You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I wonder if it isn't feasible to have operation mode that does not require a job registry (nor a job tracker to keep that job registry up to date): just query the cluster API directly to get the state of a job.
This mode would probably not be recommended for production use cases, but having it available could lower the bar considerably to get a working deployment for testing purposes.
It could also provide a fallback code path when the EJR code path is broken for some reason (EJR down, job tracker not running)
The text was updated successfully, but these errors were encountered:
soxofaan
changed the title
Operation mode with batch job support but without a job tracker
Operation mode with batch job support but without a job registry or job tracker
Jan 17, 2025
A functional deployment with full batch job support currently requires a job registry (e.g. EJR) and a regular (e.g. minutely) run of the job tracker.
These extra components can make it quite challenging to get a working deployment (e.g. in EOEPCA https://github.com/EOEPCA/roadmap/issues/332).
I wonder if it isn't feasible to have operation mode that does not require a job registry (nor a job tracker to keep that job registry up to date): just query the cluster API directly to get the state of a job.
This mode would probably not be recommended for production use cases, but having it available could lower the bar considerably to get a working deployment for testing purposes.
It could also provide a fallback code path when the EJR code path is broken for some reason (EJR down, job tracker not running)
The text was updated successfully, but these errors were encountered: