Bug 27251 - Docker does not run get error cgroup mountpoint does not exist
Summary: Docker does not run get error cgroup mountpoint does not exist
Status: NEW
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: Cauldron
Hardware: All Linux
Priority: High major
Target Milestone: Mageia 8
Assignee: Bruno Cornec
QA Contact:
URL:
Whiteboard:
Keywords: Triaged
Depends on:
Blocks:
 
Reported: 2020-09-06 00:02 CEST by Mike Crecelius
Modified: 2021-01-06 12:56 CET (History)
5 users (show)

See Also:
Source RPM: docker-19.03.11-2.mga8.src.rpm systemd-246.4-1.mga8.src.rpm
CVE:
Status comment:


Attachments

Description Mike Crecelius 2020-09-06 00:02:29 CEST
Description of problem: Running Docker gives an error
docker: Error response from daemon: cgroups: cgroup mountpoint does not exist: unknown.
ERRO[0001] error waiting for container: context canceled


Version-Release number of selected component (if applicable):
Testing with Mageia 8 Beta 1 with Docker version 19.03.0-dev, build 42e35e6
installed from the Mageia repositories.

How reproducible:  always

Steps to Reproduce:
1.Install Docker and docker-containerd from the Cauldron repositories.
2.As root run: docker run hello-world
3.

There is a workaround observed from various other sites:
mkdir /sys/fs/cgroup/systemd
mount -t cgroup -o none,name=systemd cgroup /sys/fs/cgroup/systemd
Comment 1 Aurelien Oudelet 2020-09-06 00:14:07 CEST
Hi, thanks for reporting this bug.
This is strange behavior: 

ll /sys/fs/cgroup/
-r--r--r--  1 root root 0 sept.  5 20:46 cgroup.controllers
-rw-r--r--  1 root root 0 sept.  6 00:06 cgroup.max.depth
-rw-r--r--  1 root root 0 sept.  6 00:06 cgroup.max.descendants
-rw-r--r--  1 root root 0 sept.  5 20:46 cgroup.procs
-r--r--r--  1 root root 0 sept.  6 00:06 cgroup.stat
-rw-r--r--  1 root root 0 sept.  5 23:56 cgroup.subtree_control
-rw-r--r--  1 root root 0 sept.  6 00:06 cgroup.threads
-r--r--r--  1 root root 0 sept.  6 00:06 cpuset.cpus.effective
-r--r--r--  1 root root 0 sept.  6 00:06 cpuset.mems.effective
-r--r--r--  1 root root 0 sept.  6 00:06 cpu.stat
drwxr-xr-x  2 root root 0 sept.  5 23:13 init.scope/
-r--r--r--  1 root root 0 sept.  6 00:06 memory.stat
drwxr-xr-x 89 root root 0 sept.  6 00:02 system.slice/
drwxr-xr-x  3 root root 0 sept.  5 23:13 user.slice/

Seems there are all systemd related informations. These should be mounted in /sys/fs/cgroup/systemd instead for Docker compatibility?

Does this is Docker bug or wrong systemd config?

As there is no maintainer for systemd package I added the committers in CC.
(Packagers: Please set the status to 'assigned' if you are working on it)

Keywords: (none) => Triaged
Assignee: bugsquad => bruno.cornec
Source RPM: (none) => docker-19.03.11-2.mga8.src.rpm systemd-246.4-1.mga8.src.rpm
CC: (none) => pkg-bugs, smelror

Aurelien Oudelet 2020-09-07 13:38:13 CEST

Target Milestone: --- => Mageia 8

Comment 2 papoteur 2020-12-24 17:05:49 CET
Hello,
I confirm this problem.
It seems that systemd in Mageia 7 creates /sys/fs/cgroup/systemd but no more in cauldron.
In Mageia 7:

ll /sys/fs/cgroup/systemd
total 0
-rw-r--r--  1 root root 0 déc.  24 16:56 cgroup.clone_children
-rw-r--r--  1 root root 0 déc.  24 16:56 cgroup.procs
-r--r--r--  1 root root 0 déc.  24 16:56 cgroup.sane_behavior
drwxr-xr-x  2 root root 0 déc.  24 16:56 init.scope/
-rw-r--r--  1 root root 0 déc.  24 16:56 notify_on_release
-rw-r--r--  1 root root 0 déc.  24 16:56 release_agent
drwxr-xr-x 57 root root 0 déc.  24 16:56 system.slice/
-rw-r--r--  1 root root 0 déc.  24 16:56 tasks
drwxr-xr-x  3 root root 0 déc.  24 16:56 user.slice/

Adding joequant in CC

CC: (none) => joequant, yves.brungard_mageia

Comment 3 Joseph Wang 2020-12-25 06:21:12 CET
I've been using the workaround.  The issue is that the kernel and cgroups moved things around.  It's really something that upstream should address, but I haven't kept track of what's going on there.

Not sure what the right solution is to make it easy on users.

CC: (none) => joequant

Comment 4 Samuel Verschelde 2021-01-06 12:56:24 CET
I can confirm the issue.

Likely this bug : https://github.com/docker/for-linux/issues/219

See also https://www.linuxuprising.com/2019/11/how-to-install-and-use-docker-on-fedora.html regarding docker and compatibility with cgroups2 (not sure what version of cgroups we have).

Above references suggest upgrading docker could solve the issue.

I'm raising severity and priority here because a non-functional docker out of the box should be avoided.

Priority: Normal => High
Severity: minor => major


Note You need to log in before you can comment on or make changes to this bug.