You could use ESX over NFS and step away from the 100% overcommit/reserve for iSCSI... it would help with capacity issues.
-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Raj Patel Sent: Monday, November 24, 2008 6:14 PM To: Premkumar Subramanian Cc: toasters@mathworks.com Subject: Re: FAS2050 with 20 disks
We use the predecessor to the 2050 - a 270c - we ended up settling for a 50/50 split of disks.
We do Exchange & SQL iSCSI luns from one head and ESX iSCSI luns from the other head (about 30VM's). They're clustered for failover, we gain on having both heads working for us and if there is a performance hit associated with a split in the spindles no one has noticed.
Cheers, Raj.
On Tue, Nov 25, 2008 at 8:42 AM, Premkumar Subramanian prem@x-or.co.uk wrote:
Hi,
I plan to implement a FAS2050 with 20 internal SAS disks. I intend to
have
12 disks on one of the controller and 8 on the other. the idea being to make use of processing power of both controllers.
Now the
trade-off is the number of spindles in each aggregate possibly
affecting
performance. I have seen a few people use just 2 disks on the second controller with no spares and have the other 16 disks+2 spares on a
single
controller.
which of these two would be a better option?
--
Best regards, Premkumar Subramanian