Project

General

Profile

Actions

Bug #12296

closed

[API] [Node Manager] [Docs] Fix recommendations and explanations for various max-nodes configs

Added by Tom Clegg almost 5 years ago. Updated over 2 years ago.

Status:
Closed
Priority:
Normal
Assigned To:
-
Category:
-
Target version:
-
Start date:
09/20/2017
Due date:
% Done:

0%

Estimated time:
Story points:
-

Description

In source:services/api/config/application.default.yml, the explanation of max_compute_nodes is wrong. It suggests that it limits how many nodes will be created, but it doesn't. The correct value is more like "something much bigger than node manager's max-nodes config."

Install docs should recommend TreeWidth=65533 in slurm.conf.

In source:services/nodemanager/doc/gce.example.cfg the max_nodes comment is a copy-paste fail. Others aren't as bad, but they should all say something like "Before raising this, ensure API server's max_compute_nodes config is much larger than the new value, and restart API server if needed to ensure it is using its new config."

# Node Manager will not start any compute nodes when at least this
# running at all times.  By default, these will be the cheapest node size.
max_nodes = 8
Actions #1

Updated by Tom Clegg almost 5 years ago

It is necessary to have

{nodemanager_max_nodes} ≤ {apiserver_max_slots} ≤ {slurm_conf_nodes} ≤ {slurm_conf_tree_width}

Actions #2

Updated by Peter Amstutz over 2 years ago

  • Status changed from New to Closed
Actions

Also available in: Atom PDF