Project

General

Profile

Actions

Bug #5642

closed

[Crunch] subprocesses get SIGKILL when they exceed Docker memory limit on compute nodes

Added by Brett Smith over 9 years ago. Updated over 9 years ago.

Status:
Resolved
Priority:
Normal
Assigned To:
Category:
Crunch
Target version:
Story points:
0.5

Description

You can reproduce this outside Crunch, using any Docker image:

compute2.qr1hi:/home/brettadmin# docker run -ti --memory=4M <IMAGE> bash
crunch@e5291ea9db1c:/$ dash
$ x=x
$ seq 1 10000000 | while read n; do echo $n; x=$x$x; done
1
2
…
21
22
Killed
$ echo $?
137

We expected code to get access to swap after exceeding the memory limit, but that's clearly not happening on the compute nodes. One possible improvement would be to upgrade Docker. #5425-10 suggests that is observed behavior from older Docker, and that more recent Docker is better about this. Our compute nodes are standardized on just-pre-1.2.


Subtasks 1 (0 open1 closed)

Task #5664: Review 5642-crunch-job-swap-limits-wipResolvedPeter Amstutz04/03/2015Actions
Actions

Also available in: Atom PDF