Feature #51566
closed
Added by Sebastian Wagner almost 3 years ago.
Updated about 2 years ago.
Description
ceph-ansible allowed customization of CPU limits for containerized daemons per each type (so you could have different setting for MON, OSD, MGR, etc)
cephadm should offer same option.
This is a follow-up on the osd-memory-target feature
To be done here:
- You have to extend the Cephadm's data structure to contain a cpu limit per service
- Needs be verified that the limit is sane
- you need a UX component (CLI, orchestrator interface, you name it)
- Needs to be discussed in an orchestrator weekly call for the technology
- default value should be none. (no limit)
- just expose what podman gives you.
- then you need to add this limit to the daemon deployment (has to be transferred from mgr/cephadm module to the actual daemon as deployed by the /src/cephadm)
- and the you should report this limit
ceph orch ps (plain and yaml)
- Then you need to add either unit testing or integration testing
- Status changed from New to Fix Under Review
- Assignee set to Adam King
- Pull request ID set to 44517
- Status changed from Fix Under Review to Pending Backport
- Pull request ID changed from 44517 to 44459
- Status changed from Pending Backport to Resolved
Also available in: Atom
PDF