AWS Thinkbox Discussion Forums

Slave idle for too long due to Launcher not running anymore (crashed?)

Hi,

i’ve been observing this for a while now - roughly since 7.1.0.35 (but i can’t tell for sure) and it’s still happening with 7.2.0.10 on Windows 7.

Sometimes we have Slaves being idle for much too long (many hours) despite the idle time being set to 5 minutes. When checking those machines it’s always the deadlinelauncher process that doesn’t exist (anymore) on the machine. So i’m assuming it for some reason closed/crashed. So far this seems pretty random to me and i’m not sure where to look for any hints to what might be going on. There’s nothing to be found in the logs of the Launcher. Can this kind of unexpected crash/quit be found in any Windows logs? And is there maybe some workaround to make sure the launcher process is being restarted when it crashed?

Cheers,
Holger

Hey Holger,

On Windows specifically, when Deadline hard-crashes, there can be sometimes be useful info in Windows’ Event Viewer. It might not always be helpful, but can sometimes at least point you in the right direction. How long do the Launchers normally run before crashing out? Or does it seem pretty random?

As for ensuring that the Launcher re-starts when it crashes, we don’t really have a built-in solution for this. Some of our clients set up a cron job (Linux/mac) or a Scheduled Task (Windows) to periodically run a script that makes sure the Launcher is running (and start it if it isn’t), which is probably the best way I can think of doing this.

Hi Jon,

i had a look at the Windows Event log but couldn’t find anything.
So i guess i need to cook something to check for the existence of deadlinelauncher.exe as you suggested. Just wanted to see if there’s any ‘official Deadline way’ of how to deal with this.

Cheers,
Holger

Privacy | Site terms | Cookie preferences