CentOS: docker fails to start with "no sockets found via socket activation"
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
kolla-ansible |
Fix Released
|
Medium
|
Jeffrey Zhang |
Bug Description
The kolla-ansible baremetal roles sets up a systemd docker unit override:
This is what ends up being set up on my system:
====
# cat /etc/systemd/
[Service]
MountFlags=shared
ExecStart=
ExecStart=
-H fd://
====
This yields an error when starting docker:
====
Dec 21 16:21:32 centos-testing systemd[1]: Starting Docker Application Container Engine...
Dec 21 16:21:32 centos-testing docker[13361]: time="2016-
Dec 21 16:21:32 centos-testing systemd[1]: docker.service: main process exited, code=exited, status=1/FAILURE
Dec 21 16:21:32 centos-testing systemd[1]: Failed to start Docker Application Container Engine.
Dec 21 16:21:32 centos-testing systemd[1]: Unit docker.service entered failed state.
Dec 21 16:21:32 centos-testing systemd[1]: docker.service failed.
====
In multiple references to this issue, the solution seems to not use "-H fd://", ex:
- https:/
- https:/
- https:/
I've confirmed (manually) that removing it makes Docker work again.
I don't see "-H fd://" mentioned in Kolla documentation, such as here http://
Can we take that out of the baremetal role ? CentOS is broken when using the baremetal role right now.
Changed in kolla-ansible: | |
milestone: | none → ocata-3 |
-H fd:// should be removed.
now in docker-engine rpm package, there is not docker.socket file. and add this parameter will break the docker-engine service.