I've 3 Nics of a Dell R610 running ESXi 4.1 connecting to all 8 nics in the MD3200i SAN.
Each San nic has a seperate subnet, and I've set up a VMK for each port on the host for each port on the san, and restricted the vmk to use only one nic on the host.
All was working fine (except the long boot, I'll come to later) until last night I think one of the VMs got busy and this appeared to slow down the whole host to the point it needed rebooting (couldn't kill the running VMs) - It appeared to be SAN related as activity on this went crazy.
My hunch was that it was getting busy on one path and then when it tried to use the next, something went awry and so why I think it's related to the long boot time:-
When booting the host it gets stuck on each of these for about 10mins each:
iscsi_vmk loaded successfully
vmw_satp_lsi loaded successfully
cbt loaded successfully
Once the last one has come up I can SSH into the box and I can see in the messages lots of these coming through each second:
Jan 23 21:20:17 iscsid: Notice: Assigned (H41 T0 C23 session=50c, target=7/2c)
Jan 23 21:20:17 iscsid: cannot make a connection to 192.168.37.107:3260 (101,Network is unreachable)
Jan 23 21:20:17 iscsid: Notice: Reclaimed Channel (H41 T0 C23 oid=7)
Jan 23 21:20:17 iscsid: session login failed with error 4,retryCount=0
What I suspect is happening is that every VMK is trying to talk to every IP on the SAN - Which they can't as they're on diff subnets. I've tried to
change the Scsi.CRTimeoutDuringBoot value to 5000, but makes no difference.
In the iSCSI Software adapter I do see too many paths here too, and not all have a target.
Sorry for the long post but to cut a long story short, is there a way to permanently disable these paths? ... or anything other pointers you can give me to help?