pauliew1978
Enthusiast
Enthusiast

islolation response das.isolationaddress and esx/san config

Hi all,

I have two esx servers and two sans. Both esx servers can see both sans and the sans are replicating between on a dedicated fibre link. Each esx server has two nics and each san has two nics.

esx1 nic 1 goes to live network (and is for vms and service console)

esx1 nic 2 is for iscsi access and vmotion to both sans

esx 2 has the same config.

However the link between the esx servers and the iscsi sans eventually goes down to one fibre link. If this fibre link breaks or the switch goes down then the vms hang on the esx server that has lost its iscsi access. The vms wont restart on the other esx server because the service console is still pingable.

My question is twofold:

1.) Could I set the das.isolationaddress on the esx hosts to be something on the iscsi network (i.e. one of the two switches that I have between the two esx servers and the sans) so that if a link goes down it will power down the vms on the esx server that has lost its iscsi connection?. Off the top of my head I dont think this is possible anyway as there would still be a disk lock on the san

2.) If i can't do the above or something like the above what would the process be in order to get the virtual machines back up and running? my thoughts would be to ssh into the esx server that has lost its iscsi connection and find the process id of the vms and kill them. the restart on the other esx server, however I am concerned that there would still be a disk lock and I wouldnt know how to get rid of this?

many thanks for any help!

cheers,

Paul

0 Kudos
1 Reply
pauliew1978
Enthusiast
Enthusiast

sorry for the multiple posts there seems to be a problem with the apache engine.... it kept erroring so I kept posting

0 Kudos