Affected Puppet, Ruby, OS and module versions/distributions

• Puppet: 3.8.7
• Ruby: 2.3.1p112
• Distribution: Ubuntu 16.04
• Module version: 3.0.0
• Gluster version: 3.10.3

How to reproduce (e.g Puppet code you use)

  class { ::gluster:
    repo    => false,
    server  => true,
    client  => false,
  }

Exported resources enabled.

What are you seeing

The first gluster node (test-gluster1.local in this example) comes up fine. The second node (test-gluster2.local) comes up and probes test-gluster1.local as expected. test-gluster1.local then completes the handshaking by probing test-gluster2.local.

However, test-gluster2.local seems to believe it needs to continue probing test-gluster1.local on each puppet run:

2017-06-27 15:22:19 +0100 /Stage[main]/Gluster/Gluster::Peer[test-gluster1.local]/Exec[gluster peer probe test-gluster1.local]/returns (notice): executed successfully

On test-gluster2.local, if you run

facter --puppet gluster_peer_list

It shows the ip address of test-gluster1.local rather than the hostname, this is what I believe is causing the peer probe to be called each time as this does not match the gluster::peer resource title in puppetdb

Looking at the code for the gluster_peer_list fact, it determines the peers from the value of Hostname: in the gluster peer status command. If you run this command though, the correct hostname appears under "Other names:"

gluster peer status
Number of Peers: 1

Hostname: 192.168.1.2
Uuid: 36ab6e63-8d4a-4ac9-a7a8-15734b76213a
State: Peer in Cluster (Connected)
Other names:
test-gluster1.local

What behaviour did you expect instead

Puppet run without any changes.