Openjdk-infrastructure: Determine differences between the 2 linux_x86-64 test machines

[smlambert]

Need to resolve configuration differences between the 2 linux_x86-64 machines:

test-scaleway-x64-ubuntu-16.04-1
and
test-packet-x64-ubuntu-16.04-1

Docker-based tests pass on the scaleway machine, and fail on the packet machine.

Failure message:
18:23:57 [exec] Status: Image is up to date for adoptopenjdk/openjdk8:latest
18:23:57 [exec] ---> 8ede58073db6
18:23:57 [exec] Step 4/12 : RUN apt-get update && apt-get -y install ant apt-transport-https ca-certificates dirmngr curl git make unzip vim wget
18:23:57 [exec] ---> Running in 556463860d06
18:23:57 [exec] container_linux.go:265: starting container process caused "process_linux.go:368: container init caused \"process_linux.go:351: running prestart hook 0 caused \\"fork/exec /usr/bin/dockerd (deleted): no such file or directory\\"\""
18:23:57 [exec] oci runtime error: container_linux.go:265: starting container process caused "process_linux.go:368: container init caused \"process_linux.go:351: running prestart hook 0 caused \\"fork/exec /usr/bin/dockerd (deleted): no such file or directory\\"\""

I checked that they both have the same version of Docker installed (Docker version 17.09.1-ce, build 19e2cf6).

[pnstanton]

Apparently the error fork/exec /usr/bin/dockerd (deleted): no such file or directory can be caused by something deleting and recreating /usr/bin/dockerd whilst it's still running, so that the inode of /usr/bin/dockerd doesn't match the inode of the existing dockerd process.
Looking on test-packet-x64-ubuntu-16.04-1 I can see that dockerd was started on October 18th last year, and the docker-ce package was updated on December 11th.
I've restarted the Docker service. Does that fix the issue?

[gdams]

ping @smlambert

[smlambert]

Saw this, tried rerunning yesterday but Jenkins reboot blocked launching new builds... will rerun now.

[smlambert]

Thanks, they are now running on both machines.

[smlambert]

BTW, does it imply that after an ansible refresh an additional step should occur (restarting services)?

[bblondin]

not a bad idea... it is currently set to started, I will update it to restarted
This will ensure the service gets bounced.