Failover and remote jobs

Hi

If a second server is running in standby ready to fail over, can it be used as a remote agent at the same time?

Cheers

No, it cannot be used as agent. It doesn't have the network listener running. In standby mode, it is like a one way street, standby to master.

Thanks :-)

Hello again

Also on failover, if the master scheduler fails over to the standby, what context and server do the jobs then run in for JAL action and pre and post actions? Especially if they are 'local' jobs?

The standby simply replicates job definitions and system settings. There is no separate context. In the fail-over event, the standby becomes a normal scheduler. While running jobs, it checks the value of the job "host" property. If that value is "Local" or blank, the job is run local. If for example, it is "ABC", the job is run on the ABC agent.

So I just need to replicate the supporting applications and scripts from the master scheduler to the standby box separately.

I guess if a job is in midflight then the standby isn't going to be able to continue it or anything.

Correct and Correct with a small note. Jobs having network file based dependencies might be restarted by the Standby in case if the dependency is not removed before the job start.

What sort of notification can be triggered if it fails over?

A message goes to the Windows event log. If you are on Windows 2003 or later, you can attach email notification or SNMP trap to this event. If you are using some earlier Windows version, you can add 24x7 based event monitor to watch for this event in the log. In the latest case you can do any wild stuff - email, batch processes, SNMP, etc..., for example, attempt to restart the mater remotely and push the processing back to the master.

The event log on the master or standby?

Standby

Cool thank you

The job db gets replicated from the master to the standby every so often right? What about the settings ie email details?

What you see in Tools/Options is not replicated, this is not considered as a dynamic data that changes after the initial setup. Everything else is replicated including jobs, queues, agent profiles, database profiles, script library and internal security settings.

What are my options with failover for when the original master scheduler comes back up? I would prefer to have scheduler #1 as the master whenever it is available with scheduler #2 only taking over when the other is not available.