Troubleshooting clustered server issues
I have a two-node clustered server. Both servers have two host bus adapters (HBAs) connected to two fabric switches and to a redundant SAN controller. My problem is only one path has connectivity to storage device, HBA 1 of server 1 and HBA 1 of server 2 has no connectivity to storage device. Both fabric switches are working fine. What could be wrong, can you help me?
Since I do not know enough about your environment to give you definitive answers, here are some things you can look at. Let's start with some procedural questions. Can your Fibre Channel (FC
1 in both servers
? If not, that would explain your problem. What multipathing software are you running on your servers? (MPIO, PowerPath, DMP, etc.) Is it compatible with your target storage? If you do not have multipathing software, that could explain your problem. If you do, and it is not compatible with your target storage, that also could explain your problem. How are your HBAs and target storage ports zoned
? If HBA 1 of both servers are not correctly zoned to the right target storage ports, that too will cause your problem.
There is currently only one software package on the market that can instantly diagnose your problem. It is called SANScreen from Onaro. It is a bit pricey and typically only cost justifies for SANs larger than 128 ports.
You can also contact your server, HBA, multipath software, FC switch or target storage vendor for troubleshooting help.
This was first published in December 2005