fix: Say no logs were outputted when pod is done

[simster7]

Fixes #2372. Thanks @foobarbecue!

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (master@d3a3f6b). Click here to learn what that means.
The diff coverage is n/a.

@@            Coverage Diff            @@
##             master    #2373   +/-   ##
=========================================
  Coverage          ?   13.07%           
=========================================
  Files             ?       72           
  Lines             ?    25058           
  Branches          ?        0           
=========================================
  Hits              ?     3276           
  Misses            ?    21369           
  Partials          ?      413

---

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 d3a3f6b...9af4bd8. Read the comment docs.

[foobarbecue]

This might have caused another issue. I now sometimes get "Pod did not output any logs." but then when I go to Artifacts for the same Pod, there are logs there. I think this might have to do with log rotation settings.

[foobarbecue]

Actually I think what was happening was that I'm using retryStrategy and the logs only show up on the child nodes representing retries, and not on the parent. Note that I'm using a version without the #2384 fix so all of my nodes have a (0) version even if they worked the first time. These are the ones where "logs" shows output on the (0) node but not on its parent. Both the parent and (0) node have archived logs accessible.

[simster7]

Actually I think what was happening was that I'm using retryStrategy and the logs only show up on the child nodes representing retries, and not on the parent. Note that I'm using a version without the #2384 fix so all of my nodes have a (0) version even if they worked the first time. These are the ones where "logs" shows output on the (0) node but not on its parent. Both the parent and (0) node have archived logs accessible.

Hi @foobarbecue, this is actually the correct behavior. The parent node is something we call a "virtual node" of type RetryNode which basically means that the node itself will not run a pod, but will point to other nodes that will – so the RetryNode should never contain logs. #2384 should help in addressing this, but the full solution to this problem should be #2406 if you're interested in contributing 😄

[foobarbecue]

Hi @simster7 , I see. I haven't been able to get my build environment set up yet, but I'll probably work on it some more this evening and if I can get it going I'll see if I think #2406 is within my abilities!

[simster7]

Back-ported to 2.6