Network Connection Time-Out over Server

Status
Not open for further replies.

ABAMOTO

Posts: 27   +0
Dear all:
I have an in-house W2K Server OS that has finance application software running on it. Recently a lot of my users are experiencing connection time-out problem when running that finance application's client software on the client machines. All the users can still go on internet and access local file sharing.

Is it a NIC hardware problem or it is a W2K Server DHCP problem? I have users who obtain IP from DHCP as well as fixed IP. Or the bandwidth is not big enough to handle all clients' usage (current 10/100Mbps base for 20 concurrent usage)?

Thank you all... >__<
 
If other things work OK for the same users then the problem is not with the NICs or the DHCP.

Maybe the server is under heavy load or the number of simultaneous connections to it is limited? Check the server logs to find out and check the CPU.

When the timeout occurs, can the clients still ping the server? Maybe you hav some faulty cables or networking gear somewhere between the clients and the server?
 
Additionally, in Windows 2000 inactive drive mapping connections disconnect after a short while. If the application is accessed via a network drive mapping, then often re-running the script, or simply opening the network drive will solve your application problem.

If the application is constantly contacting the server then this wouldn't be your issue. I would say check your cableing and make sure you aren't running any unnecessary services or applications.

Is the server only for the application, or does it do other things as well?

Edit: me grammar good.
 
inable caching for static info
internet description:
A Web cache sits between one or more Web servers (also known as origin servers) and a client or many clients, and watches requests come by, saving copies of the responses — like HTML pages, images and files (collectively known as representations) — for itself. Then, if there is another request for the same URL, it can use the response that it has, instead of asking the origin server for it again.

There are two main reasons that Web caches are used:

* To reduce latency — Because the request is satisfied from the cache (which is closer to the client) instead of the origin server, it takes less time for it to get the representation and display it. This makes the Web seem more responsive.
* To reduce network traffic — Because representations are reused, it reduces the amount of bandwidth used by a client. This saves money if the client is paying for traffic, and keeps their bandwidth requirements lower and more manageable.
 
Samstoned said:
inable caching for static info
internet description:
A Web cache sits between one or more Web servers (also known as origin servers) and a client or many clients, and watches requests come by, saving copies of the responses — like HTML pages, images and files (collectively known as representations) — for itself. Then, if there is another request for the same URL, it can use the response that it has, instead of asking the origin server for it again.

There are two main reasons that Web caches are used:

* To reduce latency — Because the request is satisfied from the cache (which is closer to the client) instead of the origin server, it takes less time for it to get the representation and display it. This makes the Web seem more responsive.
* To reduce network traffic — Because representations are reused, it reduces the amount of bandwidth used by a client. This saves money if the client is paying for traffic, and keeps their bandwidth requirements lower and more manageable.


The issue with this is it only pertains to internet or DMZ traffic. A proxy or cacheing server will only cache info from web sites because to get to the web site you are required to go through the cacheing server. If his internet access is fine, and the server is behind the gateway this won't help his situation.

Edit: I doubt his internet traffic is overwhelming his LAN traffic.

Edit again: Application information is hardly ever proxied in any case. This is because if the application reports data it will not be current. Additionally, the application would have to specifically support it. Application data isn't usually subject to proxy because, unlike web pages, there are no specific files to cache. It's just data that the application understands.
 
you may also want to try a compression manager
this help bandwidth also.
check your setting for dhcp ip renewal times may be to short
I run apache and use GZIP mod to compress the files on the fly
I also have a caching enabled Read
Use Server Cache Control to Improve Performance

Caching is the temporary storage of frequently accessed data in higher speed media (typically SRAM or RAM) for more efficient retrieval. Web caching stores frequently used objects closer to the client through browser, proxy, or server caches. By storing "fresh" objects closer to your users, you avoid round trips to the origin server, reducing bandwidth consumption, server load, and most importantly, latency. Apache server caching to save bandwidth and improve performance.

Caching is not just for static sites, even dynamic sites can benefit from caching. Graphics and multimedia typically don't change as frequently as (X)HTML files. Graphics that seldom change like logos, headers, and navigation can be given longer expiration times while resources that change more frequently like XHTML and XML files can be given shorter expiration times. By designing your site with caching in mind, you can target different classes of resources to give them different expiration times with only a few lines of code.
 
Samstoned said:
you may also want to try a compression manager
this help bandwidth also.
check your setting for dhcp ip renewal times may be to short
I run apache and use GZIP mod to compress the files on the fly
I also have a caching enabled Read
Use Server Cache Control to Improve Performance

Caching is the temporary storage of frequently accessed data in higher speed media (typically SRAM or RAM) for more efficient retrieval. Web caching stores frequently used objects closer to the client through browser, proxy, or server caches. By storing "fresh" objects closer to your users, you avoid round trips to the origin server, reducing bandwidth consumption, server load, and most importantly, latency. Apache server caching to save bandwidth and improve performance.

Caching is not just for static sites, even dynamic sites can benefit from caching. Graphics and multimedia typically don't change as frequently as (X)HTML files. Graphics that seldom change like logos, headers, and navigation can be given longer expiration times while resources that change more frequently like XHTML and XML files can be given shorter expiration times. By designing your site with caching in mind, you can target different classes of resources to give them different expiration times with only a few lines of code.


They run an application on a local lan. Caching will not work. Applications do not typically pass files. What you are talking about is caching web pages. Web pages consist of several individual files that can be cached.

Client/Server applications do not pass cacheable files. There is no logo image or header file or any of that stuff when the client app talks to the server and vice versa.

Additionally, the server would have to be on the other side of the local gateway, or they would have to have some serious DNS changes for anything to touch a proxy server in the first place.

Remember he is not talking about Web traffic. There isn't a bandwidth problem so long as his switches, and NICs are functioning correctly. This is a local lan/application issue.

Edit: Need spell checker
 
well since ABAMOTO does not use a program through a browser client
maybe it is how many client license's on machines?
also ,a question ,why even run a server if it is setup as closed network
if need a server for web app's keep that OS seperate from the
other one. no mention how the app run's
how ever if the clients are using an internet connect from somewhere other then the local net( dsl or dialup)
the server is still using or could use a cacheing system ,I would think it depended on the app ,needed to access.
I use access as data base and all static info is cacheable
is the app running on it's own built in server
do they (client) need to log in with password is the password set tp expire
my W$K server OS is sitting on a shelf ,way to many access issues.
some day I get time I'll load it on a small pc and play around
one thing to check all the nics should be the same for dynaminic ip
or they all should be set to same speed and duplex
just a n'other thought
 
ABAMOTO said:
Dear all:
I have an in-house W2K Server OS that has finance application software running on it. Recently a lot of my users are experiencing connection time-out problem when running that finance application's client software on the client machines. All the users can still go on internet and access local file sharing.

Is it a NIC hardware problem or it is a W2K Server DHCP problem? I have users who obtain IP from DHCP as well as fixed IP. Or the bandwidth is not big enough to handle all clients' usage (current 10/100Mbps base for 20 concurrent usage)?

Thank you all... >__<
All 3 of my 2K servers started doing just that, the only thing which worked for a temporary fix was turning off the screensavers, for whatever reasons the screensavers were halting the network connections despite no power saver modes were used and were actually disabled even in the drivers.

The permanent fix was installing 3 weeks old newly released network drivers.

It could be your problems and mine are the same.
 
I think Nein might be on to something. As Nodsu said, check the CPU utilization.

Check the CPU utilization on the servers and on the PC's both when you are having problems and when things seems to be OK.

Also, check the event logs on the server for error and warning messages pertaining to network, authentication, and your specific app. The program may have been written to write messages to the event log. You may also want to check the event log on the machines that are exibiting the errant behavior.

Does the problem only occur on a select few machines, or all of them?

well since ABAMOTO does not use a program through a browser client
maybe it is how many client license's on machines?

What do you mean by client licenses? The only client licenses that will cause connection problems in a Windows environment that I know of are terminal service CALs. If you are running terminal services and have machines that do not have a CAL or aren't already built in (Windows 2000 have built in CALs for Windows 2000 Terminal server) those machines will not be able to connect to the terminal server after the Temporary license expires. I think Windows temp CALs expire in 120 days.

also ,a question ,why even run a server if it is setup as closed network

I don't understand this question. All client/server applications use a server.

if need a server for web app's keep that OS seperate from the
other one. no mention how the app run's

Sorry, I don't understand this either. What do you mean?

how ever if the clients are using an internet connect from somewhere other then the local net( dsl or dialup)
the server is still using or could use a cacheing system ,I would think it depended on the app ,needed to access.

The server would have to be the local machine's gateway for caching to work. Or there would have to be some serious server side DNS changes to route that traffic to a caching server. Traffic on a lan stays on the lan until it can't find what it's looking for. Then it sends traffic to the default gatway. Regardless of the internet connectivity. You are correct in that an application has to specifically support caching to do so.

I use access as data base and all static info is cacheable

What static info is cacheable? I'm curious. I've never used caching with access.

is the app running on it's own built in server
do they (client) need to log in with password is the password set tp expire

If they work at all then there isn't an authentication issue. If the password expired it wouldn't let them into the app ever until the password was reset.

one thing to check all the nics should be the same for dynaminic ip
or they all should be set to same speed and duplex

This is probably not the issue as they can access other files on the network. The problem seems to be application specific, but it may be worth checking out.


Another thing to do is to plug the server into a different switch if possible, or another port on the same switch. I still think the issue isn't your network, but rather the app that you're running. Do you know if the app has any updates? Is your server up-to-date? Definitely check the Logs and processor utilization on the server.

Any news on this ABAMOTO? I'm curious if you've found your problem, and what it was/is.
 
Dear all:
Thank you for the generous knowledge sharing.

I am running MAS200 on a Win2K Server, which also acts like DHCP and domain controller. However, we currently do not have any network drive mapping.

I eliminated the root cause of Win Home Edition license (for some clients who got the connection time out warning) because I got the same warning msg while running on W2K OS.

I did not think it was the switch or router problem because all users can access internet instantenously (no delay found), and the connection time out problems occur to random users (therefore root cause is not because of particular switch or router port or cable being defective).

It could be my network card, therefore I reroute the traffic to network adapter #2 (original is #1). Since then, its been almost two days and no Connection Time Out Error is reported to me yet.

Speaking of CPU usage, the main process of MAS200 is running high. I am not sure whether it is due to heavy users' usage (14 users and more than 2/3 are in application program windows) or due to file lock caused by improper client-side shutdown. Currently I have AGAIN made a reminder to all my users to shut down the MAS200 program properly to prevent any file locking or data corruption.

I did not see any error msg regarding MAS200 in event logs.

I am not sure whether it was the change of NIC or CPU usage that is causing this Connection Time-Out. At least for now, the problem seems to come from the Server side. I will update the NIC driver, and monitor user behavior from now on. I will also look into concurrent connection limitation (it allows 50 now).

Thank you all for your input. I learned a lot -)
 
Hm. Cool. I'm glad you've got it working now.

I was just wondering, did you experienced any other problems on your network? It seems that if the NIC was going bad you would have had some authentication issues. I suppose since you are running XP home none of those machines are on your domain, but perhaps you could have had some printing problems.

Anything like that?

Thanks, and congrats on fixing it.
 
I have experienced other mysterious networking problems, but I am not sure whether it was the defective NIC that caused it:

1. client machine cannot see workgroups:
Under Network Neighborhood, client machine sometimes (like 1/5 times) can only view few of the total machines, and all the others' share folders are invisible as well.

*hotfix: In order to access other machines' share folders, just remember other machines' name and type in "\\machine_name". That will do.

2. Similar problem as #1. When client machine accesses server share folders, sometimes only root directories are visible. Once enter the root folders, either files or subfolders or both are not visible.

*hotfix: remember path name and type in address bar, praying that the files will be displayed.

3. Even for some w2k machines, I failed to add them onto domain: But maybe that is because i didnt do it right? I just select domain name and type in the username and password under Domain Server's active directory (that account belongs to domain controller group as well). But that does not do nothing. I keep getting an error either path is unreachable or something else...

My previous IT manager tried to upgrade all the P3 to P4. However, due to budget concern, he did not push the management to approve purchasing Windows XP Pro. With Home Edition, my workload is heavy on the Help Desk part... arghhh~~~~

At last but not least, I have a question about gigabyte LAN. If I add a gigabyte LAN onto my current Server, will it allow more bandwidth under the environment in which only 10/100Mbps based Switch and Router are used? Sometimes i feel 100Mbps LAN is not enough, especially when I do VNC (it sure is lagging like a turtle).

My current set up is

Server (10/100Mbps LAN) <---> 2x 24 Ports CISCO SWITCH (10/100Mbps), where my computer is plugged in to one of the ports <--> D-Link Firewall 4 port router (10/100Mbps) <--> Internet/Intranet

Because upgrading SWITCH might be difficult due to the fact it costs a lot of money, I am thinking, will the below setup work?

Server (adding Gigabyte LAN adapter card) <-->bypass switch and go directly to a Linksys Gigabyte LAN 4 ports router <--> Gigabyte router has one port connecting to CISCO Switch, and one port connect directly with my computer.

Will this give me Gigabyte LAn performance???

Thank you. Have a great weekend~
 
Sorry for such a late response.

3. Even for some w2k machines, I failed to add them onto domain: But maybe that is because i didnt do it right? I just select domain name and type in the username and password under Domain Server's active directory (that account belongs to domain controller group as well). But that does not do nothing. I keep getting an error either path is unreachable or something else...

Could you explain how you add machines to the domain again? I don't quite follow you.

When adding machines to a Windows 2000 or greater domain you do not have to do anything on the server really. You don't have to add machines manually. You add them on the pc through system properties. On the computer name tab click change and it's there where you change it's domain membership. When you supply the user account info when adding the machine it creates all the necessary entries in AD.

At last but not least, I have a question about gigabyte LAN. If I add a gigabyte LAN onto my current Server, will it allow more bandwidth under the environment in which only 10/100Mbps based Switch and Router are used? Sometimes i feel 100Mbps LAN is not enough, especially when I do VNC (it sure is lagging like a turtle).

My current set up is

Server (10/100Mbps LAN) <---> 2x 24 Ports CISCO SWITCH (10/100Mbps), where my computer is plugged in to one of the ports <--> D-Link Firewall 4 port router (10/100Mbps) <--> Internet/Intranet

Because upgrading SWITCH might be difficult due to the fact it costs a lot of money, I am thinking, will the below setup work?

Server (adding Gigabyte LAN adapter card) <-->bypass switch and go directly to a Linksys Gigabyte LAN 4 ports router <--> Gigabyte router has one port connecting to CISCO Switch, and one port connect directly with my computer.

Will this give me Gigabyte LAn performance???

Short answer is "Partly". Only the items directly connected to the gigabit ports on those switches will run at gigabit speeds. That is provided that all the machines connected have GB NICs in them. In any case, all the machines connected to the Cisco will still be at 100mbps max. So you and the server will be at 1000mbps max (provided you have a gb nic too) but not any of your client machines.

From what you've said about your LAN I doubt that your regular traffic is causing a problem. You may want to make certain that you don't have any zombies on your network, and make sure all your machine's virus defs are up to date. If you have any of those problems on your current network, upgrading the lan speeds isn't really going to help.

I don't think that you will notice a drastic change, if any, with VNC on a GB network unless you run bunches of sessions at a time. Those applications that provide remote sessions aren't designed with GB in mind. I run SMS 2003 at work and all the servers are on a GB switch and I haven't noticed a difference between server to server and server to client. I know the applications are different, but the principle is the same.

Hope this helps.
 
I added workstations to domain through workstations' system property. For server side, do I have to create computers under Active Directory first or?

Regarding the networking problem, is there anyway to detect whether there is so called "zombie" in my network?

I am pretty sure all my virus scans are up-to-date. Plus I am using symantec CE. I don't think there is a problem with virus scan occupying CPU unlike the Home Edition version (for general Anti-Virus softwares)

Also, any networking software such as monitoring LAN traffic by users that you recommand? But my dept does not have huge budget though...

Thanks a bunch =)

p.s. Do you work in a company as MIS/ IT? If so, does your company use EDI?
 
Sorry for the delay.

You do not have to do anything on the server to add a machine to the domain. Not unless you are running a flawed NT4 domain. You just need to provide an account that has permissions to add machines to the domain.

The easiest way to detect massive amounts of traffic is if you have a gateway server, to monitor the traffic on that server. You can use a product like NetMon that comes with windows (you get the full version with SMS), or ethereal which, if I recall correctly, is free. If you can find a machine that is constantly sending data to the internet, you may want to take that machine offline for a full virus/spyware sweep. There are certainly other ways to do this, and some cursory searching for 'network monitoring' should net you some decent ideas.

I use Symantec CE at my main customer's site and have had no problems with it. Sometimes it's worth checking to make sure the machines are getting the most up to date defs possible from the server. Additionally if you have any laptops that are out of the office more often than not they may have been set up in 'unmanaged mode'. This means that they don't get defs automatically from the server, nor is there an option to use symantec's live update to schedule updates. Someone has to open the client up and click update to get them.

For discovery and monitoring of machines I use Solarwinds Engineer's edition. The drawback to this is that it basically works on SNMP so if you don't have that running properly it's pretty much useless except in the discovery aspect. In addition to that I use netmon frequently as well. The only problem there is that it only monitors local NICs. That is to say that if you have it installed on a server it can only monitor traffic to and from a NIC on that server. So unless it's broadcast traffic, or if the server is the gateway for your machines you won't see much aside from local, and print traffic. I've heard good things about snort but I haven't used it.


I do work in the IT industry, but we don't really use any EDI. Most of our customers are not internet sales oriented. Sorry.
 
Previous IT set up a VPN to connect our Sales Branch with our Headquarter. However, this setup seems to be abandoned for ages. Meanwhile, our gateway is still set to that CISCO Server which is connected to a Loop machine (I am not really familiar with this VPN setup; I thought adding a VPN + firewall router would be good enough)

Because I do not have any light-loaded servers, if I take out this VPN gateway, my gateway will be the D-Link router (Yeah i know, i know..). As for this D-Link router, there is not much statistics on LAN traffic that I can monitor.

Thank you for the reply. =)
 
Status
Not open for further replies.
Back