April 24, 2014 at 10:28 am
I am being asked by management to find a solution (if there is one) :
Scenerio : To keep a very large system running optimally in a VM cluster, Take PR01 and make PR02, PR03, PR04. Distribute the 45 databases and 9T+ of disk across multiple VM guest.
Each PR## is a SQL Server 2012 Enterprise guest on a VM 5.1 cluster.
So instead of PR01 needed 16 core and 128g of memory, each one will have 4 core and 32g of memory. Making VM HA more manageable. (yes, DRS rules will apply). Also provides more HBA paths and distributes i/o over more physical disk on the SAN.
Instead of a connection string having to know PR01.dbo.UserDB01, PR02.dbo.UserDB03, ect the connection would be PRDB.dbo.UserDB01. That way if needed 1) UserDB can be moved to any of the PR## 2) new PR05, PR06 can be added as needed. The end user and processes are not allowed to touch system databases, no PR## will have a user DB called the same name.
There are seperate VM guests on other VM clusters and Citrix servers that need access to PRDB. As things expand and move around, none of the connection strings need to be changed.
I am looking into RadWare and modifing level 7 information, but that is iffy and $$$$$$.
Any ideas?
April 24, 2014 at 10:35 am
DNS Redirection is one way to go.
Users use the PRDB.+ and DNS actually routes it to the correct VM PR01. or PR02, etc.
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply