1285550 - Make sure the run-mozharness binary has been setup before trying to call it from interactive wizard

Status: RESOLVED FIXED

(Testing :: General, defect, P1)

Description

[Andrew Halberstadt [:ahal]]

There is a race condition in the interactive wizard. If a developer chooses the 1 or 2 options very quickly after being presented the wizard, it's possible the 'test-linux.sh' process hasn't gotten around to setting up that binary yet, resulting in an OSError.

The wizard should make sure that binary exists before trying to invoke it.

Comment 1

[Andrew Halberstadt [:ahal]]

Attachment: Bug 1285550 - Ensure 'run-mozharness' binary exists before invoking from interactive wizard,

This fixes a race condition between the 'test-linux.sh' process and the 'taskcluster-interactive-shell'
process in interactive tasks.

Review commit: https://reviewboard.mozilla.org/r/63218/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/63218/

Comment 2

[Armen [:armenzg]]

I will look at this tomorrow.

Comment 3

[Armen [:armenzg]]

Comment on attachment 8769251 [details]
Bug 1285550 - Ensure 'run-mozharness' binary exists before invoking from interactive wizard,

https://reviewboard.mozilla.org/r/63218/#review60640

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/8b58bd55988b
Ensure 'run-mozharness' binary exists before invoking from interactive wizard, r=armenzg

Comment 5

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/8b58bd55988b

Comment 6

[Maja Frydrychowicz :impossibus (was :maja_zf) (needinfo me)]

I saw this today. 

1) Resume task - Resume the original task without modification. This can be
                 useful for passively monitoring it from another shell.
2) Setup task  - Setup the task (download the application and tests) but
                 don't run the tests just yet. The tests can be run with a
                 custom configuration later. This will provide a mach
                 environment (experimental).
3) Clone gecko - Perform a clone of gecko using the task's repo and update it
                 to the task's revision.
4) Exit        - Exit this wizard and return to the shell.
Select one of the above options: 1
Timed out after 30 seconds waiting for the 'run-mozharness' binary
run-mozharness
Traceback (most recent call last):
  File "./run-wizard", line 170, in <module>
    sys.exit(wizard())
  File "./run-wizard", line 163, in wizard
    ret = func()
  File "./run-wizard", line 36, in resume
    call(['run-mozharness'])
  File "./run-wizard", line 20, in call
    return subprocess.call(cmd, **kwargs)
  File "/usr/lib/python2.7/subprocess.py", line 523, in call
    return Popen(*popenargs, **kwargs).wait()
  File "/usr/lib/python2.7/subprocess.py", line 711, in __init__
    errread, errwrite)
  File "/usr/lib/python2.7/subprocess.py", line 1343, in _execute_child
    raise child_exception
OSError: [Errno 2] No such file or directory

Comment 7

[Andrew Halberstadt [:ahal]]

Interesting, was this reproducible? Did run-mozharness eventually show up?

Comment 8

[Maja Frydrychowicz :impossibus (was :maja_zf) (needinfo me)]

Yes, I could reproduce reliably: start the terminal session and run the wizard pretty soon after I got notified that the loaner was ready. (maybe 20 seconds?) 

run-mozharness did eventually show up, yes. 

In case it's useful, the loaners I tried this with were for this job: https://treeherder.mozilla.org/#/jobs?repo=try&author=[e-mail]&exclusion_profile=false&filter-tier=1&filter-tier=2&filter-tier=3.com&selectedJob=71510930

Comment 9

[Andrew Halberstadt [:ahal]]

Attachment: Increase timeout to 60 seconds

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/40580c92c633
Increase timeout to 60s while waiting for run-mozharness in a one-click loaner, r=maja_zf

Comment 11

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/40580c92c633