Symptoms

All nfs cluster deployment files placed on MN in /root/deploy_nfs_cluster folder.

After successful cluster creation, diagnostic tool crm_mon shows, that quorum host in OFFLINE state:

# crm_mon
Defaulting to one-shot mode
You need to have curses available at compile time to enable console mode

============

Last updated: <DATE>
Current DC: <HOSTNAME1> (1efdbbf8-3a9f-43b6-bc48-a95724661cc4)
3 Nodes configured.
4 Resources configured.

============

Node: <HOSTNAME1> (e5639963-7820-4af3-aa9c-bbde7aff4f63): OFFLINE
Node: <HOSTNAME2> (1efdbbf8-3a9f-43b6-bc48-a95724661cc4): online
Node: <HOSTNAME3> (1cb19be7-1852-4705-9c5e-7f759fd69680): online

Cause

Unsupported network configuration

Resolution

According to deployment guide schema:

A quote from corresponding page of deployment guide:

Each NFS server has two interfaces - eth0 and eth1 - enabled.

  • For the eth0, configure a static IP address (from BackNet) that will be used by the configuration file for the Heartbeat.
  • For the eth1, configure a static IP address (from StorageNet) that will be used by the DRBD.

The quorum server has one interface eth0 with configured static IP address from BackNet.

This explicitly shows that both NFS server and Quorum should be connected to backnet using the same interface name - eth0.

The NIC names are hardcoded in deployment scripts, A feature request POA-72997 exists to allow configuring them. Please contact your account manager to track its status.

Internal content