We currently have a 2-node cluster - each node has 60 DBs spread over three SQL Server instances running on 2GB RAM, 2 Pentium III 700Mhz CPUs and 100GB storage. We are experiencing performances problems and the servers are too old to reinvest in, so we are looking to buy new replacements. I have seen good advice given in here about hardware needs so I was hoping you all could advise.
The big problem is RAM - always running at at least 85% - sits at 95+% several times a day. CPU runs around 60% most of the time but can flatline around 100% a few times a day. There's an average of at least 50 (so 150 on each node) processes running on each instance simultaneously. When things get really bad there are delays and timeouts. Storage is also about to become a problem.
We were thinking along the lines of each new node having 8GB RAM, 300GB storage and 2 Pentium 4 2.0Ghz CPUs. Do you think that is enough to keep us covered for the next 3 years at 20% growth each year and failover considerations? Thanks!
The big problem is RAM - always running at at least 85% - sits at 95+% several times a day. CPU runs around 60% most of the time but can flatline around 100% a few times a day. There's an average of at least 50 (so 150 on each node) processes running on each instance simultaneously. When things get really bad there are delays and timeouts. Storage is also about to become a problem.
We were thinking along the lines of each new node having 8GB RAM, 300GB storage and 2 Pentium 4 2.0Ghz CPUs. Do you think that is enough to keep us covered for the next 3 years at 20% growth each year and failover considerations? Thanks!