• Blue Ocean 1.2-beta1, Blue Ocean 1.2-beta2, Blue Ocean 1.2-beta3, Blue Ocean 1.2-beta4, Blue Ocean 1.2, Blue Ocean 1.3, Blue Ocean 1.4 - beta 1, Blue Ocean 1.5 - beta 3

      Using Blue Ocean to visualize our pipeline, one of the steps [Prepare] shows that it has been running for an absurdly long time, especially since it just started 5-10 seconds ago....

      Prepare .... 17339d 18h 6m 25s

      I believe I've seen this on other steps as well.

      Please see attached image as an example of the output.

          [JENKINS-45080] Blue Ocean displays wrong step time

          James Dumay added a comment - - edited

          abayer it looks like the fix you made in JENKINS-42636 has regressed

          Would you mind taking a look?

          James Dumay added a comment - - edited abayer it looks like the fix you made in JENKINS-42636 has regressed Would you mind taking a look?

          James Dumay added a comment -

          barnett2 could you please capture a HAR file for a session where you are able to reproduce this issue and upload it to this secure file drop please? This will help us diagnose what step it is and pinpoint the appropriate fix.

          abayer I will let you know when I have received the file.

          James Dumay added a comment - barnett2 could you please capture a HAR file for a session where you are able to reproduce this issue and upload it to this secure file drop please? This will help us diagnose what step it is and pinpoint the appropriate fix. abayer I will let you know when I have received the file.

          Dave Barnett added a comment -

          jamesdumay I've not forgotten, but it is not a consistent thing.

          Earlier today, I started a build. The first few steps in Blue Ocean view appeared fine. Once the build moved on to a later step, the screen update seemed to hang.

          My "Shell Script" item showed no contents [little black line like it was waiting to populate the output box] and the time at the end of the header had 17xxx days as seen before. And there it sat for quite some time [I wandered away, was the same when I came back several minutes later].

          If I reload the page, the page then loaded the output section correctly and the header showed a sensible time for the step.

          If I get into the state that the elapsed time entry is wrong and things seem hung, would recording from that point help, or does it need to be recorded from the time I open the page...?

          Perhaps doing a "view source" of the hung page load would give you something useful since a reload "fixes" the issue...?

          Best guess is that some communication hiccup causes the initialization process to not reset the timestamp since both the timestamp and the dialog contents are empty....

          Dave Barnett added a comment - jamesdumay I've not forgotten, but it is not a consistent thing. Earlier today, I started a build. The first few steps in Blue Ocean view appeared fine. Once the build moved on to a later step, the screen update seemed to hang. My "Shell Script" item showed no contents [little black line like it was waiting to populate the output box] and the time at the end of the header had 17xxx days as seen before. And there it sat for quite some time [I wandered away, was the same when I came back several minutes later] . If I reload the page, the page then loaded the output section correctly and the header showed a sensible time for the step. If I get into the state that the elapsed time entry is wrong and things seem hung, would recording from that point help, or does it need to be recorded from the time I open the page...? Perhaps doing a "view source" of the hung page load would give you something useful since a reload "fixes" the issue...? Best guess is that some communication hiccup causes the initialization process to not reset the timestamp since both the timestamp and the dialog contents are empty....

          James Dumay added a comment -

          barnett2 recorded from anytime before you saw the problem would suffice in this instance.

          James Dumay added a comment - barnett2 recorded from anytime before you saw the problem would suffice in this instance.

          Dave Barnett added a comment -

          jamesdumay I don't know if I was able to capture it or not. I've saved what I got as requested. Perhaps it will help...

          Dave Barnett added a comment - jamesdumay I don't know if I was able to capture it or not. I've saved what I got as requested. Perhaps it will help...

          Dave Barnett added a comment -

          jamesdumay I added an "attempt_2.har" file this morning. This one should be fairly short and I think I captured the exchange which should show the issue.

          Dave Barnett added a comment - jamesdumay I added an "attempt_2.har" file this morning. This one should be fairly short and I think I captured the exchange which should show the issue.

          James Dumay added a comment -

          Thanks barnett2.

          abayer I shared the HAR with you.

          James Dumay added a comment - Thanks barnett2 . abayer I shared the HAR with you.

          Cliff Meyers added a comment -

          Duration corresponds to time since the epoch, so there must be a null / 0 start time getting passed somewhere.

          Cliff Meyers added a comment - Duration corresponds to time since the epoch, so there must be a null / 0 start time getting passed somewhere.

          Karl Shultz added a comment -

          Testing Notes:

          • Once implemented, unit tests should be included in the fix
          • Predictively testing the UI for step runtimes is probably not possible in any reliable way.

          Karl Shultz added a comment - Testing Notes: Once implemented, unit tests should be included in the fix Predictively testing the UI for step runtimes is probably not possible in any reliable way.

          Nicolae Pascu added a comment -

          I looked into this and could not replicate the problem. I tried sending 0/null/undefined and all cases were handled as expected. We also updated moment recently and that might have fixed a problem in the library that caused this. I will close the issue for now and we can re-open it in the future if someone will encounter this problem again.

          Nicolae Pascu added a comment - I looked into this and could not replicate the problem. I tried sending 0/null/undefined and all cases were handled as expected. We also updated moment recently and that might have fixed a problem in the library that caused this. I will close the issue for now and we can re-open it in the future if someone will encounter this problem again.

            nicu Nicolae Pascu
            barnett2 Dave Barnett
            Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

              Created:
              Updated:
              Resolved: