Understanding how to...

wedge1988

Hi all, This post is for me to understand a couple of things; i hope you can all give me an idea of how to set them up and understand them...

1. Network Load Balancing & Server Clustering..

-> I understand you have to turn on NLB in the TCP/IP properties and use the admin mmc console to add nodes, i just dont understand the hardware requirements and how you set it up properly... does the hardware need to be the same? can i do it over a switch??? etc etc..

2. Domain controller load balancing..

-> maybe a little bit from the last question, i have not the foggiest about how you set it up. do i run the dcpromo and choose to add a new dc to the forest etc? in turn does it configure another primary DNS server etc etc?

3. DFS..

-> I know dfs is used for searches, but how the hell do you structure the servers for DFS? how do you sync data?

4. Clustering..

-> What exactly is involved? how do the servers sync data? do i have to install windows separately and then configure or can i configure 2 servers that were separate to share the same files... do they sync???

Ill have a few more questions no doubt but if anybody knows any of these, post a reply! cheers all!

royal

#1 - Hardware doesn't have to be the same. You can adjust the load weight accordingly if it's not. You can attach both NICs to a switch but it's most efficient if you attach the NLB enabled NICs to their own switch or use a VLAN since they use the same MAC and therefore will have the switch revert to sending the traffic to all ports of the switch turning it into a HUB.

#2 - Read the following posts for what I typically do when adding a new DC:
http://www.techexams.net/forums/70-291-windows-2003-network-infrastructure/30075-setting-up-active-directory-dns.html
http://www.techexams.net/forums/70-291-windows-2003-network-infrastructure/17259-ad-dns-zone-replication.html

#3 - Namespace servers are for the virtual namespace and you add folder targets which point the real phsyical part of the namespace path to a real folder. So for example, \\techexams.net\files would be the beginning namespace. You can then have \files\software point to FileServer01 and \files\software to also point to FileServer02. You then right-click on the Target Folder and choose to create a replication group for that folder and choose which server is Primary just for the sync so if the data exists in both places, the Primary is the authoritative copy. You then set up the replication options within DFS.

#4 - Shared Storage - aka SAN. You present one server with shared storage with whatever tools the vendor wants you to. You then shut the node down, bring up the second server, get shared storage working on that. You then install Cluster Services and make sure that disk is added to the cluster. The cluster takes control of the disk which allows you to bring back up the other node. It'll still have the disks but won't be able to access it. The reason you do it this way is to prevent disk issues which can include corruption.

There are some other things like using a cluster with local storage and using application replication like CCR for Exchange 2007. But don't worry about this for the exam and it's something you need to know based on an enterprise application.

wedge1988

Thanks royal, that was really helpful!

Now to try them out when i have a couple of hours lol...

I do have another question, but its probably quite a long one... (Lets assume i have a server, a switch and a cisco PIX router)

If i wanted to set up a web server, i assume i would need to connect the cisco PIX router to the internet then tell the PIX router to forward queries to the local host? But how exactly is this done? and what sort of security am i forgetting about?

thanks!