Containerize Covizu Front-End for Docker Swarm

[GopiGugan]

We currently have to completely stop/kill and start pm2 to get the front-end to use the updated data files (see #475). We need to look for alternatives to pm2 for zero downtime deployments

[GopiGugan]

Currently looking into using one of the following:

• Kubernetes
• Docker Swarm

[ArtPoon]

@GopiGugan reported that since we are using a single system, it makes more sense to use Docker Swarm

[ArtPoon]

VM webserver has 32GB RAM and is running on Intel Xeon Gold 5218 CPUs (2.3GHz, 4 cores) Old webserver had 64GB RAM so memory may be an issue

[ArtPoon]

We could ask IT services to reprovision this VM with more RAM

[ArtPoon]

@ArtPoon to inquire about RAM upgrade

[ArtPoon]

Heard back from WTS, turns out we are paying a lot for the VM:

there is a cost increase of $4/GB, new billing is $318/month vs $190/month currently so we should probably have a hard look at third-party alternatives

[ArtPoon]

We should prioritize getting the old webserver back up

[ArtPoon]

Old webserver (coreceptor, now ccr5) is back up and on the network.
I have also installed OS for a backup server (cxcr4) that was once part of the cluster.

[ArtPoon]

@GopiGugan has ccr5 ready for traffic, @ArtPoon to update record to new IP address

[ArtPoon]

Records updated

[ArtPoon]

Waiting on validation of mongoDB to bring up second instance

[ArtPoon]

@GopiGugan created Docker file for MongoDB, as well as CoVizu Docker files, so have everything needed to run Swarm

[GopiGugan]

• Covizu seems to be slow when running an instance of the node docker container inside another docker container in my test environment. But this could be because I was running an instance in an environment with limited resources. I would like to set up another test environment on Paphlagon (?)
• Looking into create a service (using Docker Compose) to avoid having to use a container within a container

[ArtPoon]

go ahead and set up a test environment on Paphlagon

[GopiGugan]

• Looking into create a service (using Docker Compose) to avoid having to use a container within a container

Using containers within a container impacts performance. Currently setting up a service using docker compose to deploy the front-end (with mongodb) on the web server

[ArtPoon]

@GopiGugan reports that front-end should be updated by this afternoon

[GopiGugan]

Front-end has been updated