Already-scheduled instances should be rescheduled after attaching a new compute node
Bug #719592 reported by
Christian Berendt
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| OpenStack Compute (nova) |
Won't Fix
|
Wishlist
|
Unassigned | ||
Bug Description
A spawned a lot of new instances (>100). They got scheduled on the already attached compute nodes. After attaching a new compute node there went none of the scheduled instanced to this node. I think it would be useful to reschedule the already scheduled instances after attaching a new compute node to direct use the new node.
I think this is a bug, because if I see a resource problem in my openstack environment and I try to resolve this problem by attaching a new compute node I have no advancement at the moment.
Revision history for this message
| Thierry Carrez (ttx) wrote | #1 |
| Changed in nova: | |
| status: | New → Incomplete |
Revision history for this message
| Thierry Carrez (ttx) wrote | #2 |
| summary: |
- no rescheduling after attaching a new compute node + Already-scheduled instances should be rescheduled after attaching a new + compute node |
| Changed in nova: | |
| importance: | Undecided → Wishlist |
| status: | Incomplete → Confirmed |
Revision history for this message
| Mark McLoughlin (markmc) wrote | #3 |
| Changed in nova: | |
| status: | Confirmed → Won't Fix |
To post a comment you must log in.
By design, already-scheduled instances means instance start requests have been queued for specific compute nodes. Support for rescheduling means we'd need to keep some extra state to allow rewriting history... I'm not convinced by the net gain, but i'll let scheduler experts comment too :)