fix: Only save memoization cache when node succeeded

[terrytangyuan]

I am not sure if this is the intended behavior but currently cache is saved regardless of node's status, which is not useful and counter-intuitive to our users.

For example, a failed node may have no output like {"nodeID":"memoized-simple-workflow-fzvkz","outputs":null,"creationTimestamp":"2021-04-19T13:17:28Z"} and then if a new template retrieves the cache, it will become successful and subsequent steps would fail due to the lack of outputs from the previous step.

Let me know if this currently behavior is indeed desired. Then we probably need to introduce a toggle for this. We have no use case to save cache for unsuccessful nodes.

Signed-off-by: terrytangyuan [email protected]

Checklist:

• My organization is added to USERS.md.

[codecov]

Codecov Report

Merging #5711 (54571fa) into master (91c08cd) will decrease coverage by 0.00%.
The diff coverage is 50.00%.

@@            Coverage Diff             @@
##           master    #5711      +/-   ##
==========================================
- Coverage   46.91%   46.90%   -0.01%     
==========================================
  Files         244      244              
  Lines       15209    15210       +1     
==========================================
- Hits         7135     7134       -1     
+ Misses       7169     7168       -1     
- Partials      905      908       +3     

Impacted Files	Coverage Δ
workflow/controller/operator.go	70.85% <50.00%> (-0.21%)	⬇️
cmd/argo/commands/get.go	56.95% <0.00%> (+0.64%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 91c08cd...54571fa. Read the comment docs.

[terrytangyuan]

Seeing # Wait for workflow controller until lsof -i :9090 > /dev/null ; do sleep 10s ; done Error: The action has timed out.

I think #5696 should fix/relieve this. Am I right? @alexec