Bug 1469441 - [RFE][Swift] Set partition power based on number of disks
Summary: [RFE][Swift] Set partition power based on number of disks
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Red Hat OpenStack
Classification: Red Hat
Component: puppet-swift
Version: 14.0 (Rocky)
Hardware: Unspecified
OS: Unspecified
low
low
Target Milestone: ---
: ---
Assignee: Christian Schwede (cschwede)
QA Contact: Mike Abrams
URL:
Whiteboard:
Depends On: 1469435
Blocks:
TreeView+ depends on / blocked
 
Reported: 2017-07-11 09:35 UTC by Christian Schwede (cschwede)
Modified: 2020-08-24 12:21 UTC (History)
6 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2020-08-24 12:21:46 UTC
Target Upstream Version:
Embargoed:
cschwede: needinfo-


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
OpenStack gerrit 529708 0 None ABANDONED Add Ansible playbook & Mistral workflow for Swift to update ring devices 2020-08-21 13:20:48 UTC

Description Christian Schwede (cschwede) 2017-07-11 09:35:41 UTC
As an operator with a lot of storage nodes/disks, I want the partition power for Swift to be larger than today.

Today the partition power in director installations is fixed to 10. This is most 
likely fine for most standard installations; however if there is a large number of disks and/or storage nodes, it might not be sufficient and should be larger by default. There is an option to define this during the initial deployment, but operators might not be aware of this and therefore it should be set based on the number of available disks.

See https://access.redhat.com/articles/2745961 for some best practice numbers.

Comment 5 Christian Schwede (cschwede) 2020-08-24 12:21:46 UTC
Closing this BZ. We have a doc section covering this (https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/16.0/html/deployment_recommendations_for_specific_red_hat_openstack_platform_services/svc-objectstorage#svc-objectstorage-partition), and in case of very large Swift clusters operators need to decide which partition power is the right one. This should likely not be automated. Additionally, the upstream bug has been abandoned as well.


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