autolanddb (postgres) container hangs on exit, can not be killed

NEW
Unassigned

Status

MozReview
General
2 years ago
2 years ago

People

(Reporter: dminor, Unassigned)

Tracking

Details

(Reporter)

Description

2 years ago
After a test run I sometimes end up with one or more autolanddb containers that can not be stopped or killed through docker. The postgres process within runs at 100% CPU and can not be killed through kill. The only solution is to hard reboot my laptop, as these unkillable processes prevent normal shutdown. Since they run at 100% CPU, the processes eventually accumulate and make my system unusable, at least for running the mozreview tests.

This problem really sounds like what is described at [1] - a problem with recent linux kernels and aufs, however applying the patch at [2] and rebuilding the aufs module does not seem to have fixed things, so maybe not.

[1] https://github.com/boot2docker/boot2docker/pull/1113
[2] https://github.com/sfjro/aufs4-linux/commit/5e439ff30c92143d9a9ee3401a84e34c9852533b

Comment 1

2 years ago
I get failures stopping the bmodb mysql containers on boot2docker. Although not with 100% CPU AFAIK. I have to restart the vm to get them to clear.

I didn't have this problem with older versions of boot2docker / Docker. Sadness.
(Assignee)

Updated

2 years ago
Product: Developer Services → MozReview
You need to log in before you can comment on or make changes to this bug.