DFSR: What happens when there is a long network outage.

The question:

Do you know of some official Microsoft documentation that describes how surviving DFS-R group members behave if one member is offline long enough that the change cache becomes full?

Background:

In a recent change review board meeting I was asked what happens with DFS-R if a server is offline for a long period of time.

I know that changes are cached and updates are exchanged when the member rejoins the group, but in these modern times with outsourced administration and cloud hosting, we get a whole world of interesting horrors that must be considered.

Such as

  • We replicate to a cloud server, then someone closes the cloud account without disabling DFS-R, so the server disappears. 

I know, I know - there will be errors reported in the event logs that someone should notice, but lets assume that this server is somehow not being monitored. 

In the case mentioned above I expect that:

  • The remaining DFS-R servers will cache changes until the cache fills;
  • The remaining servers declare that replication to the lost server is now broken; 
  • If the lost server reappears, replication must be re-established from scratch;
  • The surviving group members continue to replicate and operate normally.

In particular, I do not want the remaining server to stop working in some way because its buddy cannot be contacted.

I cannot find any posting or documentation that describes such a scenario.

While I personally value your experience, to satisfy the change board I need a link to some Microsoft documentation.

Thank you for your help.



June 18th, 2015 10:33pm

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics