Why Is My Interactive SSH Job Pending?
Interactive SSH jobs may remain in a Pending state when quotas are exceeded or requested resources are unavailable. This page explains how to check the reason in the CLI and common causes.
Investigating Pending Jobs in the CLI
To see why your job is Pending:
$ runai-bgu describe <job-name>
Common Reasons and Fixes
Exceeding User Quota
Rule: For interactive SSH jobs, each user has a limited CPU, memory, and GPU quota. If you request more than allowed, the job will stay Pending.
Fix:
-
Request resources within your quota.
-
For heavier workloads, submit a training job instead.
Example 1. Example of a job stuck in Pending due to GPU quota exceeded
Submitting a job with too many GPUs → job stays Pending
$ runai-bgu submit pycharm -n my-new-work -c 2 -m 4Gi -g 4
Creating workspace my-new-work...
Waiting for workload to appear...
Waiting for workload to run...
Error: Timeout while waiting for workload my-new-work to run. Current status: PendingDescribing the job → shows quota error in the Events section
$ runai-bgu describe my-new-work
Name: my-new-work
Phase: Pending
Project: myproject
...
Events
2025-08-25T08:20:24Z Unschedulable runai-scheduler Node-Pool 'default': NonPreemptibleOverQuota:
Workload requested 4 GPUs, but quota is 2 GPUs.
Use a preemptible workload to go over quota.Resources Are Not Available for Project
If your project is restricted to CPU-only, and you request GPUs, the job will remain Pending. Likewise, in a GPU-only project, CPU-only jobs will not start.
Fix:
-
Match your job request to your project type.
-
Remove invalid GPU/CPU flags, or request your admin to adjust project settings.