Server 2008 Standard R2

[falconmark]

Does Standard R2 support failover clustering? If not, what options do I have when running two servers to ensure that if server 1 fails server 2 will be in use.

 

Server 1 will be setup as a primary DNS, primary DC, and AD-DS Master

 

Server 2 will be setup as a secondary DNS, secondary DC, and AD-DS Replicator

 

All user data is stored on a SAN.

 

So basically if either server fails, data will be available so long as SAN is functioning.

 

Now if server 1 fails, server 2 needs to let clients login via Domain and ensure enforcement of user/group policies.

 

How do I set this up?

[jamitupya]

Have you thought about virtualizing these roles with hyperV?FYI: Standard editions of Server do not support Failover Clustering and only support a 1+1 Rule for licensingcheck out this for licensing and this for comparisonI included licensing as this will likely be your next question. :-)

 

 

 

 

Also, what you advise here is fairly common practice in 2000-2005. The basic way AD works is like you have above, a few thing to think about is user data storage. If you need this replicated, i suggest you use something like DFS/NFS to mount the user shares and have them published in AD instead of on the individual machines.

[jamitupya]

Have you thought about virtualizing these roles with hyperV?FYI: Standard editions of Server do not support Failover Clustering and only support a 1+1 Rule for licensingcheck out this for licensing and this for comparisonI included licensing as this will likely be your next question. :-)

 

 

 

 

Also, what you advise here is fairly common practice in 2000-2005. The basic way AD works is like you have above, a few thing to think about is user data storage. If you need this replicated, i suggest you use something like DFS/NFS to mount the user shares and have them published in AD instead of on the individual machines.

 

apologies....

use this for the comparison

[falconmark]

Have you thought about virtualizing these roles with hyperV?FYI: Standard editions of Server do not support Failover Clustering and only support a 1+1 Rule for licensingcheck out this for licensing and this for comparisonI included licensing as this will likely be your next question. :-)

 

 

 

 

Also, what you advise here is fairly common practice in 2000-2005. The basic way AD works is like you have above, a few thing to think about is user data storage. If you need this replicated, i suggest you use something like DFS/NFS to mount the user shares and have them published in AD instead of on the individual machines.

 

 

Not really familiar with Hyper-V. Data storage is on a DroboPro device via iSCSI. Its going to be connected to the network switch. It should appear on the servers as a local drive I believe.

 

Since i've never set up AD or DC, i'm not sure if what i'm saying is true or not. Just going by what I read about. If server 2 is a DC and ADDS replicate, how does the client know which server to access if one or the other goes down?

 

Does that information get passed when the client logs into the domain? Information as in which servers are currently available.

[jamitupya]

Not really familiar with Hyper-V. Data storage is on a DroboPro device via iSCSI. Its going to be connected to the network switch. It should appear on the servers as a local drive I believe.

 

Since i've never set up AD or DC, i'm not sure if what i'm saying is true or not. Just going by what I read about. If server 2 is a DC and ADDS replicate, how does the client know which server to access if one or the other goes down?

 

Does that information get passed when the client logs into the domain? Information as in which servers are currently available.

 

Yes, basically all the information is stored in the AD and when the client goes to connect it looks up the domain and not the specific server. once the domain is queried the server assigned to the site will respond to the client.

 

For DNS, i suggest using DHCP and when setting your DHCP options to include your IP's of Primary and Secondary DNS and DNS Suffix. this will mean theres redundancy of this service in the clients. Doing this on windows raises other concerns around IP allocation.... for example, if you use DHCP on the Primary Server and it dies, you wont get ANY ip's. A solution to this is to virtualize your additional services onto a pair of Hyper-V VM's in a cluster and hosting the services there. you can host things like DHCP/FileShares/additional DNS/SNS/WINS/Print services... this is one way to provide uptime on those services.

 

 

Hyper-V is relatively easy to install and maintain and providing you have sufficient licensing i'd recommend going down this path to get better use out of your servers. the drobopro seems an OK device and should be able to accommodate this function, though a few design changes would need to occur if you wanted to go down. Post back here and we'll (i'll) help as much as we (i) can

[falconmark]

Yes, basically all the information is stored in the AD and when the client goes to connect it looks up the domain and not the specific server. once the domain is queried the server assigned to the site will respond to the client.

 

stored in the AD as in locally (the client) and then knows which servers to connect to?

 

For DNS, i suggest using DHCP and when setting your DHCP options to include your IP's of Primary and Secondary DNS and DNS Suffix. this will mean theres redundancy of this service in the clients. Doing this on windows raises other concerns around IP allocation.... for example, if you use DHCP on the Primary Server and it dies, you wont get ANY ip's. A solution to this is to virtualize your additional services onto a pair of Hyper-V VM's in a cluster and hosting the services there. you can host things like DHCP/FileShares/additional DNS/SNS/WINS/Print services... this is one way to provide uptime on those services.

 

Hyper-V is relatively easy to install and maintain and providing you have sufficient licensing i'd recommend going down this path to get better use out of your servers. the drobopro seems an OK device and should be able to accommodate this function, though a few design changes would need to occur if you wanted to go down. Post back here and we'll (i'll) help as much as we (i) can

 

For the client computers, should I set static IP's for them as well and have the DNS entries point to both servers. Or should I let them obtain that (IP and DNS) information from DHCP? I was going to run DHCP from the Netgear UTM25 VPN device. Is it better to run DHCP from the server instead? I'm worried like you had mentioned if the server dies, then we have no IP's.

 

I'm not familiar with Hyper-V and hosting services virtually. I get the concept of running a virtual OS on top of the physical OS. I do that with my Mac at home running XP via VmWare Fusion.

[jamitupya]

stored in the AD as in locally (the client) and then knows which servers to connect to?

No, the client looks up the domain via DNS and finds out which server it should authenticate to. I wouldn't worry to much about this particular issue unless you run into real issues. checkout this for an easy to understand guide.

 

 

For the client computers, should I set static IP's for them as well and have the DNS entries point to both servers. Or should I let them obtain that (IP and DNS) information from DHCP? I was going to run DHCP from the Netgear UTM25 VPN device. Is it better to run DHCP from the server instead? I'm worried like you had mentioned if the server dies, then we have no IP's.

I Prefer running DHCP on windows as we use other tools like SCCM/MDT etc which functions better when running windows to windows... If your worried you can run it on your netgear but its still a single point of failure. Given you seem to be a beginner, building a cluster to support this may seem a bit to much at this time. Again if you go down this alley post your problems here and we'll help out. Building a cluster over iSCSI is quite easy.

 

 

 

I'm not familiar with Hyper-V and hosting services virtually. I get the concept of running a virtual OS on top of the physical OS. I do that with my Mac at home running XP via VmWare Fusion.

Yep, well, its exactly the same, just on a different scale and in a more permanent role.

 

 

Greg

 

 

 

 

[falconmark]

No, the client looks up the domain via DNS and finds out which server it should authenticate to. I wouldn't worry to much about this particular issue unless you run into real issues. checkout this for an easy to understand guide.

 

 

 

I Prefer running DHCP on windows as we use other tools like SCCM/MDT etc which functions better when running windows to windows... If your worried you can run it on your netgear but its still a single point of failure. Given you seem to be a beginner, building a cluster to support this may seem a bit to much at this time. Again if you go down this alley post your problems here and we'll help out. Building a cluster over iSCSI is quite easy.

 

 

 

Yep, well, its exactly the same, just on a different scale and in a more permanent role.

 

 

Greg

 

 

Ok, so if I don't enable DHCP on the netgear, can both servers run DHCP or would that cause a conflict?

 

We also have a Dell PowerConnect 6248p managed switch that they were using previously before I came into the company. I think its not configured properly as our network seems to run very sluggish, even over gigabit connectivity. I checked all the client workstations and they all report 1GB connections. So i'm not sure if its the managed switch, router, cabling, or what.

 

I also want to be able to give priority to Peachtree traffic so that the accounting team has better response time with using the application, because right now its very slow to process transactions. A lot of employees stream music so i'm sure its killing a good amount of bandwidth.

 

Basically my goal is to avoid a single point of failure, unless the power goes out in the building.

[jamitupya]

OK, so your looking to avoid any large scale business outages.

 

You can run 2 x DHCP services however 1 would have to be offline. I would build the system like this

 

 

 

I dont have any experience with Peachtree or dell Switches however you should look up QoS (Quality of Service) and prioritize this traffic on your network. this can be done by the L3 Switching properties on the Dell. If you can do a config dump of the switch or if your able to factory reset it, this may help. I have experienced similar issues on cisco and we would just rebuild the QoS on the LAN/WAN. to do this we need more information on Peachtree but thats a whole different story.

More of an issue due to its nature as an accounting app.... if its http(s) based its easy, just prioritize http(s) to the specific IP address of the server are highest priority. I suggest also whilst your in QoS mode, limit the amount of bandwidth your users can use to stream audio. this can likely be completed on the UTM device quite easily.

 

 

one method to test your bandwidth is using a tool called iperf.exe. Run this on your server network and then from your client's do some testing with it. it will give you an idea as to if it is network throughput limitations. (im guessing a mis-configured switch) :-)

[jamitupya]

blagh, UTM25 not 250