Oracle Linux: PaceMaker Cluster Fails To Authorize: "Error: Unable to communicate with <node>" (Doc ID 2393647.1)

Last updated on MAY 04, 2023

Applies to:

Symptoms

Configuring a PaceMaker cluster fails because the authorization does not complete.  The firewalld ports had been opened and authorization continued to fail.  Authorization still failed with the firewall disabled.

$ sudo pcs cluster auth node1 node2
Username: hacluster
Password:
Error: Operation timed out
node1: Authorized
Error: Unable to communicate with node2

Node2 responds to ping(1) attempts and nc(1) can connect to host2 port <port>.  Connection to the authorization still fails by timing out:

$ unset http_proxy
$ unset https_proxy
$ curl --insecure https://node2:<port>/remote/check_auth

curl: (28) Operation timed out after 300363 milliseconds with 0 out of 0 bytes received

# unset http_proxy
# unset https_proxy
# curl --insecure https://node2:<port>/remote/check_auth
{"notauthorized":"true"}

Changes

 This is a new deployment of a two-node Corosync/Pacemaker cluster.

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.