First off, i will say i was recently thinking about clustering.
now the details
I've got a Windows Server 2003 box that acts as the vCenter server for license management etc, as well as (more importantly) a cheap man's SAN. There's a SCSI Array attached to this box that I use for storage of all of my VM's virtual disks, setup as a shared datastore for all 4 of my esx hosts to utilize vmotion/HA etc.
In an effort to tinker with stuff and minimize downtime should a hardware failure occur, i'd like to setup a failover for this system aside from it's built in precaution of raid 1 drives for the OS and raid 10 for the array. I don't have any major objections to switching to linux aside from the time required to rebuild the box and learn the finer points of the OS than i already have broadly understood, but let's put that aside and pretend i hate llinux and want to stick with windows (crazy right?)
What might be your suggestion that you've implemented for something like this?
thanks all!
Surely you will need another box acting as your backup datastore when your main datastore failed. Look for NAS solution which support iscsi/nfs then make it as your backup datastore and continuously replicate your VMs using vReplicator(hybrid) like software to replicate selected VMs images. You can straightly power on the VMs later on using this backup datastore when main datastore failed.
vcbMC-1.0.6 Beta
vcbMC-1.0.7 Lite
sorry, that's the one thing i didn't cover. i do have the VMs backed up to the esx hosts, in such a way as to load balance them. so i can turn them on on the local esx boxes if i needed to.
im looking for a close to seamless instant failover solution.
side note: ill check out vreplicator for fun, thanks for the suggestion
I'm not a Windows guy and the only solution I think most right now, DRBD and Linux HA. Unfortunately we pretending you hate Linux right mate..
vcbMC-1.0.6 Beta
vcbMC-1.0.7 Lite
yeah, thanks for playing along with me