I would like to configure High Availability for Media Agents.
I thought about a shared disk (FC) between two media agent, that way it will still restore in case one of them fails; But I didn’t understand what a SAN Data Server is.
Then I read about GridStor, not sure if it is the same thing.
My question is, what is the best way to assure high availability for my Media Agents? Thanks in advance for your valuable help.
Best Regards
Best answer by Jos Meijer
When you use LUN's which are assigned to 2 media agents in a non-clustered environment you should use mount path sharing as both OS cannot claim the file system at the same time.
Depending on the OS you will get optimal results using:
Transport type Data Server - IP for linux media agents
Transport type regular and point to an UNC path for windows media agents
This way you will still have a dependency on the media agent which has the LUN assigned to the OS. If the OS is down, access to the LUN is not possible.
Therefore you might want to shift your method towards a SAN Data Server setup which allows different media agents to access the same LUN's simultaneously, keep use of global dedup, allows FC connectivity and removes the dependency on specific Media Agent to access the LUN:
Actually, I got a bit lost. The GridStor allows changing the path for a backup in case a Media fails.
What I am looking for, is how to make sure that both Media Agents have access to the same LUN without corrupting the data: I guess this is a prerequisite for the grid store. Or maybe it shouldn’t have access to the same LUN, I don’t know.
When you use LUN's which are assigned to 2 media agents in a non-clustered environment you should use mount path sharing as both OS cannot claim the file system at the same time.
Depending on the OS you will get optimal results using:
Transport type Data Server - IP for linux media agents
Transport type regular and point to an UNC path for windows media agents
This way you will still have a dependency on the media agent which has the LUN assigned to the OS. If the OS is down, access to the LUN is not possible.
Therefore you might want to shift your method towards a SAN Data Server setup which allows different media agents to access the same LUN's simultaneously, keep use of global dedup, allows FC connectivity and removes the dependency on specific Media Agent to access the LUN:
At the first position, it is the backup; However, I still think that we must be able to restore at any moment even if one Media fails. So I would say both.
Based on your requirements this would be an affective method to provide a shared storage location via FC to multiple media agents and ensuring HA for both backup and recovery.
Other methods will require IP based solutions or will lack HA.
We use 3 different kinds of cookies. You can choose which cookies you want to accept. We need basic cookies to make this site work, therefore these are the minimum you can select. Learn more about our cookies.