changed load balancer ip

[scottyhq]

had to re-deploy production cluster manually because it seems that circleci failing due to a timeout on the first deployment lead to a helm list STATUS=FAILED. apparently this condition leads to a number of issues (helm/helm#3353 (comment)). So we solved this pickle by removing the failed jupyterhub and re-deploying locally:

1. remove failed cluster

helm delete icesat2-prod --purge
kubectl delete namespace icesat2-prod

2. comment out this file
   https://github.com/pangeo-data/pangeo-cloud-federation/blob/prod/deployments/icesat2/config/prod.yaml

3. initial helm deploy (run from local pangeo-cloud-federation directory)

RELEASE=icesat2
helm upgrade --install --namespace $RELEASE-prod $RELEASE-prod pangeo-deploy -f deployments/$RELEASE/config/common.yaml -f deployments/$RELEASE/config/prod.yaml -f deployments/$RELEASE/secrets/prod.yaml --set pangeo.jupyterhub.singleuser.image.tag=2035c99 --set pangeo.jupyterhub.singleuser.image.name=783380859522.dkr.ecr.us-west-2.amazonaws.com/pangeo --timeout 3000

4. added in loadBalancerIP (from kubectl --namespace=icesat2-prod get svc) to
   https://github.com/pangeo-data/pangeo-cloud-federation/blob/prod/deployments/icesat2/config/prod.yaml

5. re-run helm command from step 3)

RELEASE=icesat2
helm upgrade --install --namespace $RELEASE-prod $RELEASE-prod pangeo-deploy -f deployments/$RELEASE/config/common.yaml -f deployments/$RELEASE/config/prod.yaml -f deployments/$RELEASE/secrets/prod.yaml --set pangeo.jupyterhub.singleuser.image.tag=2035c99 --set pangeo.jupyterhub.singleuser.image.name=783380859522.dkr.ecr.us-west-2.amazonaws.com/pangeo --timeout 3000

this might prevent the situation from reoccurring in the future:
berkeley-dsep-infra/hubploy#17