Just for archival purpose, this issue is now worked on at gh
https://github.com/corosync/corosync/issues/527
Hi Corosync Specialists!
I have a production cluster with two nodes (node0/1). And I have setup
for debugging this issue a completely virtual cluster also.
Both are showing the same pattern that I do not understand:
Printing link status.
Local node ID 0
LINK ID 0
addr = 192.168.2.132
status = 33
LINK ID 1
addr = 192.168.1.132
status = *n*3
What is this "n" indicating? The "n" occurs always at the second ring,
independent of the odering of the interfaces, oder the ring/IP
association, or the interface states.
Without the -b switch the status becomes even more unclear.
Printing link status.
Local node ID 0
LINK ID 0
addr = 192.168.2.132
status:
node 0: link enabled:1 link connected:1
node 1: link enabled:1 link connected:1
LINK ID 1
addr = 192.168.1.132
status:
node 0: link enabled:0 link connected:1
node 1: link enabled:1 link connected:1
What is an "enabled" vs. an "connected" link? At first I thought about
something like spanning tree, where some interfaces are deliberately are
shut down to prevent circles. But this does not correlate with my
findings when I disabled interfaces.
If I disable the interface 192.168.2.132 on Node0 I get on Node0
Local node ID 0
LINK ID 0
addr = 192.168.2.132
status:
node 0: link enabled:1 link connected:1
node 1: link enabled:1 link connected:0
LINK ID 1
addr = 192.168.1.132
status:
node 0: link enabled:0 link connected:1
node 1: link enabled:1 link connected:1
while I get on node1
Printing link status.
Local node ID 1
LINK ID 0
addr = 192.168.2.134
status:
node 0: link enabled:1 link connected:0
node 1: link enabled:1 link connected:1
LINK ID 1
addr = 192.168.1.134
status:
node 0: link enabled:1 link connected:1
node 1: link enabled:0 link connected:1
This is awkward. I had assumed that both nodes indicate for Link0
node 0: link enabled:1 link connected:0
Any help appreciated.
Volker
corosync.conf
totem {
version: 2
cluster_name: mail
token: 3000
token_retransmits_before_loss_const: 10
clear_node_high_bit: yes
crypto_cipher: none
crypto_hash: none
interface {
linknumber: 0
knet_transport: udp
knet_link_priority: 20
}
interface {
linknumber: 1
knet_transport: udp
knet_link_priority: 10
}
}
nodelist {
node {
ring1_addr: 192.168.1.132
ring0_addr: 192.168.2.132
nodeid: 0
name: mail3
}
node {
ring1_addr: 192.168.1.134
ring0_addr: 192.168.2.134
nodeid: 1
name: mail4
}
}
_______________________________________________
Manage your subscription:
https://lists.clusterlabs.org/mailman/listinfo/users
ClusterLabs home: https://www.clusterlabs.org/
_______________________________________________
Manage your subscription:
https://lists.clusterlabs.org/mailman/listinfo/users
ClusterLabs home: https://www.clusterlabs.org/
