Is it possible currently or in future versions, if there are x many errors on x different machines then stop the job. This would allow the odd machine with issues to report a load of errors and the job wouldn’t stop but if loads of machines had the same issue the job would stop.
Hey Dave,
What you’ve described is essentially “Bad Slave Detection”:
thinkboxsoftware.com/deadlin … detection/
For example, if you have the “Mark a Slave as Bad…” setting set to 5, then if a Slave produces 5 consecutive errors on the same job, it will no longer pick up that job. So a machine with an issue with a particular job won’t bring that job down, and can move on to other jobs.
Cheers,
- Ryan
But what if say more than 10 bad slaves get detected? Can the job be marked as a bad job?
No, but if you configure the failure settings to mark a job as failed after 50 errors (for example), you’re essentially getting the same results.
Cheers,
- Ryan