AWS Thinkbox Discussion Forums

Slave Status Inaccurate

In the Slaves panel in Monitor, machines that have been down (some for hours), still show that they are rendering or idle.

Deadline’s housecleaning feature should detect when slaves haven’t been updated for a while, and mark them as stalled. First, can you open your Repository Options and under Slave Settings, let me know what the “Number of minutes before an unresponsive Slave is marked as Stalled” is set to?

Thanks!

  • Ryan

Its set to 120

So that means it will take at least 2 hours before a slave is detected as stalled, and since you guys are running Pulse, it shouldn’t be much longer than that. Was it the case that these slaves hadn’t been running for a period longer than 2 hours?

Yes. There are some machines that have been down since last Thursday that are still showing status of “Rendering” or “Idle.”

Thanks for the info. In the slave list, what is shown for the Last Status Update for these machines? Is that timestamp also much older than 2 hours? That’s what Deadline would be using to compare against the current database time to determine if a slave has become stalled.

Interestingly enough, the Last Status Update for the machines that I know have been down are within the last few minutes. What could cause that?

Is there maybe a slave process running in the background on those machines? The only thing that should be updating that information is the slaves themselves…

Privacy | Site terms | Cookie preferences