Resource Allocation Logic

JupyterHub
,
1

Hi all,

Can some one point me to how the resource allocation in Hub is done? I assume I can provide some limits while creating. But how does this work.

For example, if I request a notebook for 4GB RAM and 1 GPU( Can’t do fractional) and I’m using only 2GB of RAM, will the 2GB RAM be allocated to another notebook requestor?

2

Hi! It depends on the spawner you’re using. Some spawners take account of your resource requests, others ignore it completely.

3

Hi, What all type of spawners are there? I didn’t see any resource for the same?

All I see is some custom docker images which we can add for each user and resource limits. This dynamic resource allocation is this something new?

4

There are some technical details for how spawners work, including on resource limits, here: Spawners — JupyterHub 2.3.1 documentation