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
As the following log shows, the default number of worker threads used by controller is 1. This value is too small for production environment because it will cause extremely delay of processing Job if there are many Jobs at the same time.
And I found this can't be configured.
1.6717679978118246e+09 INFO Starting workers {"controller": "pytorchjob-controller", "worker count": 1}
1.6717679978129413e+09 INFO Starting workers {"controller": "tfjob-controller", "worker count": 1}
1.6717679978130171e+09 INFO Starting workers {"controller": "mpijob-controller", "worker count": 1}
As the following log shows, the default number of worker threads used by controller is 1. This value is too small for production environment because it will cause extremely delay of processing Job if there are many Jobs at the same time.
And I found this can't be configured.
Referring to point4 of #1703
The text was updated successfully, but these errors were encountered: