Our current plan in order to take advantage of all resources given to us is to auto-build #34 three containers (I think, need confirmation):
Web Server/Job Submission Container (reader-ws)
A Job Running Container (reader-j)
Solr Index container (reader-s)
This means that for a given set of resources the web server will submit a job, which will spin up a container of the job container type, both will interact with Solr.
Solr meanwhile will have to be configured to be cross datacenter. #101
We will build two distant readers to start West and East, later we could add an India cluster. reader-ws will be deployed to Jetstream in East and Microsoft in West, reader-j will run on Slurm, Jetstream in East and LANL in West, Solr will run on Jetstream in East and Microsoft in West.
This means that for me, what I should do is use the future West implementation as the test bed for containerization. Maybe starting with a simple https://azure.microsoft.com/en-us/resources/templates/slurm/ Slurm implementation in the West.
Our current plan in order to take advantage of all resources given to us is to auto-build #34 three containers (I think, need confirmation):
This means that for a given set of resources the web server will submit a job, which will spin up a container of the job container type, both will interact with Solr.
Solr meanwhile will have to be configured to be cross datacenter. #101
We will build two distant readers to start West and East, later we could add an India cluster. reader-ws will be deployed to Jetstream in East and Microsoft in West, reader-j will run on Slurm, Jetstream in East and LANL in West, Solr will run on Jetstream in East and Microsoft in West.
This means that for me, what I should do is use the future West implementation as the test bed for containerization. Maybe starting with a simple https://azure.microsoft.com/en-us/resources/templates/slurm/ Slurm implementation in the West.