03-06-2014 11:50 PM
The manual is a bit vague on this. It says you shouldn't set this to "Backup media server" (the default) if you want the policy validated. It says you must set it to an usable media server. Which is best once the policy has been validated
Solved! Go to Solution.
03-07-2014 01:04 AM
Backup Media Server in this case means one that is listed as a VMware Access Host in the Master Servers Host Properties
If you need to pin the policy down to using just one Media Server due to the storage unit used then select that media server
The ideal behind having the new feature of Backup Media Server was to allow you to use a Storage Unit Group in the policy so that you have a level of resilience and failover for the VMware policies
Hope that makes sense
03-07-2014 01:04 AM
Backup Media Server in this case means one that is listed as a VMware Access Host in the Master Servers Host Properties
If you need to pin the policy down to using just one Media Server due to the storage unit used then select that media server
The ideal behind having the new feature of Backup Media Server was to allow you to use a Storage Unit Group in the policy so that you have a level of resilience and failover for the VMware policies
Hope that makes sense
03-07-2014 04:45 PM
Thanks, that made sense because that what I attempted. All the media servers (7) have VMware access. I have "prioritized" storage unit groups and thought it would be best to change them from the primary media server to "Backup Media Server" then they started failing. They did not all fail at once but did it over time, once failed they never came right.
The error codes were related names resolution : 58 Server unable to locate client & 48 A gethostname failure
When the VMware backup host was set back to a specific media server they started working again.
Its a bit weird.
03-07-2014 10:35 PM
Have you double-checked that all media servers in the STUG can resolve VM client hostnames?
03-08-2014 12:34 PM
They all have the same DNS servers and nothing has changed in the environment that area (a lot of other people would be complianing as well). The problem was specific to a small number of policies (approx 5 in about 300) where that setting was changed. Each policy had about 16 clients with only four jobs can run at any one time.
Where I can use the default "VMware hostname" as the selection criteria to validate DNS. Its my understanding that if you use "VMware hostname" it will do a reverse lookup of the vCenter IP address of the client. vCenter does not know anything about the property "VM Hostname".
Because the problem is fixed by using a specific media server hostname this tells me DNS is working with no timeouts. The TCP backup traffic can block or slow down the UDP packets of DNS when the network get maxed. If that was occurring then it would have been seen in other policies.
Copied a policy to see if was corrupt and created a new policy for some of the same machines with the same settings. Both attempts failed with same errors