The following are the possible causes that a host is disconnected. Make sure that you check whether:
- The management node can normally access port 22 of the host.
- ZStack Custom ISO can be used to install the system properly.
- The libvirt service of the host works normally.
- The network attached to the cluster where the host is running can be mounted normally.
If eth1 is specified to be the physical NIC of the L2 network that is attach to the cluster, but the NIC device number of the host has been changed, which means that eth1 is no longer existed, the host cannot be reconnected.
- The system disk of the host is out of storage capacity.
- Assume that you must attach the host to the NFS primary storage. If you fail to attach it to this primary storage, check the link from the host to this primary storage.
- Assume that the host uses a distributed storage. If this distributed storage has a storage network, the host can access the storage network directly.
- If you use a VXLAN network, when you attach it to the cluster, the specified VTEP IP should be existed. If this IP is not existed, the host cannot be reconnected.