[SPARK-31418][CORE][FOLLOW-UP][MINOR] Fix log messages to print stage id instead of the object name

[venkata91]

What changes were proposed in this pull request?

Just few log lines fixes which are logging the object name instead of the stage IDs

Why are the changes needed?

This would make it easier later for debugging.

Does this PR introduce any user-facing change?

No

How was this patch tested?

Just log messages. Existing tests should be enough

[venkata91]

@tgravescs This is just a minor log line changes which I found printing object names instead of the IDs.

[tgravescs]

ok to test

[tgravescs]

what did the log look like before? I thought printing the taskSet should give you something like: TaskSet 2.0. where 2 is the stage id and 0 is the stage attempt id. Now its only printing the stage id without the attempt id?

[SparkQA]

Test build #126776 has finished for PR 29279 at commit 39e22c5.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[venkata91]

Infact its not even printing stageId rather printing the TaskSetManager object name. Do you think instead of printing stageId its better to print name taskSet.name?

For eg:

20/07/29 16:10:04 INFO YarnScheduler: Notifying ExecutorAllocationManager to allocate more executors to schedule the unschedulable task before aborting org.apache.spark.scheduler.TaskSetManager@f031375.

20/07/29 16:10:04 INFO YarnScheduler: Waiting for 120000 ms for completely blacklisted task to be schedulable again before aborting org.apache.spark.scheduler.TaskSetManager@f031375.```

[tgravescs]

what if you explicitly put ${taskSet.toString}

[tgravescs]

oh sorry, I was looking at the wrong thing, those are TaskSetManagers not TaskSet

[venkata91]

oh sorry, I was looking at the wrong thing, those are TaskSetManagers not TaskSet

Yes thats right. Do you think its better to add toString to `TaskSetManager?

I think ${taskSet.stageId} would be good enough. Though I tried with ${taskSet.name} which prints like this:

20/07/29 18:49:31 INFO YarnScheduler: Waiting for 120000 ms for completely blacklisted task to be schedulable again before aborting TaskSet_0.0.

With stageId:

20/07/29 18:52:44 INFO YarnScheduler: Waiting for 120000 ms for completely blacklisted task to be schedulable again before aborting stage 0.```

[tgravescs]

no, I think this is fine.

[SparkQA]

Test build #126811 has finished for PR 29279 at commit ad847b1.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[HyukjinKwon]

retest this please

[SparkQA]

Test build #126848 has finished for PR 29279 at commit ad847b1.

• This patch fails due to an unknown error code, -9.
• This patch merges cleanly.
• This patch adds no public classes.

[HyukjinKwon]

retest this please

[SparkQA]

Test build #126873 has finished for PR 29279 at commit ad847b1.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[HyukjinKwon]

Merged to master.