Disaster Recovery / Restart tips for a Server Crash with Data Manager

[shawn1272]

Hi,
I've been tasked with developing a recovery plan for Data Manager in the case where there is a server crash.  The issue that I am struggling with is that in the case where there is a server crash, First, Data Manager is not going to know to restart itself, and Second, even if I were to use Windows Scheduler to restart Data Manager on a start up Data Manager is not going to know to restart only a previously running JobStream. 

So, here are the scenarios:
1.  Server Crash and Data Manager has no JobStream running.  I want nothing to happen.
2.  Server Crash and Data Manager has a JobStream running.  I want to restart the JobStream. 

Also, the JobStream is currently setup to restart at the earliest node that did not successfully complete. 

So, does anyone know how to work a server crash?  'Cause the Windows Scheduler doesn’t seem robust enough for the job.  Maybe a java script?  Any input or direction would be great!

Thanks,

Shawn

[UseCog]

some thoughts...

create a table for handling the rerunnability, means before and after every jobstream, keep an entry for that job stream in the table. similearly for the main job stream as well. So this will tell you that whether the any job is currently executing and where exactly it is crashed.

to handle the first scenario, check whether the main job stream is running while you start the server again. if not , then call the main jobstream ...this will navigate to the failed job stream and start from that point.

i hope this work fine.

[shawn1272]

I'll give it a shot.  Thanks for the help

Shawn

[UseCog]

Please keep us updated the status :)