We encountered an issue on our Master DS server that caused the eDir
database to become unresponsive during SSI backup. When this happened
we quickly had reports from all across campus that users were no longer
able to log in despite the fact that we had other replicas of all
partitions available. The problem seems to be that once the client got
the list of replicas from the DA it used ping to "cost" the connections
and choose a server. Because the master server was still responding to
ping many clients continued to attempt to authenticate to the failed
server. It would, after a very long timeout period the client would
timeout with an error and the user would not be able to authenticate.
While we believe we've fixed the server issue with FTF2 for eDir 8.8.2 I
was wondering if there were any client settings that could mitigate
the issue if something similar happened in the future. Would any of the
settings below help?

> 1) Bad Address/Bad Server name: Will these settings actually apply
> to this situation since the client is able to contact (just not
> connect to) the server?
> 2) There is a Name Resolution timeout but again, if the server is
> ping-able, I'm not sure this would apply to this situation.
> 3) SLP Resolve Nearest Server: This seems promising? If this
> causes the client to look at more than ping it might actually notice
> that eDir is not responding before attempting to connect. Would
> enabling this setting indeed do that?