Can Navision Financials 3.01 be run on two or more Windows 2000 Advanced Servers with Load Balancing installed? I talked with Microsoft and Navision at COMDEX Las Vegas and neither could verify whether or not the above would work. Here is what I would like to do: Two HP NetServers (LP1000R or LP2000R) both with Windows 2000 Advanced Server both running “Load Balancing” component within the Windows 2000 Advanced Server. Both machines would access a SAN device. This way either machine could go down and we would still be operational on the other. We can add servers as # of users increases. We have about 50 users total. 20 are in Chino, Calif. 10 in Canton, Mass. 10 in Marlboro, Mass. 5 in Santa Barbara, Calif. 5 in Crowborough, England Right now, the Navision server is in Chino. Both Mass locations are via frame relay. All others go via DSL / Internet. Remotes come in on Citrix Metaframe now. Advice anyone? Thanks! Scott Faucher Anytime Tech Support, LLC Scott@AnytimeTechSupport.com (760)846-0410 24/7 cellular
Hi Scott did you set up that system ? Richard
No. In fact, I have been trying to find anyone at Microsoft who can show me some real world numbers, benchmarks, etc. so I can make an informed decision on what hardware to buy. Navision says “buy as much RAM as you can afford” and “buy the fastest CPU you can” but this does not help much. It is possible that a P3 server for $5,000 does 95% of what we want it to and that extra 5% may cost an additional $45,000 for a supercomputer. In addition, I’m trying to figure out if a “beefy” Windows 2000 Advanced Server using RDP can take the place of our existing two servers (one for Navision and one for Citrix Metaframe using ICA). From what I have read, the RDP protocol will hog too much bandwidth and we are best to stay with two separate servers. Hopefully someone from Microsoft will work with me to figure this out and make this information available to everyone. I believe there is a total lack of information when it comes to Attain server sizing and how much bandwidth Attain requires via RDP vs via ICA. - Scott
Hi Scott, firstly, this is not a really big system, so you don’t need to much of a server. Though users are not everything, you may have 49 people entering sales orders in which case it could be a big system. All Navision NSCs have the experience to tell you the size of machine you need, so speak with your NSC, and get them involved, but just guessing from what your brief spec here. Your Navision server needs about : 0.5-1 G ram. 10 hard drives. - 2 mirrored for the OS, - and the Other 8 as 4 mirrored pairs for 4 database segments CPU single (ALWAYS) and just about anything, say a Piii 500Mhz is plenty Get a really fast network card to link to Citrix there may be value in a 1 G link. Citrix Metaframe on W2k Terminal Server- 2-3 CPUs 1.5 - 2 G ram 2 drives mirrored. Use ICA clients 30 users coming in over the net on a T1 should be OK, but think about haveing some reserve if you expand that. Also you need to address the printing issue, that will eat your bandwidth pretty fast.
Sorry, I forgot about the load balancing. SANs and Navision do not mix well. You will definitely see performance issues here. You are going to spend a lot of money getting the connection fast enough to make sense. I have seen Navision running on dual servers sucessfully, but most people duplicate the whole machine, and this si for redundancy only, not for performance. I can’t see how Navision could make use of multiple servers, since the server only launches one thread when it runs. My recommendation would be to do a cost analysis of the down time, and get your self into a 2 hour rebuild situation, and save the money now.
I’ll second David’s advice on this one. From my observation, if you have any communication problem at all, it’s liable to create havoc with the application because the load balancing doesn’t handle that situation very elegantly. You also need to consider the ongoing administrative costs (time) of dealing with a more complex infrastructure, as well as the usual business issues of how much down time you can stand, etc. As long as your Navision server isn’t your PDC, you can probably put a backup machine on line a lot faster and easier - and minimize the complexity of the system.
And one other note, DONT put Citrix and Navision on the same machine, no matter how powerfull that machine is. Also Allen hinted to it, but definietely make sure that the PDC is another machine. You can use it as your file server and if you absolutely must (though I don’t recommend it) you could also put Excange on it if it is a beefy enough machine. If you need to use one of the machines as a BDC, then use the Citrix, not Navision.
I don’t really understand what goes down, is it the server or the client part. Her is my experience: In our case, we have the NF server on one machine and the client part is on the citrix server. To be sure that users can access Database, we have 2 citrix servers using load balancing, so if a server go down, the second is use. To prevent the bandwith saturation for printing i use Thinprint server on each citrix server. Now if the Database server goes down, i just restart it manually. My problem today is =================== How to restore the database if the database became corrupt, i made a simulation and i need about 4 hours to restore and reindex with the fbk (the only sure backup).
Scott, we are running Attain2.6 load balanced over IBM X350 Quad Xeon processors with redundant device Windows 2000 advanced server. Delivering via 4 IBM X330 servers running Terminal Services (TS). This is into 8 x 18Gb raid 1 disc cluster. All fronted by a 1Gb switch then to the network. We deliver to 180 TS clients across the state of Western Australia and 40 plus LAN connections. It works for us. Costly exersize but it depends on your needs. Talk to your NSC and hardware vendor.
Hi, I can tell you about our system and by that give you an idea what you need for yourself. We use Navision Attain 3.10A (C/SIDE). About 50 users where about 40 are putting in orders all day long. 1 Navision Server: 1 cpu(normal CPU when we bought it a year ago) with 1.5GB RAM. We use a SAN device where we have 8 disks in RAID 1(Database splitted on four drives). The San is only connected to the navision server. Connections is with Fibre channel. 2 citrix servers (2cpus and 1.5GB ram) with loadbalancing. We have Gigabit between the three servers. We think we have very performance for the users with this solution. We have also made an Navision server installation on one of the metaframe servers, so that if the Navision server breaks down then we will be able to replace it very fast because that the database are on the SAN and not on the actual Navision server. Our only concern is if the SAN breaksdown. But then we have made a plan for taking out the disks from the san, inserting them one and one into the Navision server and copy the databasefiles on one big harddrive that will fit all the files. Then we are able to start Navision again, but of course with poor performance(4 databasefiles are on one harddrive). But atleast we can keep up work. Hope this can be helpful. Regards Daniel
We have a problem our configuration - 2 servers Citrix v1.8 win2000srv MétaFrame ( 2GO ram , 2 Processors 800Hmz ) in load-balancing - 1 server Navision (Navision 2.60d version) win2000srv ( 2GO ram , 4 Processors 800Hmz ) This server is also used as FTP server as well as data server. All are interconnected via optical fibre 1000mbps At present we have losses of connection with the server Navision 2.60d The users concerned by this errors are connected via Citrix. The others connect to the server directly without passing by Citrix. The first disconnection takes place so: A window appears with as message “Error 59 UnexpectedNetworkError” After this disconnection, the application does not start any more on the Client and we obtain Dr Watson’s message in " Event Viewer " of the Citrix Srv: The application, , generated an application error. The error occurred 26/09/2002 @ 13:53:58.390 The exception generated was c0000005 at address 0043BC3D () Not possible to open Navision app During neighbourhood 3 quarters of an hour unless rebooting the Navision server. On the other hand, it happens that an user is not ejected: we finds in the directory who are files *.zup a named file " __ temp __.@ Uu " dated on 12/09/2002. When one delete it and that this user opens again Navision this file reappears with the same date on 12/09/2002. On the 12/09/2002 is the date we reinstalled Navision from NT4’s towards Windows 2000 server: During it this reinstallation we changed is Netbios name and the IP adress. was Somebody already confronted with this kind of problem or with a part and could help us to give ourselves information. Thank you Thanks