/job/ExampleWorkspace/job/ExampleProject/job/Create_Environment job randomly failing on a Docker Swarm deployment

[midalias]

Some times the Create_Environment job fails with error “nginx: [error] open() ‘/run/nginx.pid’ failed (2: No such file or directory)”

We believe this error is because the nginx process is not started by the moment of the job's execution. If that's the case, Giovanni and I think we can solve this using either a loop or a sleep command. It needs to be tested.

[anton-kasperovich]

@midalias could you please provide more details, is it Swarm or Single-host, which version of Docker, Docker-compose? Some details like, how often you ran Create_Environment job, which parameter value you chosen? Didn't saw such error, need to troubleshooting.

[midalias]

Hi @anton-kasperovich, this has happened in ADOP on Azure deployments and has been random. The solution is a Docker Swarm cluster. About the Docker version, we use an Docker extension in our ARM template. In that URL says "Install latest stable version of Docker Engine on your Linux VM". I am going through the documentation to see if we capture the Docker version.

The Create_Environment job was executed once. We have a script (adop.sh) that triggers the execution of "Load_Platform" job. I believe the "Create_Environment" is executed from there using the "Dev" parameter.

I'll contact my coworker to validate this information.