Bug 1285318

Summary:	pcs/pcsd communication redundancy inside a cluster
Product:	Red Hat Enterprise Linux 7	Reporter:	Tomas Jelinek <tojeline>
Component:	pcs	Assignee:	Tomas Jelinek <tojeline>
Status:	CLOSED WONTFIX	QA Contact:	cluster-qe <cluster-qe>
Severity:	unspecified	Docs Contact:
Priority:	medium
Version:	7.2	CC:	cfeist, cluster-maint, pzimek, tojeline
Target Milestone:	rc
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2018-11-13 09:20:24 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:
Bug Depends On:	1183103
Bug Blocks:

Description Tomas Jelinek 2015-11-25 11:55:28 UTC

Currently pcs/pcsd talks to other cluster nodes using ring0 addresses only. When RRP is configured in a cluster and a node is not available using its ring0 address, a ring1 address should be used to reach the node.

Issues to think about:
* authorization to different addresses of a node
* "pcs status pcsd" should probably display status for all nodes' addresses

This has been already requested on a mailing list:
http://clusterlabs.org/pipermail/users/2015-November/001860.html

Comment 2 Tomas Jelinek 2018-11-13 09:20:24 UTC

This requires a major overhaul of code related to node to node communication and authentication. It has partially been done in pcs-0.10 branch, so it is possible to follow up with this functionality there (bz1649255). The pcs-0.10 branch cannot be included in RHEL7 as it does not support Corosync 2.x and Pacemaker 1.x. Backporting to pcs-0.9 will not happen as there are fundamental changes which are not compatible with pcs-0.9.