Cluster engine connections

This topic describes the several ways that cluster engines are connected in SAN File System.

The SAN File System cluster engines connect with each other in three ways:
These connections are used in the following ways:

The Ethernet connections use a customer LAN that does not need to be isolated. For performance reasons, the metadata server engines and the SAN File System clients should all be on the same subnet; that is, there should be no routers or gateways in their paths. If you deploy a metadata server cluster on a heavily loaded IP-network, it is possible that there could be delays in getting heartbeat traffic across the network. This could result in unnecessary metadata server fail-overs. A possible solution to situations such as these would be a dedicated VLAN.

Parent topic: Validating the existing SAN infrastructure

Library | Support | Terms of use | Feedback
(C) Copyright IBM Corporation 2003, 2004. All Rights Reserved.
IBM TotalStorage SAN File System v2.2