- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
02-13-2023 10:06 PM
when we run databricks job it take some time to get job cluster active .
I created pool also and attached with job cluster but still it take time to attached the cluster and job cluster get active to start the job run.
is there any way - we can run databricks job instantly without waiting . ( all purpose cluster can use with all time active minimum work node but all purpose is not advisable to use in production ) .
So is there any way using on demand job cluster - can minimize the waiting time and run job instantly
- Labels:
-
Databricks Job
-
Instance Pool
-
Job
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
02-13-2023 10:44 PM
Hi @vinay kumar ,
We can use serverless cluster to run the job instantly for now serverless cluster is not available in all the region but in future we can use this.
The first-time pool cluster take time for spin up.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
02-13-2023 10:58 PM
Thanks @Ajay Pandey for your quick response , pool cluster taking 60-65 seconds uptime every time .
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
02-14-2023 05:06 AM
If you want instant processing, you will have to have a cluster running all the time.
As mentioned above, Databricks is testing serverless compute for data engineering workloads (comparable to serverless SQL). This fires up a cluster in a few seconds, but is not yet GA.
The closest you can get at the moment is cluster pools.