A Storage node.

If you have replication configured through Avamar, the Avamar metadata is being replicated from the single node system on Site A to the single node system on Site B at the same time that Avamar is directing the DD on Site A to replicate its data to the DD on Site B. Avamar / DD replication replicates Avamar-to-Avamar and Data Domain-to-Data Domain.

I'm not sure I understand the problem you're trying to solve. It sounds like the solution is working as intended, except that you're not replicating backups sent to Avamar storage. If you wish to replicate these backups (and I'm not sure you do based on your post), you can just add the clients that are backing up to Avamar storage to a new or existing replication group and they will be replicated to the target.

Could you clarify what's not working for you?

2 Posts

March 30th, 2017 12:00

Ian,

Thank you for the response. I'll try to explain further:

Production Site A:

Ava single node with 4TB local.

Site A DD2500 set as target within Site A Ava.

All backups stored on Site A DD2500 are configured to replicate to Site B DD2500(configured as replication target in Site A Ava) through Ava's Replication Groups.

Repl Group consists of all pertinent client's and their backups I want replicated.

All backups stored on Site A Ava node are not configured to replicate to Site B Ava. Ava node storage is used for short term retention only(as I currently have it configured). I had no intentions of replicated this data to DR Site B.

DR Site B:

Ava single node with 4TB local.

Site B DD2500 is NOT set as target for Site B Ava. It is the replication target, as mentioned, for Site A DD2500 replication.

Site B Ava single node sits idle at this point, with no configuration.

I do not wish to replicate Site A Ava backup storage to Site B, unless it is a requirement to ensure that Site B Ava has all metadata, datasets, schedule, groups, etc. If this is the only way to ensure that Site A single node will replicate this data for DR Site B recovery with Site A Ava fails, then I'll reconsider job placement, etc.

While setting up job replication is very straight-forward and working properly for the DD to DD replication. I do not find documentation, nor configuration options, for establishing the replication of the Single Node "management data", such as meta, datasets, schedule etc. Perhaps I'm just looking for too easy of an answer such as "Establish Avamar Node Replication for Recovery" button/wizard.

I hope that clarifies where I'm stumped. And if there is documentation on how to achieve this, I'll be happy to review it before asking more questions. I feel I've been thorough in looking for answers in a single node to single node replication and I can't find what I'm looking for.

Thanks again,

JZ

No Events found!

Top