Bug
Container Linux Version
Current nightly test runs, e.g.:
NAME="Container Linux by CoreOS"
ID=coreos
VERSION=1506.0.0+2017-08-25-1813
VERSION_ID=1506.0.0
BUILD_ID=2017-08-25-1813
PRETTY_NAME="Container Linux by CoreOS 1506.0.0+2017-08-25-1813 (Ladybug)"
ANSI_COLOR="38;5;75"
HOME_URL="https://coreos.com/"
BUG_REPORT_URL="https://issues.coreos.com"
COREOS_BOARD="amd64-usr"
Environment
Happens on qemu and AWS, probably all of them.
Expected Behavior
docker run works reliably
Actual Behavior
Every once in a while, the first run has the below error:
$ journalctl -u docker -o cat | head -n 12
Starting Docker Application Container Engine...
time="2017-08-29T01:38:38.034391242Z" level=warning msg="failed to rename /var/lib/docker/tmp for background deletion: rename /var/lib/docker/tmp /var/lib/docker/tmp-old: no such file or directory. Deleting synchronously"
time="2017-08-29T01:38:38.047270072Z" level=error msg="Failed to built-in GetDriver graph aufs /var/lib/docker"
time="2017-08-29T01:38:38.244953954Z" level=info msg="Graph migration to content-addressability took 0.00 seconds"
time="2017-08-29T01:38:38.248685288Z" level=info msg="Loading containers: start."
time="2017-08-29T01:38:38.416311592Z" level=info msg="Default bridge (docker0) is assigned with an IP address 172.17.0.0/16. Daemon option --bip can be used to set a preferred IP address"
time="2017-08-29T01:38:38.539319520Z" level=info msg="Loading containers: done."
time="2017-08-29T01:38:38.562007032Z" level=info msg="Daemon has completed initialization"
time="2017-08-29T01:38:38.562264173Z" level=info msg="Docker daemon" commit=874a737 graphdriver=overlay2 version=17.06.1-ce
Started Docker Application Container Engine.
time="2017-08-29T01:38:38.571944666Z" level=info msg="API listen on /var/run/docker.sock"
time="2017-08-29T01:38:39.756677347Z" level=error msg="Handler for POST /v1.30/containers/5b2752feaa6cff592771206bb4d5fe27a14bf5936bfdd167abc470648df25165/start returned error: error creating overlay mount to /var/lib/docker/overlay2/ef86febd7609b1508ef12b147c5aaf07ed750d09ca90b4d362b7e81543f33a3e/merged: device or resource busy"
Reproduction Steps
kola run "docker.base" and get really (un)lucky
Other info
We, unfortunately, bumped the kernel to 4.13 at around the same time we bumped docker to 17.06.
I expect it'll be one of those two changes.
This happens very infrequently. When it does happen, the impact is minimal; the first container run fails, but subsequent ones are fine. I think it can only happen as the first container run, so anything that's retrying appropriately should be fine.
Bug
Container Linux Version
Current nightly test runs, e.g.:
Environment
Happens on qemu and AWS, probably all of them.
Expected Behavior
docker runworks reliablyActual Behavior
Every once in a while, the first run has the below error:
Reproduction Steps
kola run "docker.base"and get really (un)luckyOther info
We, unfortunately, bumped the kernel to 4.13 at around the same time we bumped docker to 17.06.
I expect it'll be one of those two changes.
This happens very infrequently. When it does happen, the impact is minimal; the first container run fails, but subsequent ones are fine. I think it can only happen as the first container run, so anything that's retrying appropriately should be fine.