Bug 735912

Summary:

Cman RRP - Include mcast per NIC patch and set default threshold to 3

Product:

Red Hat Enterprise Linux 6

Reporter:

Jan Friesse <jfriesse>

Component:

cluster

Assignee:

Fabio Massimo Di Nitto <fdinitto>

Status:

CLOSED ERRATA

QA Contact:

Cluster QE <mspqa-list>

Severity:

medium

Docs Contact:

Priority:

medium

Version:

6.2

CC:

ccaulfie, cluster-maint, jkortus, lhh, rpeterso, teigland

Target Milestone:

Keywords:

TechPreview

Target Release:

6.2

Hardware:

Unspecified

OS:

Unspecified

Whiteboard:

Fixed In Version:

cluster-3.0.12.1-19.el6

Doc Type:

Technology Preview

Doc Text:

Improve integration between cman and corosync for Redundant Ring Protocol

Story Points:

---

Clone Of:

Clones:

917773 (view as bug list)

Environment:

Last Closed:

2011-12-06 15:06:04 UTC

Type:

---

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:

Bug Blocks:

743047, 917773

Attachments:

Description	Flags
proposed patch	none
new patch	none

Description Jan Friesse 2011-09-06 06:55:14 UTC

Description of problem:
1) There is need for corosync RRP to have different mcast address for all NICs
2) Default threshold should be set to more reasonable value 3

Version-Release number of selected component (if applicable):
Newest for 6.2

How reproducible:
See https://bugzilla.redhat.com/show_bug.cgi?id=722469

Additional info:
Threshold key is totem.rrp_problem_count_threshold

Expected results:
corosync-objctl -a | grep rrp_proble
-> 3

Comment 8 Fabio Massimo Di Nitto 2011-09-08 08:28:28 UTC

Created attachment 522060 [details]
proposed patch

Unit test results:

0) no altname, no <cman/>

Multicast addresses: 239.192.99.73

[root@clusternet-node2 ~]# corosync-objctl |grep rrp
totem.rrp_mode=none

1) autoselection:

no <cman/>

      <altname name="clusternet-node1-eth2"/>
      <altname name="clusternet-node2-eth2"/>

Multicast addresses: 239.192.99.73 239.192.99.74

[root@clusternet-node2 ~]# corosync-objctl |grep rrp
totem.rrp_mode=passive
totem.rrp_problem_count_threshold=3

2) force cman main multicast on primary interface

  <cman>
   <multicast addr="239.192.100.1"/>
  </cman>
  <clusternodes>
    <clusternode name="clusternet-node1-eth1" votes="1" nodeid="1">
      <altname name="clusternet-node1-eth2"/>

[root@clusternet-node2 daemon]# cman_tool status
Multicast addresses: 239.192.100.1 239.192.99.74

[root@clusternet-node2 ~]# corosync-objctl |grep rrp
totem.rrp_mode=passive
totem.rrp_problem_count_threshold=3

3) force mcast all over (note that altname need mcast per node!)

  <cman>
   <multicast addr="239.192.100.1"/>
  </cman>

      <altname name="clusternet-node1-eth2" mcast="239.192.100.2"/>
      <altname name="clusternet-node2-eth2" mcast="239.192.100.2"/>

[root@clusternet-node2 daemon]# cman_tool status
Multicast addresses: 239.192.100.1 239.192.100.2

[root@clusternet-node2 ~]# corosync-objctl |grep rrp
totem.rrp_mode=passive
totem.rrp_problem_count_threshold=3

Comment 10 Fabio Massimo Di Nitto 2011-09-08 09:18:48 UTC

Created attachment 522068 [details]
new patch

The previous patch had an error when <totem tag was present in cluster.conf.

Previous unit test is still valid.

Add the following cases:

<totem rrp_mode="active"/>

[root@clusternet-node2 xml]# corosync-objctl |grep rrp
cluster.totem.rrp_mode=active
totem.rrp_mode=active
totem.rrp_problem_count_threshold=3


<totem rrp_mode="active" rrp_problem_count_threshold="10"/>

[root@clusternet-node2 xml]# corosync-objctl |grep rrp
cluster.totem.rrp_mode=active
cluster.totem.rrp_problem_count_threshold=10
totem.rrp_mode=active
totem.rrp_problem_count_threshold=10

Comment 11 Fabio Massimo Di Nitto 2011-09-08 09:21:24 UTC

The "today I just don´t know how to copy/paste"

  <totem rrp_problem_count_threshold="10"/>

[root@clusternet-node2 xml]# corosync-objctl |grep rrp
cluster.totem.rrp_problem_count_threshold=10
totem.rrp_problem_count_threshold=10
totem.rrp_mode=passive

Comment 15 Fabio Massimo Di Nitto 2011-10-03 06:55:15 UTC

    Technical note added. If any revisions are required, please edit the "Technical Notes" field
    accordingly. All revisions will be proofread by the Engineering Content Services team.
    
    New Contents:
Improve integration between cman and corosync for Redundant Ring Protocol

Comment 17 errata-xmlrpc 2011-12-06 15:06:04 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.

http://rhn.redhat.com/errata/RHBA-2011-1516.html