Api Server uve shown as Non-functional on some setups

Bug #1372875 reported by Vedamurthy Joshi
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Juniper Openstack
Status tracked in Trunk
R1.1
Fix Committed
High
Sundaresan Rajangam
Trunk
Fix Committed
High
Sundaresan Rajangam

Bug Description

R1.10 34

On one of the setups in the field, API Server UVE on one of the three config nodes was shown as Non-functional
Its connection to Discovery was shown as Down with reason "Heartbeat -Connection Error"

Due to this, contrail-status would report the api-server as initializing.

Sundar debugged the issue and found that once a connection to Discovery is set to Down initially, on reconnecting to the discovery, the status is not updated to Up and continues to be marked Down forever.
Workaround would be to restart API Server

Tags: analytics
Revision history for this message
OpenContrail Admin (ci-admin-f) wrote : A change has been merged

Reviewed: https://review.opencontrail.org/3045
Committed: http://github.org/Juniper/contrail-controller/commit/d520c323c44d07e011d9a83e55e016eb253d929d
Submitter: Zuul
Branch: master

commit d520c323c44d07e011d9a83e55e016eb253d929d
Author: Sundaresan Rajangam <email address hidden>
Date: Tue Sep 23 12:31:56 2014 -0700

(1) When Heart beat to the discovery server fails [haproxy down/not started], the discovery client library marks the Discovery connection as DOWN.
However, on successful Heart beat, the connection status is not being marked UP. Also, reset the connection status to DOWN when the response code for
HeartBeat request is not 200.
(2) Increasing the timeout for introspect connection in contrail-status to 0.5 seconds.
In swisscom setup, changing the timeout from 0.01 to 0.1 did not work. Did not see much load on the system. Changing the timeout to 0.5 to be on the safer side.
Closes-Bug: #1372875

Change-Id: Ie8e14e698ccfec537827c95b055681c8268a5fba

Revision history for this message
OpenContrail Admin (ci-admin-f) wrote :

Reviewed: https://review.opencontrail.org/3064
Committed: http://github.org/Juniper/contrail-controller/commit/26810947dec22677f75bbe0579b960ccf38aec5d
Submitter: Zuul
Branch: R1.10

commit 26810947dec22677f75bbe0579b960ccf38aec5d
Author: Sundaresan Rajangam <email address hidden>
Date: Wed Sep 24 01:35:05 2014 -0700

(1) When Heart beat to the discovery server fails [haproxy down/not started], the discovery client library marks the Discovery connection as DOWN.
However, on successful Heart beat, the connection status is not being marked UP. Also, reset the connection status to DOWN when the response code for
HeartBeat request is not 200.
(2) Increasing the timeout for introspect connection in contrail-status to 0.5 seconds.
In swisscom setup, changing the timeout from 0.01 to 0.1 did not work. Did not see much load on the system. Changing the timeout to 0.5 to be on the safer side.
Closes-Bug: #1372875

Change-Id: I90fa290beac079b25781d5c2ac2db0d5193c0828

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.