Closed Bug 1307087 Opened 9 years ago Closed 9 years ago

ATMO V2: scheduled jobs are failing

Categories

(Cloud Services Graveyard :: Metrics: Pipeline, defect, P1)

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: mdoglio, Assigned: mdoglio)

References

Details

Attachments

(5 files)

This is what I found in the logs: Traceback (most recent call last): File "/usr/local/lib/python2.7/dist-packages/rq/worker.py", line 588, in perform_job rv = job.perform() File "/usr/local/lib/python2.7/dist-packages/rq/job.py", line 498, in perform self._result = self.func(*self.args, **self.kwargs) File "/app/atmo/jobs/jobs.py", line 8, in launch_jobs SparkJob.step_all() File "/app/atmo/jobs/models.py", line 136, in step_all if spark_join.should_run(now): File "/app/atmo/jobs/models.py", line 89, in should_run active = self.start_date <= at_time <= self.end_date TypeError: can't compare offset-naive and offset-aware datetimes
Another error similar to the one in comment 0: Traceback (most recent call last): File "/app/.heroku/python/lib/python2.7/site-packages/rq/worker.py", line 588, in perform_job rv = job.perform() File "/app/.heroku/python/lib/python2.7/site-packages/rq/job.py", line 498, in perform self._result = self.func(*self.args, **self.kwargs) File "/app/atmo/jobs/jobs.py", line 8, in launch_jobs SparkJob.step_all() File "/app/atmo/jobs/models.py", line 139, in step_all if spark_join.should_run(now): File "/app/atmo/jobs/models.py", line 90, in should_run active = self.start_date <= at_time TypeError: can't compare offset-naive and offset-aware datetimes
I can't reproduce it anymore. I think it was fixed as a side-effect of fixing bug 1309227
No longer blocks: 1248688
Status: ASSIGNED → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
I am still unable to successfuly run a scheduled job. It looks like jobs are not started at all on EMR.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
Severity: normal → major
:rvitillo I dug a bit and found that spark job clusters aren’t launched with VisibleToAllUsers=True but False which is why I wasn’t able to see the cluster in the AWS console. Should I add that to the spark jobs?
Flags: needinfo?(rvitillo)
(In reply to Jannis Leidel [:jezdez] from comment #5) > :rvitillo I dug a bit and found that spark job clusters aren’t launched > with VisibleToAllUsers=True but False > which is why I wasn’t able to see the cluster in the AWS console. Should I > add that to the spark jobs? Go ahead.
Flags: needinfo?(rvitillo)
This also explains why I wasn't able to reproduce this issue locally where the jobs are spawned with my own credentials. Good spot!
Scheduled jobs appear to be still failing (tested on stage). I scheduled this simple notebook [1] which should not fail but it still did. As there are also no logs (filed Bug 1312749 for that one) I don't know what happened. [1] https://raw.githubusercontent.com/mozilla/telemetry-airflow/master/examples/spark/example_date.ipynb
I can reproduce this locally.
Closing this as this was fixed in staging.
Status: REOPENED → RESOLVED
Closed: 9 years ago9 years ago
Resolution: --- → FIXED
Jannis, did this land on prod?
Flags: needinfo?(jezdez)
:marco We just landed it (with an unplaned delay) in the atmo-prod.herokuapp.com environment.
Flags: needinfo?(jezdez)
This is still not working for me on stage. I don't see the jobs being scheduled from the AWS console and the UI doesn't give me any indication that it tried to run my jobs.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
Now that bug 1316623 is solved we can close this as well.
Status: REOPENED → RESOLVED
Closed: 9 years ago9 years ago
Resolution: --- → FIXED
Product: Cloud Services → Cloud Services Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: