Quantcast
Channel: High Availability (Clustering) forum
Viewing all articles
Browse latest Browse all 6672

Pre purchase advice on Dell VRTX Cluster config

$
0
0

Hi, I have read quite a few posts on this box and the more I read the more confused I am getting.  

The required end result is an SQL 2012 std AlwaysOn 2 node cluster; a highly available general file and profile share; and a 2012 hyper-v server.

Am I correct in thinking the following...

2 blades (4 procs, 8 cores, 32gb ram) each blade running Server 2012 std, hyper-v enabled running 2 vm's of server 2012,  vm1 (2 cores) file/profile share, vm2 sql (4 cores) all installed onto the blade HDD/SSD. 85 users, business critical for SQl and file / profile share.

1 blade (2 proc, 12 cores, 32gb ram) server 2012 ent, hyper-v enabled.  This will run assorted linux and windows vm's. low usage, non business critical machines.

1 blade (1 proc, 4 cores, 16 gb ram) server 2012 std, scale out file server using 4 of the shared drives in one raid 10 LUN (15k drives), smb 3.x for SQL. 12 shared drives in a separate raid 50 LUN (7.2k NL) for the file/profile share vm's and hyper-v. This leaves hdd's free for global spares and future expansion.

My questions are;

Is this a viable configuration or am I looking at the box the wrong way, having a single blade controlling the LUNS? Should I be looking at putting Server 2012 enterprise on 2 blades, upping the memory to 64gb on them, and running the other 2 blades at a base level as a file cluster?  SQL speed is the most important factor for us, we are about 50/50 read write and use FILESTREAM to store data, this will not change and writes are likely to increase going forwards.

If I go for the 16port internal switch using all M502P blades will I gain a benefit from using Intel SFP+ cards for the 2 SQL blades ?  I have a Netgear GS752TXS stack that supports DA connections.

At the moment our SQL, File Share and profiles are all on separate physical boxes.  We have moved our desktops to VDI on vmware / citrix, expanding that with another Equalogic and couple of nodes will cost considerably more than a VRTX.  I want to remove the remaining single points of failure. I accept the VTRX is still a single box but it will replace a T510, T610 and PE2950 all out of warranty.

Many thanks for any suggestions or advice.

Kane.


Viewing all articles
Browse latest Browse all 6672

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>