strider
strider copied to clipboard
Strider-CD
Cancelling a job does not invoke the clean up phase
When cancelling a job it does not run the clean up method, when using Docker this leaves problematic images/containers clogging the system up.
Also, if an unexpected error occurs it does not invoke the clean up phase, this might be verbiage in the custom scripts plugin stating "Cleanup always runs."
@davemackintosh I don't think anyone is looking into it. You're right, the problem still exists from back when we built the main guts of the docker runner (about 5 months ago).
I just granted you write access to the main repo as well as strider-docker-runner and assigned you to this issue. good luck let me know if i can be of any further assistance: I'm also always in the IRC channel as keyvan and get push notifications on my phone if you write my name.
I'm stopping work on this as I can't resolve an issue with the docker-modem integration while using boot2docker on OSx.
docker-modem doesn't seem to support the tcp protocol, I can access my docker rest api but for some reason docker-modem cannot and I spent wayy too long trying to figure out what was going on before stopping
It does, https://github.com/apocas/docker-modem/issues/31#issuecomment-68103138
Interesting, I guess the dependency in dockerode is out of date then as I did a manual install of the latest dockerode and still had the same problem. I'll try a manual install of the latest docker-modem in a bit and continue work if it fixes the problem and I'll submit a PR to dockerode I can see it says 0.1.x but maybe it should be ~0.1.22 so it can still get patches.
Doesn't that mean you'll get minor revisions as well which might have breaking changes?
Technically minor shouldn't have breaking changes, since those are new features. Breaking changes should increment major. But every project is different and it depends how dockerode is managed.
I'll look in a bit, some client work to do before hand which is more important.
