Allow compute nodes to use shared storage provider for DISK_GB resources
Allow compute node resource providers to use DISK_GB inventory configured in a shared storage resource provider by using aggregate relationship.
Shared resource providers are marked with a standard trait called `MISC_SHARES_
For example, assume we have two compute nodes, CN_1 and CN_2, each with inventory of VCPU and MEMORY_MB but not DISK_GB (in other words, these are compute nodes with no local disk). There is a resource provider called
"NFS_SHARE" that has an Inventory of DISK_GB and has the MISC_SHARES_
Earlier, in Rocky release, an attempt was made to add this support but it wasn't a clean fix for LP bug 1784020.
Blueprint information
- Status:
- Started
- Approver:
- None
- Priority:
- Undefined
- Drafter:
- Tushar Patil
- Direction:
- Needs approval
- Assignee:
- Shilpa Devharakar
- Definition:
- Pending Approval
- Series goal:
- None
- Implementation:
-
Started
- Milestone target:
- None
- Started by
- Eric Fried
- Completed by
Related branches
Related bugs
| Bug #1414432: nova hypervisor-stats shows wrong disk usage with shared storage | Confirmed |
| Bug #1824858: nova instance remnant left behind after cold migration completes | Fix Released |
Sprints
Whiteboard
Gerrit topic: https:/
Addressed by: https:/
Allow compute nodes to use DISK_GB from shared storage RP
Gerrit topic: https:/
Gerrit topic: https:/
Addressed by: https:/
Allow compute nodes to use DISK_GB from shared storage RP
Deferring this out of Train since it wasn't approved and we're about 2 weeks from feature freeze. -- mriedem 20190829
[efried 20200213] My understanding is that we've agreed this needs a Placement API change, and that's not going to be possible in U, so deferring.
