This content has been marked as final. Show 7 replies
It is not required, however it will be installed in almost all configurations.
During the setup you can select it not to be configured.
Getting rid of the Webcache is easy after the installation, except when using Portal.
Well I guess I'm going to have some problems because we are using the Portal. We are going from 1 middle tier to what I believe is called a clustered middle tier and we are going to use an F5 to load balance those two servers. We are going to have the infrastructure and middle tiers on the same two boxes.
I guess they should get an AS admin and not a developer to implement this solution. I'm not even sure if I'm suppose to install the two servers and then set them up for clustering or set them up while the install is going.
Clustering the middle tier should not be a problem.
Do not put the infrastructure and the MT on the same machine, as clustering will become very difficult.
You should definitely have somebody with experiance on board instead of a developer. AS clustering and loadbalancing are difficult to configure.
Andreas I would like to thank you for your suggestions in this matter. If I shouldn't be putting the infra and midtier on the same box then I believe they I believe they don't have enough hardware. Here's a better description of the situation:
The person that set this up didn’t do his research before he started the project. After the hardware was bought he left the company. He bought two new v240s and an F5 load balancer and was planning to use the F5 to balance the two loads between the two servers . Right now we have the Midtier and infra installed on one v240 and we have a 2nd box with just the webcache on it sitting in front as a reverse proxy. I was told by my unix admin that he could get me two more v100s for the midtier boxes, but we aren’t sure on what type of further licensing we will need.
Some more information after another conversation today with my unix guy... We have decided to run Solaris 10 and use zones to seperate the mid and infra tiers.
Message was edited by:
My suggestion is to make sure that you have an architecture that corresponds with your goals.
Usually you want to use several machines when you either want high availability or scalability.
As the middle tier and infrastructure offer different services and are different in their configuration when it comes to clustering you will need at least three machines (2*mt, 1*infra) for scalability of your portal, or four machines (2*mt, 2*infra). More machines are possible.
I'm working in a project now where we use 16 machines (8 per site) to get the high availability working.
Have a look at the High Availability Guide of the 10.1.2.0.2 AS http://download.oracle.com/docs/cd/B14099_19/core.1012/b14003/toc.htm and the Portal advanced sections http://download.oracle.com/docs/cd/B14099_19/portal.1012/b14037/toc.htm.
This will give you a lot of stuff to read - especially the HA of the infra is very difficult.
Try the following: start with clustering the MT, then use the loadbalancer, then use HA of the infra (if necessary).
IMHO scalability and/or HA with 10.1.2.0.2 requires more that two machines. It is possible but very complicated with two machines, three is easier, four is best practice.
Yes I've been up all night reading everything that has to do with the HA and advanced Portal setups. I'll be using 5 logical machines by taking advantage of the zones of solaris 10. I will have two web cache boxes clustered, two Portal/Wireless/Reports boxes, and 1 infra. Our environment is pretty small at this time and this is already going to be an overkill. The v240s we setup are a huge overkill in my opinion, but they will provide for plenty of growth. I do thank you very much for taking the time out of your busy day/morning to reply to me. One step at a time and it will come together. I will keep you posted on how I do hehe.
I wasn't aware that the zones in Solaris are an equvalent to VMWare - where you can create virtual machines.
If you still have room on the machines you might consider creating 6 or 8 machines.
But this is a very advaced setup using RAC/ ASM/ CRS/ MRCA/ Infrastructure HA, etc. Keep this for a rainy sunday ;-)
Glad that I could help.