I have successfully tested failover between 2 seperate computers (running in GUI mode). However when I setup 24x7 to run as a service under the domain administrator account it does not 'failover' correctly. Also, I would like for the service to failover between machines as they are rebooted (master fail to backup, and backup able to fail to master).
Is there a seperate configuration needed for this?
Thanks

What do you mean by "it does not 'failover' correctly?" If you get an error, what is the errror text?

You can control when the failover takes place by changing number of synchronization attempts and timing. Please note that failover is one way and it is don so on purpose. This is because you most likely need to review first what failed, what needs to be rerun and what is the best way to do that before you switch back to the old master.

: I have successfully tested failover between 2 seperate computers (running in
: GUI mode). However when I setup 24x7 to run as a service under the domain
: administrator account it does not 'failover' correctly. Also, I would like
: for the service to failover between machines as they are rebooted (master
: fail to backup, and backup able to fail to master).
: Is there a seperate configuration needed for this?
: Thanks

: What do you mean by "it does not 'failover' correctly?" If you get
: an error, what is the errror text?

: You can control when the failover takes place by changing number of
: synchronization attempts and timing. Please note that failover is one way
: and it is don so on purpose. This is because you most likely need to
: review first what failed, what needs to be rerun and what is the best way
: to do that before you switch back to the old master.

I mean that when monitoring (remotely) the 'master' and 'backup' nodes - when the service is stopped on the master, I do not see the backup start? Due to the app running as a service (under a domain admin account) how can I successfully test a failover scenario (with remote access)?
Thanks

The standby service must be running all the time. Don't expect this service to somehow start magicaly when you stop the other service. The failover is not a Windows NT service level failover, it is a functional failover that requires 2 active instances. When the primary instance goes down the secondary (standby) instance takes over job running process.

: I mean that when monitoring (remotely) the 'master' and 'backup' nodes - when
: the service is stopped on the master, I do not see the backup start? Due
: to the app running as a service (under a domain admin account) how can I
: successfully test a failover scenario (with remote access)?
: Thanks

: The standby service must be running all the time. Don't expect this service
: to somehow start magicaly when you stop the other service. The failover is
: not a Windows NT service level failover, it is a functional failover that
: requires 2 active instances. When the primary instance goes down the
: secondary (standby) instance takes over job running process.

Understood, I guess I am having a difficult time testing/seeing this due to the fact that I now have it running as a service and have to connect to the servers via the 'remote client.' Both instances are running, I am simply trying to test the service failover. In other words, If I shutdown one machine validate the other took over. I could 'watch' this happen when I was running it through the GUI but with it set to run as a service I don't see where you can monitor this. Does that make more sense?
Thanks

Here is a tip, when GUI is not available, either lookup last records in schedule.log or enable HTML reports options and use these reports to figure out if failover happens and all other things.

: Understood, I guess I am having a difficult time testing/seeing this due to
: the fact that I now have it running as a service and have to connect to
: the servers via the 'remote client.' Both instances are running, I am
: simply trying to test the service failover. In other words, If I shutdown
: one machine validate the other took over. I could 'watch' this happen when
: I was running it through the GUI but with it set to run as a service I
: don't see where you can monitor this. Does that make more sense?
: Thanks