Fix extraEnvVars in helm chart

[callumforrester]

Fixes #581

extraEnvVars was being treated as a string in some places and a list in others. This PR makes it a list everywhere, any values files that look like this should be updated to remove the |:

extraEnvVars: |
  - name: BEAMLINE
    value: i04

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.86%. Comparing base (6d0b50f) to head (0bda41b).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #582   +/-   ##
=======================================
  Coverage   92.86%   92.86%           
=======================================
  Files          40       40           
  Lines        1793     1793           
=======================================
  Hits         1665     1665           
  Misses        128      128           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[DominicOram]

Maybe naive question but is possible to have written a test that would have caught this? For hyperion we occasionally run our deploy script into /tmp for testing. Maybe a job that deploys the latest version into the central k8s cluster?

[callumforrester]

Yes, but I need to look into it further: #559

[DiamondJoseph]

I think this was required for other edge cases where we're injecting something in values.yaml into extraEnvVars as part of the tpl call? Or else maybe if extraEnvVars is empty?

[callumforrester]

Hmm, it may have been needed for rabbitmq password injection?

[DiamondJoseph]

Looking at i22,

extraEnvVars: |
    - name: SCRATCH_AREA
      value: {{ .Values.scratch.containerPath }}
    - name: RABBITMQ_PASSWORD
      valueFrom:
        secretKeyRef:
          name: rabbitmq-password
          key: rabbitmq-password

Has this been checked if it worked with the {{.Values.scratch.containerPath}} injection? Or are we going to replicate the value here?

[callumforrester]

With #520, the $SCRATCH_AREA variable is no longer needed, I assume the tpl was because of $RABBITMQ_PASSWORD

[callumforrester]

@DiamondJoseph and @DominicOram are we happy for this to go in?