Michael Hobbs
Michael Hobbs
@eljojo not sure what's happening here. I built a custom herokuish image that has supervisor installed already so my deployments bypass this execution path. if you run `docker logs` on...
I think this should be updated to remove the same check from the plugin hook. No need to do this during runtime too.
You can run a custom [buildstep](https://www.github.com/progrium/buildstep) image that has the supervisord package already installed.
it looks like something is already running on port `18000`. you can run something like the following then post that output here. ``` $ netstat -lnpt | grep 18000 tcp...
can you provide the logs for the logspout container itself? ``` docker logs logspout ```
As far as I understand it, plugins are installed at image build time and thus out of scope for this plugin. However, `dokku-logspout` could be easily modified to support any...
@yabawock I've seen the issue almost solely under high resource utilization.
I definitely agree in principle. On a more meta level, I'm not sure how to allow container RW access to a volume without opening it up to all users since...
An idea from IRC...what if buildstep (`/exec`) exported the user and group?
Though I suppose the `stat` method is sufficient.