This post is more than 5 years old
3 Posts
0
8890
AX150i and failover cluster problems
Hi all.
I have tried to set up failover cluster on Win2003 x64 with AX150i.
We have already one cluster configured with AX so I now it is working properly.
The problem is that when I create cluster one node has ownership of the disks and another one should see disks but not have any rights on them except wait for them to be released from the first node, but on my configuration when first node take ownership on the disks the disks on other node are beginning to failing, they become unavailable, cluster service sees that and stops, so second node becomes unavailable too.
I have the latest firmware installed, the latest PowerPath and the latest Microsoft iScsi initiator installed.
Any suggestions would be appreciated.
Thanks
Regards
Amar
laibhari
19 Posts
1
February 18th, 2009 10:00
you need to have connection to both SPs from both the servers in cluster, this will ensure redundancy and proper access to disk as well.
Dev Mgr
4 Operator
4 Operator
•
9.3K Posts
0
September 18th, 2008 13:00
Are you going through a switch/vlan with 1 link to SPA and 1 link to SPB?
In a cluster, all nodes need to communicate to the same storage processor or processors (so both to SPA, both to SPB, or both to both storage processors; not 1 node to SPA and the other to SPB).
If you do have redundant connections from both servers, can you clarify a bit more what happens, as I'm not sure I fully understand your explanation of what's happening.
Dzoner
3 Posts
0
September 18th, 2008 13:00
At the moment one server is connected to SPA and other on SPB because first two servers are connected in same way. The first two are connected to SPA1 and SPB1 ports, while those new ones are connected to SPA2 and SPB2.
I will try to connect them to switch and configure switch so they can comunicate on both ports SPA and SPB.
Thanks for reply.
Dell-Jeff G
156 Posts
0
September 18th, 2008 14:00
Cluster documentation for AX150 is located here.
The LUN ownership is per SP so when the LUN is one by SPA, the other cluster node conneced to SPB can't access is properly.