I have an issue when containers are running for several weeks/months on a host. After the mentioned amount of time, the health checks that are configured for these containers are failing. If you check
journal -u docker, you see a lot of messages about:
error: unable to find user consul: no matching entries in passwd file. For each container that has a health check an entry is logged for that specific user on which the process is running on.
This is happening via the different images (Alpine/Debian/CentOS) we use and all users are created in the
I can’t open a tty or run a command when this is happening, as when I do:
docker exec -it consul bash unable to find user consul: no matching entries in passwd file
For the services that hasn’t have a health check yet, I can’t open a tty too.
When I restart the
docker service and thus restart all containers, all is fine. Only restarting the containers is not sufficient. The containers are running fine when this is happening. Almost each service register itself to Consul and the health checks that Consul performance are ok.
Any idea what is happening? I don’t want to restart the
docker service every time I see this problem occur.