June 5, 2008 at 1:20 am
I have a 2 nodes cluster running three sql 2005 instances in production already. I have to build a 4th instance for an upcoming new release but unfortunately the hardware guys would not make themselfs available for discussing the new ms cluster group that has to be build (we're outsourcing).
I would appreciate if someone there had this experience before; I'm highly interested how long it would take to build the ms cluster with the hardware (logical) design provided - storage size/type, CPU, memory, all calculated to fit in the new app module - and if this new build would affect the other cluster groups already online and for how long.
Thanks everyone in advance.
June 5, 2008 at 10:06 am
if this instance is going on the physical nodes running the other 3 instances, then you need some common storage to hold the data for this instance as well as the install files
June 5, 2008 at 5:11 pm
The new LUNs would be added as resources for the new cluster group; it's included already with the hardware design requirements. What I do not understand is if, by adding new storage, the existing online cluster groups would be affected or they can still run as usual.
June 5, 2008 at 5:26 pm
in our case when the SAN admin maps the new LUN's to the server we have to reboot for Windows to see the new disks
June 5, 2008 at 6:11 pm
Reboot is usually required.
However, the other questions you ask is a bit harder to answer. Do you have enough physical RAM to support a 4th instance? Keep in mind that if one of the node fails, all 4 instances will be running on the one remaining node. You want to have enough memory to ensure you don't have a memory short situation.
The same consideration needs to be made with respect to the processors. When the 3 current nodes are on one server, how does the system handle the load?
And then consider the fabric of the shared storage. This is usually not an issue, but make sure it isn't.
K. Brian Kelley
@kbriankelley
June 5, 2008 at 6:36 pm
Thanks all for your responses. It's a relief to know that the outage would be for a reboot only.
June 16, 2008 at 9:19 pm
On EMC SANs:
- it it NOT necessary to reboot the server to see a new LUN
Rather:
- open Computer Management;
- click on Disk Management;
- select Rescan Disks from Actions menu.
When the volume shows up:
- partition and format the drive
- assign it as a Cluster resource
- etc.
This might not work in your environment, but try this before you down your server(s) unnecessarily.
June 17, 2008 at 12:59 am
Hi Michaela!
I have no response for your question, but i Have a simple question for you.
You have 3 instances over a cluster of 2 nodes. Do you have a resources group for each instance? What do you think about to have two instances into a resource group?
Thanks in advance
Viewing 8 posts - 1 through 7 (of 7 total)
You must be logged in to reply to this topic. Login to reply