cap-pipeline-config

Building ontology pipeline configurations for Cell Annotation Platform

Deployment

Auto-deployment

Ask admin to grant you with Storage Object Admin role.

To auto-deploy solr.json index please keep gs://cap-gke-rc1-ols/solr/indexes/solr.json up to date. Every time CAP is deployed CI/CD pipeline is uses gs://cap-gke-rc1-ols/solr/indexes/solr.json to migrate indexes for OLS.

To upload local solr.json use the following commands:

export ENV=[sandbox_env_here]

gcloud auth login

gsutil cp solr.json gs://cap-${ENV}-ols/solr/indexes/solr.json

sandbox_env_here can be either one of gke-prod, gke-rc1, gke-rc2

Manual deployment

To access live Solr instance on a GCP project follow instructions described down below in CAP live Solr instance.

To deploy initial version of index run the following command:

./solr_init.sh -h [HOSTNAME] -p [PORT] -c [COLLECTION]

e.g.

./solr_init.sh -h localhost -p 8983 -c ontology

To deploy fresh version of index run the following command:

./solr_index.sh -h [HOSTNAME] -p [PORT] -c [COLLECTION]

e.g.

./solr_index.sh -h localhost -p 8983 -c ontology

CAP live Solr instance

Auth into GCP:

gcloud auth login

List available projects:

gcloud projects list

Before running further commands connect to remote solr instance:

export PROJECT_ID=[put_you_project_id_here]

e.g.

export PROJECT_ID=capv2-gke-rc1 # for RC1 taken from `gcloud projects list`

export PROJECT_ID=capv2-gke-prod # for Prod taken from `gcloud projects list`

Connect to solr:

ES_INSTANCE_NAME=$(gcloud compute instances list --project $PROJECT_ID | grep solr | awk '{ print $1 }')

ES_INSTANCE_REGION=$(gcloud compute instances list --project $PROJECT_ID | grep solr | awk '{ print $2 }')

gcloud compute ssh --project $PROJECT_ID --ssh-flag="-L 8984:localhost:8983" --zone $ES_INSTANCE_REGION $ES_INSTANCE_NAME