Corosync error "No interfaces defined" in a cluster member

Question

I am having an error starting corosync on a cluster member:

May 16 00:53:32 neftis corosync[19741]:  [MAIN  ] Corosync Cluster Engine ('2.3.4'): started and ready to provide service.
May 16 00:53:32 neftis corosync[19741]:  [MAIN  ] Corosync built-in features: dbus systemd xmlconf snmp pie relro bindnow
May 16 00:53:32 neftis corosync[19741]:  [MAIN  ] parse error in config: No interfaces defined
May 16 00:53:32 neftis corosync[19741]:  [MAIN  ] Corosync Cluster Engine exiting with status 8 at main.c:1278.
May 16 00:53:32 neftis corosync: Starting Corosync Cluster Engine (corosync): [FALL�]
May 16 00:53:32 neftis systemd: corosync.service: control process exited, code=exited status=1
May 16 00:53:32 neftis systemd: Failed to start Corosync Cluster Engine.
May 16 00:53:32 neftis systemd: Unit corosync.service entered failed state.
May 16 00:53:32 neftis systemd: corosync.service failed.
May 16 00:54:06 neftis systemd: Cannot add dependency job for unit firewalld.service, ignoring: Unit firewalld.service is masked.
May 16 00:54:06 neftis systemd: Starting Corosync Cluster Engine...
May 16 00:54:06 neftis corosync[19773]:  [MAIN  ] Corosync Cluster Engine ('2.3.4'): started and ready to provide service.
May 16 00:54:06 neftis corosync[19773]:  [MAIN  ] Corosync built-in features: dbus systemd xmlconf snmp pie relro bindnow
May 16 00:54:06 neftis corosync[19773]:  [MAIN  ] parse error in config: No interfaces defined
May 16 00:54:06 neftis corosync[19773]:  [MAIN  ] Corosync Cluster Engine exiting with status 8 at main.c:1278.
May 16 00:54:06 neftis corosync: Starting Corosync Cluster Engine (corosync): [FALL�]
May 16 00:54:06 neftis systemd: corosync.service: control process exited, code=exited status=1
May 16 00:54:06 neftis systemd: Failed to start Corosync Cluster Engine.
May 16 00:54:06 neftis systemd: Unit corosync.service entered failed state.

Here is my config on the three nodes but it is failing just in netfis which I added recently.

totem {
    version: 2
    secauth: off
    cluster_name: cluster-osiris
    transport: udpu
}

nodelist {
    node {
        ring0_addr: isis.localdoamin
        nodeid: 1
    }

    node {
        ring0_addr: horus.localdoamin
        nodeid: 2
    }

    node {
        ring0_addr: netfis.localdoamin
        nodeid: 3
    }
}

quorum {
    provider: corosync_votequorum
}

logging {
    to_syslog: yes
}

I am running a pacemaker, corosync, pcs cluster on CentOS 7.1 64. bits.

I searched on internet but It is not clear what is going on.

Could you help me?

Solution

Do the hostnames you have listed in your Corosync configuration resolve correctly? I would start by verifying that.

# host isis.localdoamin

Since, "domain", appears to be spelled incorrectly (or in a language I am ignorant of), I am going to guess that command fails? ;-)

Also, you could use the short hostname (without ".localdomain") or the IP addresses of the interfaces you'd like Corosync to bind to instead.

Comments

[Matt Kereczman]

Ah, "doamin" turns out to be Spanish!! Try the shortname or IP address; it definitely sounds like it can't find an interface to bind to.

[Anon]

You are right. I tryied with neftis.localdomain and I could not reach it. in my /etc/hosts file I have this line 10.10.10.230 neftis.localdomain Pretty strange, isnt it? Thank you!