subnet allocation does not scale
Bug #1382149 reported by
Robert Kukura
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Group Based Policy |
Triaged
|
High
|
Robert Kukura |
Bug Description
In gbp/neutron/
# REVISIT(rkukura): This is a temporary allocation algorithm
# that depends on an exception being raised when the subnet
# being created is already in use. A DB allocation table for
# the pool of subnets, or at least a more efficient way to
# test if a subnet is in-use, may be needed.
The current trial-and-error approach slows down as the L3P's supernet becomes exhausted, and also results in lots of errors being logged. We probably need to implement allocation pools using the DB.
Changed in group-based-policy: | |
importance: | Undecided → High |
importance: | High → Medium |
description: | updated |
Changed in group-based-policy: | |
assignee: | nobody → Robert Kukura (rkukura) |
milestone: | none → juno-release |
Changed in group-based-policy: | |
status: | New → Confirmed |
Changed in group-based-policy: | |
milestone: | juno-release → kilo-gbp-1 |
Changed in group-based-policy: | |
milestone: | kilo-gbp-1 → kilo-gbp-2 |
Changed in group-based-policy: | |
milestone: | kilo-gbp-2 → kilo-gbp-3 |
Changed in group-based-policy: | |
milestone: | kilo-gbp-3 → kilo-gbp-4 |
importance: | Medium → High |
Changed in group-based-policy: | |
milestone: | kilo-gbp-4 → next |
Changed in group-based-policy: | |
status: | Confirmed → Triaged |
To post a comment you must log in.
I propose that we defer fixing this until our kilo-based release, and resolve it by having the resource_mapping driver map l3 policies to subnet pools as described in https:/ /review. openstack. org/#/c/ 135771/.