Improve Solr database delete/load for name-lookup

[gaurav]

The current sequence order is:

1. Init container: We wipe the Solr data directory (which is also where we download the backup)
2. Init container: We upload the Solr backup into the Solr data directory.
3. Init container: We start Solr.
4. Restore-job pod: Command Solr to load the Solr backup into its data directory.

However, this can cause a problem if the Kubernetes instance is restarted -- the init container will wipe the Solr data directory, but the restore-job pod is not triggered.

A better system might be:

1. Move the deletion/clear operation to the Restore-job pod.
2. Modify the download code so that the file isn't redownloaded if it hasn't been modified.

[gaurav]

@YaphetKG has added a way to do this: by running helm upgrade name-lookup ... --set dataUrl='' once name-lookup has started. This ensures that the Solr database is not re-downloaded if the node is restarted. Now the next question is how we can ensure that ITRB does this -- can we integrate this into https://github.com/helxplatform/translator-devops/blob/develop/jenkins/name-lookup/Jenkinsfile somehow? Or maybe just let them to do this when upgrading a node, and see if they have ideas about automating this?

[gaurav]

Here's what the configuration looks like in plater, but it's unclear how exactly this parameter is used:

translator-devops/jenkins/plater/Jenkinsfile

Line 13 in e158e62

choice(name: 'LOAD_DATA', choices: 'no\nyes', description: 'Load data or not, default no')