Issues launching a service on 1.12

Open Source Projects DockerEngine
1

I’m running docker 1.12.0-rc4.
If I run the following everything works as expected:

docker run -p9117:9117 --name files_jackett2 tyler43636/unraid-jackett

Removing that image then running this:

docker service create -p 9117:9117 --name files_jackett tyler43636/unraid-jackett

results in the task constantly restarting. I’m not totally sure where to find the logs for docker, but looking at my syslog I can see this error is definitely related to this logged error:

Jul 16 13:41:21 dar-docker-02 kernel: [17951.925291] aufs au_opts_verify:1597:dockerd[1515]: dirperm1 breaks the protection by the permission bits on the lower branch
Jul 16 13:41:22 dar-docker-02 kernel: [17952.220483] aufs au_opts_verify:1597:dockerd[1515]: dirperm1 breaks the protection by the permission bits on the lower branch
Jul 16 13:41:23 dar-docker-02 kernel: [17953.221603] aufs au_opts_verify:1597:dockerd[1515]: dirperm1 breaks the protection by the permission bits on the lower branch
Jul 16 13:41:23 dar-docker-02 dockerd[990]: time=“2016-07-16T13:41:23.748530154+01:00” level=error msg=“fatal task error” error= module=taskmanager task.id=eklbwzrikps66nc1q2l9s9dxc

Is there anything else I can do to track this issue down? I’m really quite new to docker so I can’t tell if it’s just a beta bug, or an issue with the image, but thought it was worth posting anyway!

2

Find the error detail from the containe rlogs. First, obtain the container id with
docker ps

Subsequently, obtain the container logs with
docker logs [OPTIONS] CONTAINER

3

AH! I couldn’t find them until I did -a. Anyway, the logs are empty for all the containers it had spun up. I tried using the toke/mosquitto image and the same thing - works fine running using docker run, but as a service I get the same error, and the logs for the containers are empty :/.

4

For listing containers that have exited -a is required. A container could exit if it has completed and does not necessarily indicate an error. What is the Dockerfile for the image?

5

Sorry, I understood that - maybe this will help!

ross@dar-docker-02:~$ docker ps -a
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
94c98031822d toke/mosquitto:latest “docker-entrypoint.sh” 5 seconds ago Created openhab-mosquitto.1.2ykb609cgvmn9hrrgd1mxm3qw
0d3f94ef7aa7 toke/mosquitto:latest “docker-entrypoint.sh” 14 seconds ago Created openhab-mosquitto.1.30ecgv6sufuki4pkqe88d7ev8
SNIP
ross@dar-docker-02:~$ docker logs 94c98031822d
ross@dar-docker-02:~$

6

i’m having the exact same issue, i have a dns server i want to run and whenever i create a service with it, the service just keep restarting. So fast that i can’t attach to a container.

i had 1 time the luck of attaching to a container and it kept that one running. As soon as i detached, it kept restarting again.

how do we get the logs to understand the issue

7

actually found my answer, my repo is private and both docker login and pulls need to be done on each server before calling the service.

i would be great to login at the swarm level

8

Use this to automatically login on all Swarm nodes to pull the image:

docker stack deploy --with-registry-auth

Doc

Send registry authentication details to Swarm agents

PS: Funny how the forum here sometimes highlights year old posts :sweat_smile:

9

It usually happens for a reason: either a topic got moved into another category, or more likely someone posted spam in their response, which got removed.