NMM 3.0.1 restore GLR Exchange using DataDomain very poor performance

Hello joe030666

Can you please provide us with some more details?

For example:

- Could you please describe in more detail what is the performance you actually get (times to complete restore, time to mount, time to browse, etc.)?

- Does it happen with all Exchange DB's?

- What is the size and number of logs in the DB?

- How many users per DB?

Did you have the chance to check previous posts related to this same topic?

Do you have a case open with EMC Support?

Thank you,

Carlos

Hi Carlos,

if i use to run GLR of DB less then 250 gb its fine and performance are good, if i run GLR with DB more bigger of 250 GB performance of restore session is 4mb/s

I've opened an SR and EMC support said to me :

Well we recommend GLR restore to customer only when the size of the DB is less than 250 GB if the size of the DB is more then doing a GLR restore will have poor performance which is expected."

"So in cases of huge DB restore we suggest a RDB."

"So since the size of the DB is more suggest you to perform a RDB restore rather then GLR."

Hi Joe,

I agree that's by default the recommendation due to some performance issues seen in the past, however in some and after deeper investigations, support managed to get better results.

My personal take on this one? I'd go for the RDB, as I assume the performance you will find with RDB is much better that GLR.

If you want you can send me the SR#, but I'm afraid I'll be able to do pretty much nothing on the case side, but I would like to review the case notes, if that's OK with you.

Could you please let me know the NMM build and also the NW server and client version and build? If you are using storage node to backup the Exchange nodes I would need to know as well the NW client version on the SN.

Thank you,

Carlos

Hi Carlos,

many thanks for your answer,  about my environment:

NMM 3.0.1 build 271

Networker Server NetWorker 8.1.1.1.Build.271 Network Edition/25

Client Networker 8.1.1.1.Build.271

No storagenode

DataDomain 860  share cifs  mounted in Networker like AFTD device

About Case I think that it’s closed but in any case SR#68566828

Many thanks in advance

Hi Joe,

I'm wondering about that NMM build, as for NMM 3.0 SP1 the GA version should be 280.

On the server and client versions, 8.1.1.1 is not really the best to have in place now, as that was the first DA version so I assume you are using "basic" DA versions.

My first advice would be for you to go for the latest versions, for example:

NW server and client: 8.1.1.9 Build 360

NMM Version 3.0.1 Build 280

This is going to be a good move for you to avoid falling into any of the known issues of the DA version besides, some of the escalations I've worked on, related to this issue, ended up by having a fix on nsrmmd, which is the process reading from the DD device so I suggest you to upgrade the environment and give it another go.

In the meantime, if you have no restrictions with the free storage on the Exchange nodes, I would suggest you to go for the RDB as an interim solution

Thank you,

Carlos