Setup ssjdispatcher service and deployment - usually run as part of gen3 roll all
When run on the admin-vm this will try to configure the ssjdispatcher secrets
in creds.json
.
- setup sa-linked roles from ssjdispatcher and its jobs
- create an upload bucket
- create sns and sqs
- setup indexd creds
- setup metadata service creds if they are present
If auto
is provided as the bucket name, then the script constructs a
safe name for the bucket. The caller must configure the DATA_UPLOAD_BUCKET
in fence-config-public
.
If an SQS URL is provided, then kube-setup-dispatcher
uses that rather than
create a new SQS.
This is one easy way to upgrade a legacy ssj deployment to use an AWS role (rather than user creds):
- remove the existing
ssjdispatcher
config from creds.json:
creds="$(cat creds.json)"
jq -r '. | del(.ssjdispatcher)' > creds.json <<< "$creds"
- run
kube-setup-ssjdispatcher
with the existing bucket name and SQS url:
gen3 kube-setup-ssjdispatcher existing-bucket-name https://existing-sqs-url