Video Screencast Help

CASO and optimized dedupe between sites for DR

Created: 14 Dec 2010 | 2 comments

Primary objective is to have backups duplicated to the secondary site in dedupe(opt) format.

Problem is what happens when the link is down between the two sites?  
Surely my ability to restore on the Primary site is gone, as are backups?  

To get around the link issue in a DR situation I have got the Primary CASO on the secondary site, so assuming the Primary site blows I can still do all my DR restores on the secondary site as it hosts the Primary CASO.   But what if the link is only down for a day, no production backups on Primary site will work as it can't reach primary CASO

If I do distributed logs I can't do optimized dedupe unless I use SSO which has similar problem of needing the primary available for the secondary to function

The setup is:

Primary site:
Managed media server with Dedupe Option

Secondary site:
Central media server with Dedupe Option

 

I don't want to do IDR to solve this problem because to do a DR test I have to go through the whole IDR process which seems like a sledge hammer approach.

 

Seems that optimized dedupe from primary to secondary site, with resilience for link failure should be a common scenario.  

Comments 2 CommentsJump to latest comment

Colin Weaver's picture

To be honest your problem is not a DeDup problem as you are basically asking how to make an MMS still run a backup job if the link to the CAS fails - obviously your Opt Duplication would not work anyway as the link is down.

If you link is that unreliable that you need to cover for more than 1 missed backup because of a link failure then you will probably need to make sure you know how to make an MMS server operate as stand alone server in an emergency - or make sure you have redundancy on your WAN link so that the failure auto switches to a different path

Your other option of course would be to put a CAS server in both sites - with both sites having an MMS server linked to the CAS at the other site. If you configure local DeDUp backup jobs to each MMS followed by Opt DeDup from the MMS servers to the remote CAS servers - in the event of a link failure you can have some jobs on hold on each CAS that can do direct backups of local data if you know the link has failed - again this is a bit of a sledgehammer approach but you seem to be asking about what is hopefully a rare failure condition and if you really need to consider long term options to protect yourself for an extended failure then maybe you need a slegehammer approach (personnally I'd investigate the option for redudancy of your WAN links)