Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1834588

Summary:	[RHOSP-13][Octavia] Active/Standby amphora are not replaced when failing
Product:	Red Hat OpenStack	Reporter:	Nilesh <nchandek>
Component:	openstack-octavia	Assignee:	Carlos Goncalves <cgoncalves>
Status:	CLOSED DUPLICATE	QA Contact:	Bruna Bonguardo <bbonguar>
Severity:	high	Docs Contact:
Priority:	high
Version:	13.0 (Queens)	CC:	cgoncalves, ihrachys, lpeer, majopela, sboyron, scohen
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	All
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2020-05-15 14:51:06 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Comment 1 Carlos Goncalves 2020-05-15 10:38:46 UTC

I could not find log entries in the Octavia Health Manager service from any of those two amphorae. The amphorae are expected to report status (heartbeat messages) every 10 seconds in a round-robin way between the three health manager service instances but that is not happening here. Without the Health Manager service knowing about the existence of these amphorae, it will never trigger an amphora failover. Such explains why the customer is not seeing amphora failovers when one of the two amphorae is shut off in Nova. This usually means that whichever of three Octavia controller services (Worker, Health Manager, Housekeeping) that provisioned the amphorae did not inject a valid "controller_ip_port_list" value to the amphora. Specifically to this case where active-standby (tech preview) *and* spare amphora pool (tech preview) are both enabled, the amphorae were provisioned by the Octavia Housekeeping service.

In practical terms, this should mean the load balancer is in ACTIVE provisioning status but in OFFLINE operating status -- could you please double-check?

If we confirm that state, the Octavia Health Manager is not aware of those amphorae and thus will not trigger an amphora failover. This has historically translated to a misconfiguration in the Octavia configuration files created by the installer. We believe it was fixed in https://bugzilla.redhat.com/show_bug.cgi?id=1709925 and released in OSP 13 z11. The customer is running OSP 13 z10.

Comment 2 Sébastien Boyron 2020-05-15 14:39:19 UTC

You are right, the customer does have an OFFLINE operating status.

This confirm your hypothesis. 

I informed the customer and he will update to z11.

Comment 3 Carlos Goncalves 2020-05-15 14:51:06 UTC

Sébastien and I agreed to close this BZ.

*** This bug has been marked as a duplicate of bug 1709925 ***