openstack/octavia
Code Issues Proposed changes

Fix the amphora base port coming up

Browse Source 
A recent change[1] broke the base port IP address coming up in the
amphora. This would cause active/standby and single topology amphora
with members on the VIP subnet to fail.

This patch resolves this issue by not flushing the eth1:0 address.

Story: 2005383
Task: 30368

[1] https://review.openstack.org/#/c/648504/

Change-Id: I52e7e9f172b7783bae09be76cc137f4e7198165f
This commit is contained in:
Michael Johnson 2019-04-05 11:24:05 -07:00
parent a728bc000f
commit 41ff43131f
2 changed files with 17 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
octavia/amphorae/backends/agent/api_server/osutils.py
View File

@ -199,7 +199,7 @@ class BaseOS(object):
return host_routes
@classmethod
def _bring_if_up(cls, interface, what):
def _bring_if_up(cls, interface, what, flush=True):
# Note, we are not using pyroute2 for this as it is not /etc/netns
# aware.
# Work around for bug:
@ -215,9 +215,10 @@ class BaseOS(object):
out = subprocess.check_output(int_up.split(),
stderr=subprocess.STDOUT)
LOG.debug(out)
out = subprocess.check_output(addr_flush.split(),
stderr=subprocess.STDOUT)
LOG.debug(out)
if flush:
out = subprocess.check_output(addr_flush.split(),
stderr=subprocess.STDOUT)
LOG.debug(out)
out = subprocess.check_output(cmd.split(),
stderr=subprocess.STDOUT)
LOG.debug(out)
@ -248,7 +249,7 @@ class BaseOS(object):
cls._bring_if_down(secondary_interface)
cls._bring_if_up(primary_interface, 'VIP')
if secondary_interface:
cls._bring_if_up(secondary_interface, 'VIP')
cls._bring_if_up(secondary_interface, 'VIP', flush=False)
def has_ifup_all(self):
return True

11
releasenotes/notes/fix-eth1-ip-flush-7fadda4bdca387b5.yaml Normal file
View File

@ -0,0 +1,11 @@
---
upgrade:
- |
To fix the issue with active/standby load balancers or single topology
load balancers with members on the VIP subnet, you need to update the
amphora image.
critical:
- |
Fixed a bug where active/standby load balancers and single topology
load balancers with members on the VIP subnet may fail. An updated
image is required to fix this bug.