I had the following cloud environment paused:
- Hail, no startup script
- Master: 8 CPUs, 30 mem, 500 GB
- Workers: 2 workers, 120 preempts, 8 CPUs, 30 mem, 100 GB
I deleted the notebook I started the environment on and uploaded a more up-to-date version of the notebook. When I went into the new notebook and tried to unpause the compute, it was stuck on the "starting notebook" config for about 20 minutes (even after a few refreshes), at which point I deleted it and created a new compute. I have a hunch that this has something to do with the fact it's a spark cluster, but I'm not sure. This took place around 3 PM Eastern, before Terra's 4:00-4:30 PM maintenance time
It wasn't a blocker for me, but this might be tricky for other users, especially if they have unsaved analysis that hasn't been saved to the bucket.
Please sign in to leave a comment.