We noticed recently that the LDAP failover was pretty slow. I'm guessing it's controlled by ldap.connect.timeout system setting. Although the report was that it took "about 2 minutes" to failover and the setting is 50,000. That could be a subjective issue from the person complaining!
I'd also like to know if anyone can clarify some of the detail on the failovers.
What happens if we have 2 AD LDAP servers, the "primary" fails and Jive flips to the "backup" and then service is restored on the primary. Will Jive naturally switch back to the primary before the backup fails or wait for the backup to fail before trying the primary.
Assuming it doesn't return to primary when it is fixed, sometime later the backup goes down. Will it keep flipping from one to the other as one fails or does it mark a server as "bad" somewhere that would need clearing?
Experience so far suggests that it does just keep flipping from one to the other every time one fails, but I'd like to be a bit more sure.