In task T190185, we added a cleanup tool, tmpreaper, with more restrictive settings than used in the production module to web nodes due to temp files becoming enough of a problem to temporarily halt puppet runs and other tasks. This task is to track how that is doing at this point and apply the strategy on all execute nodes if it is appropriate.
Description
Description
Related Objects
Related Objects
- Mentioned Here
- T190185: tmp file leak from tools.wsexport
Event Timeline
Comment Actions
Since I just caught an exec node filling tmp with old files, I think that means I should go forward with this.
Comment Actions
So this notion will clearly not be a panacea. Continuous jobs that are malfunctioning are capable of restarting on loop (even worse if they are Big Brothered), but the cleanup might be welcome, nonetheless.
Comment Actions
I've come to the conclusion that this will likely just create unnecessary load and problems. Unless someone says otherwise, I'll leave that functionality only on web nodes where it is more useful and necessary.