help with multipath msa1510i setup

[Adrian Cruz]

please excuse me if this may seem as a dumb question but i'm not too keen with this type of setup.

anyhow, i have an msa1510i that hosts the DBs of an active/passive sql 05 cluster. my problem is that the controllers seem to failover a lot and sometimes continue until a critical lockup occurs or the controller decides to disable itself and i have to reseat it. and obviously sometimes MSCS does not like all of the failovers and decides to get confused and the cluster is completely down. so it's not looking good for redundancy right now.

so i'd like to see if anybody can post their setup or a sample setup?

my setup is as follows: (i have one vlan'd procurve 2848)

sqlserver01 and 02 on these networks:
eth0 192.168.1.0/24 (LAN)
eth1 100.100.100.0/24 (heartbeat)
san0 192.168.100.0/24 (vlan0)
san1 192.168.101.0/24 (vlan1)

msa1510i on these networks:
data port sa0 192.168.100.0/24 (vlan0)
data port sb0 192.168.101.0/24 (vlan1)


any help is most appreciated.

cheers!

[Gordon Waldie]

Hi Adrian,

Did you find a solution for this? I am in the same trouble. We have a DB cluster using a mutlipath setup to our MSA1510i and find that ever couple of days the controllers crash or failover. It is improved slightly if we have only a single controller in the MSA but after 2 days it will crash again and the MSA needs restarted.

Thanks

[Adrian Cruz]

hey gordon, so actually my cluster is FINALLY stable. i had to go through 4 controller replacements and 2 controller memory replacements until the SAN decided to be happy! i wasn't the happiest customer until then. apparently, hp support kept sending bad controllers. i had to send over the revision number of both of my controllers to an HP technician and he realized that it was faulty.

but, everything has been up & running and healthy ever since. so no complaints. ::knocks on wood::

best of luck. cheers.

[Gordon Waldie]

Hi Adrian, we have been taking similar steps. We have changed the controllers 3 or 4 times, replaced the memory and even replaced the whole chassis incase the backplace was the problem. We are working with HP to try and discover the problem. I dont suppose you have your case number which I could give to our HP engineer to see if he finds any clues. Maybe we both have controllers from the same faulty batch. Thanks, Gordon