I'm installing an Xsan attached Episode Engine 6 cluster and want it operate much like EE5, so all the submitted workflows are Monitor jobs. I want these jobs to stay "bolted down". The EE6 interface is now the same as the desktop and submitted jobs seem much more vulnerable, for lack of a better word.
I need a way of submitting, monitoring and modifying these jobs in place. It's unclear how I modify a submitted monitor job. Will copying the submitted job... modifying and saving all Tasks/Workflow change the submitted job? It appears it does not. It also seems using the stop button at the bottom stops the job for good requiring a new submit. Resubmitting every time I make a change is something I want to avoid.
I also would like to have all the tasks and workflows exist in one place much like they did in the EE5 controller node. I want to be able to power down the cluster... bring it up and have the cluster back with all jobs intact and operating. The "keep history forever" pref seems to do this but is it wise to rely on this to keep all the jobs "bolted down"? Where is this history stored so I can back it up?
I understand one can use an unlicensed, non-Xsan attached machine as a non-encoding Master. Has anyone tried this? I've been using a licensed master as the place where I create and submit jobs, but if the app has problems enough that a restart is required then that brings down the cluster. My plan is to put up another unlicensed machine to create tasks and submit "bolted down" jobs from.
I don't know of any way to modify a running workflow. I understand your concern certainly, but at present I don't see any way to do this. To modify a workflow you have to stop it, modify it, then submit it.
From my experience, each time you reboot your cluster, the master always starts up (assuming you've set this preference) and the workflows will continue monitoring the same watch folders. Additionally you could easily batch submit all the workflows to the cluster via command line (it's not complicated) and all the workflows will start.
An unlicensed Mac as the Master is how we're using it, so yes it can be done. However, if your present Master is on Xsan, I'd highly suggest that your backup Mac is also on Xsan as you'd want each of them to have the same paths to your watch folders and output.
Thanks for the info. I assume your EE6 is Xsan attached? If so have you made any changes to the Node.xml to keep traffic off the MDC network or to hard code connections to avoid Bonjour? I've gotten a few instructions from Telestream on how to avoid the MDC net but don't they aren't working.
Any Mac could take over as the "master", it would just need to have a copy of the workflows. It wouldn't function as an automatic failover however. Perhaps Kevin knows of a way to do this that I'm unaware of.
Yes our master is presently SAN attached. Since the master is monitoring your input folders and output folders, it simply needs to have the same access to these folders that your nodes would have.