ci: alpine build errors

[chu11]

The new alpine builder seems to be failing at a slightly higher rate than other builders. It's possible alpine is uncovering raciness in tests that haven't been seen before. This is simply document some of the failing tests. Some of these may be completely random and unrelated to alpine.

  expecting success: 
  	id=$(flux submit -N4 -t20s cat) &&
  	echo hello from 0 \
  		| flux job attach --label-io -i0 $id >stdin-ranks.out 2>&1 &&
  	flux job eventlog -p guest.input $id &&
  	cat <<-EOF >stdin-ranks.expected &&
  	0: hello from 0
  	EOF
  	test_cmp stdin-ranks.expected stdin-ranks.out
  
  not ok 19 - attach: --stdin-ranks works

  expecting success: 
  	jobid=$(flux submit --wait-event=start sleep 300) &&
  	update1=$(expiration_add $jobid 100) &&
  	flux job eventlog $jobid &&
  	${UPDATE_LOOKUP} $jobid R > lookup3A.out &&
  	update2=$(expiration_add $jobid 200) &&
  	${UPDATE_LOOKUP} $jobid R > lookup3B.out &&
  	flux cancel $jobid &&
  	cat lookup3A.out | jq -e ".execution.expiration == ${update1}" &&
  	cat lookup3B.out | jq -e ".execution.expiration == ${update2}"
  
  1701097007.629339 submit userid=1001 urgency=16 flags=0 version=1
  1701097007.641013 validate
  1701097007.651737 depend
  1701097007.651783 priority priority=16
  1701097007.653275 alloc annotations={"sched":{"resource_summary":"rank0/core0"}}
  1701097007.657305 start
  1701097007.720380 resource-update expiration=1701097107
  true
  false
  not ok 3 - job-info: update lookup with update events works (job active)

(note, this is #5594)

  not ok 10 __main__.TestJobOutput.test_output_event_watch_nowait
  FAIL: python/t0015-job-output.py 10 __main__.TestJobOutput.test_output_event_watch_nowait
    ---
      message: |
        Traceback (most recent call last):
          File "/usr/src/t/./python/t0015-job-output.py", line 233, in test_output_event_watch_nowait
            with self.assertRaises(FileNotFoundError):
        AssertionError: FileNotFoundError not raised
    ...

[grondo]

A couple of these were addressed in #5609. The other already has a separate issue open, so closing this generic one.