I am getting an error "Internal error occurred for the current attempt" while submitting a job on google cloud ML.
Can anyone help me on this ???
Asked
Active
Viewed 1,752 times
0
-
I also have the same . https://stackoverflow.com/questions/45609164/google-cloud-ml-engine-job-failed-due-to-an-internal-error-cant-execute-a-j – Shamane Siriwardhana Aug 10 '17 at 10:25
-
Did you try this on new job which means a brand-new project and a bucket ? – Shamane Siriwardhana Aug 10 '17 at 10:35
-
Yes, I have tried it on new project and bucket. i believe it has something to do with google only. I think there servers are down. Not sure though – Appu Aug 10 '17 at 11:43
-
Yeah I tried everything . Can you comment here when it works . I'll do the same . – Shamane Siriwardhana Aug 10 '17 at 11:45
-
1Sure.By the time , I am thinking of running the job in a different region. You can try the same. – Appu Aug 10 '17 at 11:50
-
By creating a new project ? – Shamane Siriwardhana Aug 10 '17 at 11:56
-
No. Try within the same project. – Appu Aug 10 '17 at 12:00
-
Hi how to change the region in an existing project ? – Shamane Siriwardhana Aug 10 '17 at 12:04
-
when you are submitting a job, there would be one argument as region. Just put whatever region you want the job to run in. – Appu Aug 10 '17 at 12:16
-
can you kindly share a screen shot . – Shamane Siriwardhana Aug 10 '17 at 12:24
-
Hi how did it go? – Shamane Siriwardhana Aug 10 '17 at 13:42
1 Answers
2
We're experiencing a bit of a capacity crunch making it a little hard to obtain VMs. You can try running in another region such as us-east1 for the time being. We are working on making sure the error message is more informative. Thanks for your patience.
rhaertel80
- 8,254
- 1
- 31
- 47
-
How can I change the region of the job. Can you please help me. – Shamane Siriwardhana Aug 10 '17 at 14:44
-
`gcloud ml-engine jobs submit training my_job ... --region=us-east1` – rhaertel80 Aug 10 '17 at 15:00
-
1
-
So I have to put the region before any other arguments right ? – Shamane Siriwardhana Aug 10 '17 at 15:12
-
1Before any of your *own* arguments. Here's a sample command-line: `gcloud ml-engine jobs submit training us_east1 --region us-east1 --stream-logs --runtime-version 1.2 --job-dir ${GCS_JOB_DIR} --module-name trainer.task --package-path trainer/ -- --train-files $TRAIN_FILE --eval-files $EVAL_FILE --train-steps $TRAIN_STEPS --eval-steps 100` – rhaertel80 Aug 10 '17 at 15:34