corosync:[TOTEM] Digest does not match

In one two-node cluster, The following can be seen while checking the coroysync service status, and the same error is also filling the messages log file.

corosync[]:[TOTEM] Digest does not match
corosync[]:[TOTEM] Received message has invalid digest... ignoring.
corosync[]:[TOTEM] Invalid packet data

To solve the issue perform the following actions:

1. Check if the /etc/corosync/corosync.conf and the /etc/corosync/authkey files are synchronized between the two nodes. For example with md5sum, files size, and also the permission mode of authkey file. The output of the md5 check should be the same on both nodes. 
2. Check if the “transport” option is inside the totem section in the file /etc/corosync/corosync.conf for both cluster nodes. If it is outside, then modify it to be inside and stop/start the pacemaker service.

Example(“transport” option is inside the totem section):

totem {
        version: 2
        secauth: on
        crypto_hash: sha1
        crypto_cipher: aes256
        cluster_name: hacluster
        clear_node_high_bit: yes
        token: 5000
        token_retransmits_before_loss_const: 10
        join: 60
        consensus: 6000
        max_messages: 20
        interface {
                ringnumber: 0
                mcastport: 5405
                ttl: 1
        }

        transport: udpu
}

If the transport option is placed outside the totem stanza (where it belongs). This means that it is simply ignored and so corosync reverts back to the default behaviour, which is to use multicast.
To compound the issue, some clusters have the same mcastaddr directive. Two clusters configured in this way would therefore be justified in believing they are a single cluster because they are in the same multicast group. They would then attempt to merge together.