[Docs] Storage network with Ceph
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Won't Fix
|
High
|
Fuel Documentation Team | ||
6.0.x |
Won't Fix
|
Medium
|
Fuel Documentation Team | ||
6.1.x |
Won't Fix
|
Undecided
|
Fuel Documentation Team | ||
7.0.x |
Won't Fix
|
Undecided
|
Fuel Documentation Team | ||
8.0.x |
Won't Fix
|
Medium
|
Fuel Documentation Team | ||
Future |
Won't Fix
|
Undecided
|
Fuel Documentation Team | ||
Mitaka |
Won't Fix
|
Medium
|
Maria Zlatkova |
Bug Description
Documentation is not really clear about how the storage network is used with Ceph. Is it used for :
- Ceph replication?
- Communication with the monitors?
- Communnication between a VM and it's volumes?
Storage network with Ceph is used for Ceph OSD replication only.
VM to Cinder/Ceph traffic, communucation between monitors and Cinder to Ceph traffic served by Management network.
With minimal recommended for production Ceph replication factor=3, Ceph have to replicate twice data more than it was written by Ceph clients.
So, taking to account additional load from other services to Management network, such settings make close to equal load to both, Managenment and Storage networks, plus it offloads OpenStack network in case of Ceph data re-balance and PG backfills when some disaster happened with part of existing OSDs.
Changed in fuel: | |
milestone: | none → 6.1 |
status: | New → Confirmed |
tags: | added: low-hanging-fruit |
Changed in fuel: | |
importance: | Undecided → Medium |
no longer affects: | fuel/6.0-updates |
Changed in fuel: | |
status: | Confirmed → Triaged |
Changed in fuel: | |
milestone: | 6.1 → 8.0 |
tags: | added: area-docs |
Changed in fuel: | |
milestone: | 8.0 → 9.0 |
status: | Triaged → New |
Changed in fuel: | |
status: | New → Incomplete |
tags: | removed: need-info |
Changed in fuel: | |
milestone: | 9.0 → 10.0 |
Changed in fuel: | |
status: | Incomplete → Won't Fix |
This is actually high priority item, as many users hit this to some or another way. We need to ensure it's documented in 6.1.