March 15, 2011 at 2:37 am
Hi,
I would like some advice on whether this design for clustering is appropriate. The plan is to have 1 physical server and 1 vm in a cluster which will then failover with same setup. I don't see why we should include the vm in a cluster as I thought the physical machine will suffice. Any advice well appreciated.
Thanks
March 15, 2011 at 2:56 am
In clustering the two nodes should be identical. That's hardware, OS, drivers, etc. Clustering a physical box with a virtual is definitely not having two identical nodes.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
March 15, 2011 at 3:38 am
GilaMonster (3/15/2011)
In clustering the two nodes should be identical. That's hardware, OS, drivers, etc. Clustering a physical box with a virtual is definitely not having two identical nodes.
Correct, although a Physical (primary) and Virtual (Secondary) is not impossible and can be done.
There are some tricks to perform for the virtual though.
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
This thing is addressing problems that dont exist. Its solution-ism at its worst. We are dumbing down machines that are inherently superior. - Gilfoyle
March 15, 2011 at 4:28 am
Hi,
Thanks for the info, Henrico do you have any articles of how I could implement the physical/vm side? I know I will need to present the VM as raw devices.
March 15, 2011 at 6:46 am
GilaMonster (3/15/2011)
In clustering the two nodes should be identical. That's hardware, OS, drivers, etc. Clustering a physical box with a virtual is definitely not having two identical nodes.
The key here is "should" and is therefore recommended for a number of reasons such as the ability to failover and still maintain a consistant level service. Therefore there is nothing stopping you implementing a cluster with different hardware so long as you are willing to support two different servers and their potentially unique issues.
My previous company accepted that in the event of a serious hardware failure, they would accept a reduced capacity server until such a problem could be fixed. We had 3 servers of the same model admittedly. Two were identical and ran their own instance of SQL, the third was a lower spec CPU and reduced RAM.
A side point, Microsoft no longer dictate that a cluster hardware should be on "the list" to be supported as part of Windows 2008.
March 16, 2011 at 4:01 pm
TST1 (3/15/2011)
Hi,Thanks for the info, Henrico do you have any articles of how I could implement the physical/vm side? I know I will need to present the VM as raw devices.
Are these ESX server VMs?
What shared storage medium are you using?
clustering a physical box with a VM is perfectly valid, at minimum, ensure they utilise the same operating system and patch level.
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply