I have been using this MUP config
for the deployment
until recently. When I encountered an issue and I had to stop, reboot the Instance multiple times
.
Then, This causes the meteor app container to shut down and the MongoDB container is running just fine but wasn't accessible through an SSH tunnel on a MongoDB GUI
(but running systemctl status mongo
shows active status: activating
.
I troubleshoot and run docker ps -a
. It shows the MongoDB container only as a running
container and the meteor app container completely shutdown.
I tried running the MUP deployment
in an attempt to get the meteor app container up
and running.
However, I got an error Removing docker containers. Errors about nonexistent endpoints and containers are normal.
I run the mup setup
command successfully and then I tried running mup reconfig
and I got the same above error, I have attached the screenshot of the error below.
To Reproduce this error
- Create a meteor app with Iron-meteor.
- Setup an Instance (Ec2).
- Setup Deployment with Meteor-up
- Deploy your app with Meteor-up.
- SSH into the instance and run cmd
docker ps
. Should see at least two running containers, app and mongo respectively. - Run a cmd to
stop
theapp container
while themongo container
isrunning
. - Finally, Goto your project and redeployed with
mup
Should see a similar error as above. for step 6
restarting the instance in my case shut down the two containers and I was able to get the mongo container back up and running.
However, I couldn't get the app container running, so I tried redeploying with the expectation that a new app container would be created if it doesn't exist on the instance.
UPDATED!
I don't know if this will help, but in my experience, mup likes a fresh instance better than an existing one.
My first step would be a
mup stop
command. This will shut down the docker instances. Then you can remove them withdocker rm
, and you can remove the images withdocker rmi
. Then do amup setup
again, followed by amup deploy
.If the first one doesn't work, you can basically start with a fresh vm, as in the droplet or ec2 instance. This is generally quite successful.