Bug 1527660 - [Docs] Document how to override containerized Ceph OSD memory and CPU usage
Summary: [Docs] Document how to override containerized Ceph OSD memory and CPU usage
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Red Hat OpenStack
Classification: Red Hat
Component: documentation
Version: 12.0 (Pike)
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
: ---
Assignee: Kim Nylander
QA Contact: Dan Macpherson
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2017-12-19 19:30 UTC by John Fulton
Modified: 2018-01-31 15:13 UTC (History)
8 users (show)

Fixed In Version: 12.0
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2018-01-31 15:13:50 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description John Fulton 2017-12-19 19:30:34 UTC
The OSP12 documents should describe how to override containerized Ceph OSD memory and CPU usage. A proposal to do this is below. 


In the OSP12 Ceph document [1] would you please add a new section 5.3 "Reserving Memory and CPU resources for Ceph" with content like the following: 

"""
By default each containerized OSD is limited to 1 GB of RAM and 1 vCPU. These values are based on the minimum recommendations [3] but may not be appropriate depending on hardware. These settings may be overridden using an example like the following which reserves 3 GB of RAM and 1 vCPUs per OSD:

parameter_defaults:
  CephAnsibleExtraConfig:
    ceph_osd_docker_memory_limit: 3g
    ceph_osd_docker_cpu_limit: 1

To determine appropriate values for the above settings, see the Red Hat Ceph Storage Hardware Guide [4].
"""

Please also update the HCI document for OSP12 [2] to have a section 4.3 "Reserving Memory and CPU resources for Ceph" which also contains the same content or links to new section described above. 

[1] https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/12/html/deploying_an_overcloud_with_containerized_red_hat_ceph/configuring_ceph_storage_cluster_settings

[2] https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/12/html/hyper-converged_infrastructure_guide/resource-isolation#resource-isolation-cpu-ram

[3]  https://access.redhat.com/documentation/en-us/red_hat_ceph_storage/2/html/red_hat_ceph_storage_hardware_guide/minimum_recommendations

[4] https://access.redhat.com/documentation/en-us/red_hat_ceph_storage/2/html-single/red_hat_ceph_storage_hardware_guide/index#intel_hardware_guide

Comment 2 John Fulton 2018-01-04 19:06:09 UTC
I've introduced the comments I am requesting, and others, in the upstream document for review at: https://review.openstack.org/#/c/531214

Comment 4 John Fulton 2018-01-04 19:58:27 UTC
I request the following update to this docbug:

"""
Each containerized OSD may be limited in GB of RAM and vCPUs through the `docker run` command's `--memory` and `--cpu-quota` options. To override these values use an example like the following which reserves 3 GB of RAM and 1 vCPUs per OSD:

parameter_defaults:
  CephAnsibleExtraConfig:
    ceph_osd_docker_memory_limit: 3g
    ceph_osd_docker_cpu_limit: 1

To determine appropriate values for the above settings, see the Red Hat Ceph Storage Hardware Guide.

https://access.redhat.com/documentation/en-us/red_hat_ceph_storage/2/html-single/red_hat_ceph_storage_hardware_guide/index#intel_hardware_guide
"""


Note You need to log in before you can comment on or make changes to this bug.