Optimize gpe segment allocation
Bug #1817218 reported by
Naveen Joy
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| networking-vpp |
Fix Released
|
Wishlist
|
Naveen Joy | ||
Bug Description
GPE has a large pool of allocatable segment IDs (16+ Million).
The current implementation stores each unallocated segment ID
in the neutron DB for allocation. This could cause DB
issues when the entire range of segments is specified
for allocation.
Optimize segment allocation by storing only the allocated GPE VNIs
in the Neutron DB.
| Changed in networking-vpp: | |
| assignee: | nobody → Naveen Joy (najoy) |
| Changed in networking-vpp: | |
| status: | New → In Progress |
| Changed in networking-vpp: | |
| importance: | Undecided → Wishlist |
Revision history for this message
| OpenStack Infra (hudson-openstack) wrote Fix merged to networking-vpp (master) | #1 |
| Changed in networking-vpp: | |
| status: | In Progress → Fix Released |
To post a comment you must log in.
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit f661e06e4ac3b90
Author: Naveen Joy <email address hidden>
Date: Fri Feb 8 12:54:28 2019 -0800
Optimize GPE VNI allocation & train release support
GPE networks have a large pool of allocatable
Virtual Network Identifiers (VNIs) - 16+ Million. The prior
implementation stores each unallocated VNI in the specified
range, for allocation in the Neutron DB.
This patch optimizes VNI allocation in the DB, by dynamically
computing a valid VNI for allocation at network creation time,
thereby storing only the allocated VNIs in Neutron DB.
Mypy code added; couple of method signatures changes; removed
some hard work that previously generated a log about whether
a released segment was a tenant segment or not.
Also, provides GPE type driver support for the Train
release.
Closes-bug: #1817218
Change-Id: I5413c130103431