Why not use docker-compose to set up the networks, volumes, etc? Also, why not start ceod automatically in docker, but to use sleep infinity and attach? That seems like bunch of manual steps that can be automated.
Why not use docker-compose to set up the networks, volumes, etc? Also, why not start ceod automatically in docker, but to use sleep infinity and attach? That seems like bunch of manual steps that can be automated.
We could use docker-compose - I just felt like writing a shell script (feel free to modify).
As for why we sleep infinity, it's because you're going to need to manually restart ceod if you change the source code. If ceod was the foreground process, this would require killing and re-creating the container. I figured it would be easier to just manually start ceod so that we can Ctrl-C and restart it.
We *could* use docker-compose - I just felt like writing a shell script (feel free to modify).
As for why we `sleep infinity`, it's because you're going to need to manually restart ceod if you change the source code. If ceod was the foreground process, this would require killing and re-creating the container. I figured it would be easier to just manually start ceod so that we can Ctrl-C and restart it.
I updated the VM so that we can use the same ceod_dev.ini file in both: 2438b1c1d9
I updated the VM so that we can use the same ceod_dev.ini file in both:
https://git.uwaterloo.ca/csc/syscom-dev-environment/-/commit/2438b1c1d9d270aedb79b388e912df584853a319
Maybe we can have a simple superviser that restarts ceod on signal? Normally ceod restarts by itself anyways so manual restart would only be needed occasionally (using docker kill -sighup or similiar). Main thing is that I feel having to attach to each individual manually is too much manual work especially when docker can start services by itself.
I can make a pr to this branch with docker-compose and the superviser thing if you want, although I don't want to install docker on my computer right now. Doe any csc machines have docker installed?
Maybe we can have a simple superviser that restarts ceod on signal? Normally ceod restarts by itself anyways so manual restart would only be needed occasionally (using docker kill -sighup or similiar). Main thing is that I feel having to attach to each individual manually is too much manual work especially when docker can start services by itself.
I can make a pr to this branch with docker-compose and the superviser thing if you want, although I don't want to install docker on my computer right now. Doe any csc machines have docker installed?
I can make a pr to this branch with docker-compose and the superviser thing if you want
Sure, that'd be great.
Doe any csc machines have docker installed?
Unfortunately none of the general-use machines have Docker installed - we haven't figured out a way to do that securely.
> I can make a pr to this branch with docker-compose and the superviser thing if you want
Sure, that'd be great.
> Doe any csc machines have docker installed?
Unfortunately none of the general-use machines have Docker installed - we haven't figured out a way to do that securely.
Use docker-compose to manage the containers, and start ceod automatically. The service can be restarted with SIGHUP sent by docker/docker-compose.
Co-authored-by: Rio6 <rio.liu@r26.me>
Reviewed-on: #19
Co-authored-by: Rio <r345liu@localhost>
Co-committed-by: Rio <r345liu@localhost>
a268wang
r345liu
Why not use docker-compose to set up the networks, volumes, etc? Also, why not start ceod automatically in docker, but to use sleep infinity and attach? That seems like bunch of manual steps that can be automated.
We could use docker-compose - I just felt like writing a shell script (feel free to modify).
As for why we
sleep infinity, it's because you're going to need to manually restart ceod if you change the source code. If ceod was the foreground process, this would require killing and re-creating the container. I figured it would be easier to just manually start ceod so that we can Ctrl-C and restart it.I updated the VM so that we can use the same ceod_dev.ini file in both:
2438b1c1d9Maybe we can have a simple superviser that restarts ceod on signal? Normally ceod restarts by itself anyways so manual restart would only be needed occasionally (using docker kill -sighup or similiar). Main thing is that I feel having to attach to each individual manually is too much manual work especially when docker can start services by itself.
I can make a pr to this branch with docker-compose and the superviser thing if you want, although I don't want to install docker on my computer right now. Doe any csc machines have docker installed?
Sure, that'd be great.
Unfortunately none of the general-use machines have Docker installed - we haven't figured out a way to do that securely.
de18a9f293into v1 2 years agoReviewers
de18a9f293.