MSA Storage
1752586 Members
4694 Online
108788 Solutions
New Discussion юеВ

Re: Switch Redundancy Not working

 
HDZ
Occasional Contributor

Switch Redundancy Not working

Hi All,

Sorry if this is a stupid question I've not dealt with HP SAN Kit in a while. I am having trouble with the redundancy in my SAN setup.

I have an MSA2000 with dual controllers. This is connected to two HP 8/20 FC Switches. This is connected in the usual redundant manner.

Controller 1 Port 1 connected to Switch 1.
Controller 2 Port 1 connected to Switch 1.
Controller 1 Port 2 connected to Switch 2.
Controller 2 Port 2 connected to Switch 2.

I then have a Windows 2003 server with two HBA's connected as follows.

HBA1 connected to Switch 1.
HBA2 connected to Switch 2.

I have created a vDisk and explicity mapped the LUN to the Windows 2003 server using all four controller ports. When I look in disk manager it shows 4 disks. It looks like it is showing each connection on the controller as a disk rather than a single disk and when I disconnect one of the HBA's on the server it looses two disk connections.

I was expecting the disk to appear as a single disk and full redundnancy provided no matter what might fail.

What am I doing wrong here? is there a document which discusses configuring your environment for full redundnacy?

Thanks
3 REPLIES 3
P.Prokopets
Advisor

Re: Switch Redundancy Not working

You should install MPIO DSM for Windows. After this you will see only one disk.
HDZ
Occasional Contributor

Re: Switch Redundancy Not working

Excellent that worked. Thanks
Johnny von Heimerall
Frequent Advisor

Re: Switch Redundancy Not working

Hi there,

Besides MPIO, -for full redundancy you need to change cabling to as follows

Controller 1 Port 1 connected to Switch 1.
Controller 2 Port 2 connected to Switch 1.
Controller 1 Port 2 connected to Switch 2.
Controller 2 Port 1 connected to Switch 2.

After this, at a convenient time, try to shut down each of the controllers at a time via SMU to verify that the controller failover works.
Also try to pull 1(! only single point of failure !) FC cable at a time to verify cabling and switch redundancy.