We have some scripts that artists can trigger that can adjust priorities for high priority jobs via some wrangling oversight. Since switching to deadline7, it seems that some of the changes applied by the scripts are transient, and after a while the jobs ‘reset’ to their previous priority settings. We have no tool that automatically adjusts priorities per se, so we don’t know what could cause this. There are no manual priority adjustment entries in the history.

We have a script that auto adjusts task timeouts based on some custom criteria, but edited its using the command line too via -ExecuteScript (because deleteJobReports is not accessible via restAPI).

2015/05/10 16:10:45 avi.joshi LAPRO3303 (SCANLINEVFXLA\abhishek.joshi): Modified job priority to Tier 1
2015/05/11 05:00:20 root lascript0002.scanlinevfxla.com (lascript0002.scanlinevfxla.com\root): [Pro Optimizer] Task timeouts detected. Adding 2 hours to job task timeout
2015/05/11 05:00:20 root lascript0002.scanlinevfxla.com (lascript0002.scanlinevfxla.com\root): [Pro Optimizer] Task timeout reports will be deleted
2015/05/11 08:00:25 root lascript0002.scanlinevfxla.com (lascript0002.scanlinevfxla.com\root): [Pro Optimizer] Task timeouts detected. Adding 2 hours to job task timeout
2015/05/11 08:00:25 root lascript0002.scanlinevfxla.com (lascript0002.scanlinevfxla.com\root): [Pro Optimizer] Task timeout reports will be deleted
2015/05/11 09:48:33 laszlo.sebo LAPRO3067 (SCANLINEVFXLA\laszlo.sebo): Viewed History Log

In the morning , the priority was the default again.

The issue has been reported by a couple of people to us internally.

Did anybody else report such behavior?
Would it be possible for the restAPI/command line execute to mark history entries every time changes are applied through it?

(i’ve edited the post… after talking to the developer who wrote the pro-optimizer script, turns out, its not using the restAPI. So i dont think this is related to a script side caching of obsolete job values)

There is a REST command to log job history entries, so your script can use that to log what it’s doing:
docs.thinkboxsoftware.com/produc … itory.html

Cheers,
Ryan

Yes, and we are using that as much as possible. But i think it might be better handled at the API (rest/commandline) level, that would ensure that there is absolutely no way to change a job/slave etc, without a log entry showing up.

We’ll add this request to the wishlist.

Cheers,
Ryan