2 Media servers and CASO - what if one crashes

[Julien_Lux]

Hello,

I run BE2010 with 2 servers. Server A is the CASO one and is connected to dedup storage A. Server B is just a MMS and connected to dedup storage B. I run all jobs followed by a duplicate. If backup job runs on Server A / Dedup A, then there is duplicate job to put data as well to Dedup B. And vice versa. This way, the 2 dedup storages are synchronized and contains the same data. The catalog for both servers is shared.

As a test, I did shutdown Server B only ( MMS one ) and from Server A ( CASO )I was able to restore any of the jobs. Not only the ones done on Server A, but also the ones done on Server B and duplicated to Server A. In short, the catalog contained everything.

As a second test, I did shutdown Server A only ( CASO one ). And from Server B, the catalog contains only the jobs which have run form this server B, and not the ones run from Server A and duplicated to Server B.

So, I am wondering, how to see the complete catalog ? The whole data is for sureon Dedup B

I was about to do an inventory then a catalog of Dedup B, but it is a 9TB disk and it will last for hours...

Any help is more than wlecomed.

Thanks.

Julien.

[CraigV]

Hi,

I take it you have Server B's device and media information centralised on Server A? If not, this would explain why you can't see the information from Server A. Server A controls Server B, and no information is stored on Server B.

That said, if you did so and Server A died, all jobs would effectively stop on Server B until you got the CASO running again. You'd have to hack the CASO information out of the registry before trying to run jobs on Server B again.

Thanks!

[Julien_Lux]

Hi,

That's correct, on ServerB, On Media Server Properties, Tab Advanced, Device and Media database location is : Central Admin Server, that is to say Server A.

But is there a different setup ? This is greyed out. And I can only change the catalog location which is currently set to replicated. 

So, how can I have Server B fully operated in case I loose server A. I am not sure understand what you mean by "You'd have to hack the CASO information out of the registry before trying to run jobs on Server B again.".

Or maybe the whole setup is not the best one. Server/Dedup disk A and Server/Dedup disk  B are on different sites actually. They run backup in parallel and then there is a cross duplication so that the 2 storages contains the same data. And the goal is to be able to restore everything from one site A or B easily. If site/server/dedup B  is down, that's fine. But if site/server/dedup A is down, I miss half of the data in the catalog from server B even if I am sure that dedup disk B contains everything.

Thanks a lot for more details.

Julien.

[CraigV · Accepted solution]

...you can look into replicating the details between the 2 servers as you mentioned.

By hacking out the CASO server's details on Server B I mean you've got to do a search in the Symantec key under HKEY_LOCAL_Machine\SOFTWARE and replace it with the name of Server B which should allow you to resume local backups on Server B.

If you had a spare server on the site where Server A is located, you could cluster the server offering an extra level of hardware/OS redundancy.

[Julien_Lux]

Ok, thanks.

I do not like to change the keys, but I will give it a try.

[CraigV]

ONLY change the keys if your CASO is down and you need backups to run...