8 node active/active cluster

#**101**

What's your thought with a mscs?

This will give you no fail options if the enclosure goes down, that's
a big spof.

Fail cluster should only have disk in common.You need the possibility
to have single parts of the cluster down without affect, the exeption
is disk witch is shared. Of course there is possibilitys to use san
replication.

Geoff pointed out in another message that "Blade + Cluster = Low
Availability". I don't agree. Blades don't have the hottest hardware
but if a cluster is for maintaining uptime, ie just failing for
patching this is not a limitation if you separate the physical
hardware. Of course they then have to be hardware separated.

Todays blade has both performance (but rack server goes higher) and
uptime availability.

/Peter

On Wed, 4 Jun 2008 13:54:01 -0700, Leonid
<Leonid (AT) discussions (DOT) microsoft.com> wrotc:

Quote:

Hi,
We are planning to build a 8-node SQL Server cluster on 8 node blade chassis
as well.
Each server has 64Gb of memory. i am planning to install run one SQL Server
instance per server on 32GB . The remaining 32GB can be used by one failed
over SQL instance. In worst case scenario it wil be 4 servers running 2 SQL
server instances each. (Of cause I will leave some memory for OS on each
server)

Any experience - concerns would be a great help.

Thanks

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

8 node active/active cluster

microsoft.public.sqlserver.clustering microsoft.public.sqlserver.clustering

Re: 8 node active/active cluster - 06-05-2008 , 01:21 PM