Bug 1200473

Summary:	MongoDB - Fix bug with reconnection to new master node
Product:	Red Hat OpenStack	Reporter:	Eoghan Glynn <eglynn>
Component:	openstack-ceilometer	Assignee:	Chris Dent <chdent>
Status:	CLOSED ERRATA	QA Contact:	Yurii Prokulevych <yprokule>
Severity:	high	Docs Contact:
Priority:	high
Version:	6.0 (Juno)	CC:	ajeain, dnavale, eglynn, jruzicka, pbrady, rrosa, yeylon
Target Milestone:	z3	Keywords:	ZStream
Target Release:	6.0 (Juno)
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	openstack-ceilometer-2014.2.3-3.el7	Doc Type:	Bug Fix
Doc Text:	Previously, OpenStack Telemetry did not handle a specific exception (AutoReconnect) when MongoDB ReplicaSet was used. As a result, reconnections did not work properly. With this update, built-in handling in the 'pymongo' client is used to deal with AutoReconnect gracefully. As a result, connection retry is handled in an acceptable fashion in both single and ReplicaSet installations.	Story Points:	---
Clone Of:		Environment:
Last Closed:	2015-05-05 13:10:47 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 3 Chris Dent 2015-04-14 15:27:45 UTC

This hasn't been a clean backport: reconnection strategies evolved in the intervening time causing conflicts in the main code and failing tests (which relied on a previous interface). I've made the necessary adjustments in the patch (which will link up here soon after I've got local unit tests passing) but I would recommend that the patch and the resulting packaged stuff get a somewhat more than normal review.

Comment 9 Chris Dent 2015-04-17 16:20:16 UTC

I haven't got access to a replicaset to test this but it is possible to test that the reconnection functionality works in a relatively straightforward fashion:

* do an installation (packstack --allinone is what I did)
* authenticate (`. keystonerc_admin `)
* confirm meters with `ceilometer meter-list`
* stop mongod `sudo service stop mongod`
* try meters again, get failure
* start mongod `sudo service start mongod`
* try meters again, get some

The reason for doing this, even in a non-replicaset setup, is that the backport completely replaces previous reconnection logic, so doing this confirms that the new stuff has the basic expected functionality.

Comment 10 Chris Dent 2015-04-22 19:15:52 UTC

Added some DocText, but note this is still ON_QA, for now, and though my own testing has shown that "it works" confirmation is pending.

Comment 11 Yurii Prokulevych 2015-04-24 17:25:10 UTC

Verified using replicaset with 3 members.

Comment 13 errata-xmlrpc 2015-05-05 13:10:47 UTC

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHBA-2015-0930.html