kswapd0 cpu usage

[mbranyon]

Issue Report

Bug

CoreOS Version

$ cat /etc/os-release
NAME=CoreOS
ID=coreos
VERSION=1068.2.0
VERSION_ID=1068.2.0
BUILD_ID=2016-06-17-1616
PRETTY_NAME="CoreOS 1068.2.0 (MoreOS)"
ANSI_COLOR="1;32"
HOME_URL="https://coreos.com/"
BUG_REPORT_URL="https://github.com/coreos/bugs/issues"

...
BUG_REPORT_URL="https://github.com/coreos/bugs/issues"

Environment

What hardware/cloud provider/hypervisor is being used to run CoreOS?

ec2/kubernetes

Expected Behavior

kubernetes minions to run containers with out crashing and burning.

Actual Behavior

kswapd0 cpu usage skyrockets until the machine becomes unresponsive. Eventually, it will fail ec2 status checks, and be terminated.

Reproduction Steps

1. This happens when running about 50 active pods on a c3.xlarge minion

[crawford]

Upstream issue: https://bugzilla.kernel.org/show_bug.cgi?id=65201.

[skippy]

@mbranyon I'm running into this issue as well... I found that if my /run directory, which uses tmpfs, gets large relative to the mem size of the box, this process starts to spin. I create lots of short-lived docker containers, and it looked to be related to: #1081. I'm running coreOS 899.15.0, but have just upgraded to the latest stable branch and will be curious if I have the same issue. Which version of coreOS are you running, btw?