Helo, I’m using nomad to schedule my docker container. Through reserved, nomad is scheduled on the 32-191 CPU core. In the job configuration, I configured the CPU to be 500. When I started 150 containers on a node, I found that the containers were basically working on the 32 core, resulting in the CPU utilization rate of the 32 core being 100%, and the 33-191 core load was almost zero. This also led to the docker becoming very stuck after creating dozens of containers, and it was almost impossible to create new containers.
Related topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
Nomad 1.5.6 set cpu values, but can not share cpu cores | 0 | 181 | August 4, 2023 | |
Nomad 1.5.6 can not share cpu cores | 2 | 315 | August 3, 2023 | |
Docker driver core reservation behaviour in 1.1 | 1 | 399 | June 3, 2021 | |
How do I set nomad cpu oversubscription? | 1 | 482 | August 22, 2023 | |
What are recommended CPU settings for task resources? | 1 | 468 | March 29, 2022 |