Troubleshooting
Lost/unknown job ID
The first step in any investigation is knowing a job ID. If you started your job with neuro run, the job's ID was printed in the output.
However, if you can't find the initial terminal output, you can use one of these commands to find a specific job:
neuro ps prints only running jobs.
neuro ps -a prints all jobs.
neuro ps -s failed prints all jobs with the Failed status.
Image build failed
When you run neuro-flow build IMAGE_NAME, neuro-flow uploads the build context to the platform and creates a platform job that uses Kaniko to build a docker image and push it to the platform registry.
If building fails, you can check the job's status and logs to get more information.
Getting a job's status
To check a job's status, run:
The Status transitions section in the output can help you learn at which step the job failed.
Getting builder logs
To check builder logs, run:
neuro run / neuro-flow run failed
There are a few main reasons your job may fail. Here are some of the most common:
Incorrect image name
This can happen if you have a typo in the image name or if the specified image was not built before running a job. List of all images can be accessed by running neuro image ls. You can also list tags for a particular image via neuro image tags <IMAGE_URI>.
Incorrect volume mounted
You might have an invalid volume mounted to the job. For example, you've mounted a volume to the /my-project folder, but your code expects /my_project. You can double-check it in the logs.
Cluster scale up failed
If you see a Cluster Scale Up Failed error in the status, it usually means you’ve requested resources that are not available in the cluster at the moment. For example, all GPUs are busy, so your job can’t be scheduled.
Code issues
You may have an error in your python script that prevents the job from running properly.
Can’t access my job via HTTP
There are a few steps to troubleshooting such issues.
Checking for an open HTTP port
The first point of interest is whether you have an open HTTP port for your job. To check this, you can:
Use the --http_port parameter.
Checking the listening IP
Next, make sure that your web app listens on 0.0.0.0, not on 127.0.0.1 or localhost — otherwise it won't be able to accept incoming requests from the outside of the container.
Disabling HTTP authentication
And finally, if you can access your job via browser, but curl and similar tools don’t work, most likely you didn’t disable HTTP authentication. The Neu.ro platform puts an HTTP authentication layer in front of your app by default for security reasons.
You can disable this behavior manually when running jobs:
Use the --no-http-auth parameter.
Troubleshooting a running job
Just like with Docker, you can get a shell in a running job to check its state. To do this, run:
Note: In Docker you would typically add the -it parameters to the command, but they’re not necessary for neuro exec.
Navigating job statuses
A job might have one of the following statuses:
Status | Description |
pending | The job is being created and scheduled. This includes finding (and possibly waiting for) sufficient amount of resources, pulling an image from a registry, etc. |
suspended | The job's execution was temporarily suspended |
running | The job is running |
succeeded | The job terminated with the exit code 0 |
cancelled | The running job was manually terminated/deleted |
failed | The job terminated with a non-0 exit code. |
Each of these statuses have additional sub-statuses that can help you monitor the job and trace errors on an more granular level:
Sub-status | Description |
PodInitializing | Initializing a pod for the job |
ContainerCreating | Creating a container for the job |
ErrImagePull | Failed to pull the specified image |
ImagePullBackOff | Stopping image pulling |
InvalidImageName | Incorrect image name was specified |
OOMKilled | Job terminated due to an Out Of Memory error |
Completed | Job is completed |
Error | An error occured |
ContainerCannotRun | Cannot run the container |
Creating | Creating the job |
Collected | Job collected |
Scheduling | Scheduling the job execution |
NotFound | The job could not be scheduled or was preempted |
ClusterNotFound | Specified cluster was not found |
ClusterScalingUp | Scaling up |
ClusterScaleUpFailed | Failed to scale up the cluster |
Restarting | Restarting the job |
DiskUnavailable | Specified disk is currently unavailable |
QuotaExhausted | User quota was exhausted - you will need to renew it to perform more jobs |
LifeSpanEnded | The job has reached the end of its lifespan |
UserRequested | The job was terminated per user request |
Last updated