Link to home
Start Free TrialLog in
Avatar of RJLemon
RJLemonFlag for Canada

asked on

Klunky keyboard echoing in Terminal Services Sessions

Good Morning (at least for me) Experts,

I have approx 250 users that login to 4 TS boxes.  Two are TS and nothing else two are TS + DNS + AD, probably not the best scenario but we do with what we have.

The boxes are for the most part identical Dual Xeon 3.0 with 6 GB RAM and a pair of Raptor drives in a mirror configuration.

On average we have 60 users / box with each user having at least 25 MB of our apps in RAM + what ever windows uses for each session (from some MS docs I would class these users as knowledge workers so that should be about 10 MB of RAM, can anyone confirm or deny this is close?).

I have one group of customers saying that when they type larger sections of text (product descriptions and things with more than 1 or two words like names, etc) the echo response from when they type to when the text on the screen appears can have 2+ second delays.  It seems this is making it difficult to navigate as you have to keep going back to make sure the text has all appeared before moving on.

Most users are coming from behind a firewall of their own and this particualar group claim they can attach to other TS servers in other buildings and not experience this delay.  I have not really been able to experience the lengthy delays they claim to be getting.  I see little blips now and again (open  notepad and hold a key down and watch the text scroll) but they claim it actually appear chunky in this same basic test).

I have a firewall with forwards 3389 to a 2x load balancer with 4 servers, it then send ths data off to the TS box.

Anyone have any suggestions?  How can I tell if I am in a RAM shortage causing excess swapping?  My % usage of page file is no more than 10-15%, of 6G if RAM Memory/Available RAM is 4271MB with 40 users on now and CPU utilization rarely goes above 40.  If you need more please let me know.  These folks are driving me nuts cause I can't reproduce the issue and have not been able to find anything that would point me in any direction.

Begging may not be becoming or required but I am begging please someone poke a small hole in my blinders so I may see which direction to head or at least not to head.

Thanks.
Avatar of carl_legere
carl_legere

are you using office products?  disable spell check / grammar check as you type.

I built a TS with hardware mirrored drives and suffered the same problem.  Mirroring is not good for frequent writes, which your paging file is doing.  Disable paging or reconsider the mirroring, or both.

By definition a terminal server should be doing absolutely nothing else, hence you can run it with one drive, or two drives striped for enhanced speed.  Redundancy is irrelevant.  You should have a ghost or acronis image of these machines available to instantly rebuild one if there is a crash.

Make SURE your problem is based on users by comparing memory and drive queue values from times when there is 30 users with 60 users.  Try to weight the system down, you will find your problem.  (one easy way to do this is to run NT backup on the c: to a newly created folder on C: drive, generally overwhelming your drive queue.)

You have to find the circumstances under which you can observe the problem.

I found lots of cool hints from using Sysinternals' Process Explorer, which revealed that I should not have installed Mcafee Antivirus (or any of them) on a TS.  You achieve virus protection by making sure that the ts isn't storing any data for users, it is all redirected and the fileserver(s) where data is stored is protected.

Also- there is volumes of more information out there on this, I'm giving you some boiled down ideas, ymmv.
Is this one group of customers all connecting to you from the same or similar internet connections or ISPs? It could be a bandwidth or ISP issue on the their end. Find out how many users are using their internet connection and how much bandwidth they are paying for from their ISP.
Avatar of RJLemon

ASKER

Howdy pkutter,

It is one group of customers but an outside consultant, theirs not ours, has been in and said that the issue does not exist when they connect to other servers from their locations only when they connect to ours.  I have no idea where these other servers are or the configuration or load but once again I can only provide you with the information or lack there of that I have been provided.

I guess that is what makes this so frustrating.
Avatar of RJLemon

ASKER

Howdy carl legere,

I am going to see if I can remove the mirror from one of the servers and see if that makes it any better.  Then I will try the next step of moving the swap file to the old mirror drive to get it off the OS drive to see what that will do for me.  It is going to take me a couple of days to get through both steps as these are production servers, not a great way to test but on the other hand a very real test, so I have to make the changes after hours both times and then give the users a chance to try.  I will let you know as soon as I can.  In the mean time if you come up with anything else please let me know.

Thanks.
Hi RJLemon,
Maybe I'm not understanding exactly what's going on. I'm assuming that you have one group of users connecting to terminal server 1 that are functioning just fine. You have a second group of users that connect to the same terminal server that are functioning poorly. Unless the set of applications that they are running are different it doesn't sound like a server problem to me. This would seem to imply either a network problem or a client configuration problem. Is it possible that in the RDP client under the experience tab they have specified the network speed as 10Mb or higher, or any of the other options in that window?

Just some thoughts, good luck.
Avatar of RJLemon

ASKER

Hi pkutter,

You are understanding it correctly.  I have users from approximately 150 sites and of these lets say 20 to 25 sites worth of users are claiming to experience this issue.  All of these users from these 20 to 25 sites come from behind a single firewall.  No one else has communicated that they are experiencing this difficultly so that either means that they are not or they just have not yet said they are.  

Now that you mention client settings I believe these users are the only ones using the active x control to attach to our servers.  I will verify what the settings are for these users and see what happens.

Thanks for the suggestion.
You could also call some of your other clients, maybe some power users if you have any, and see if they are having any similar problems. You wouldn't have to come out and ask them if they noticed this problem, just treat it more like a courtesy call. It may give you quick access a little bit more information to find out if it's system wide or just this other set of users.
Avatar of RJLemon

ASKER

After doing some polling we have started to get the odd other person with the issue and I have been able to successfully reproduce the problem.  I had to travel 2 provinces West to do it but hey.  I hope we can now pin down the issue.
ASKER CERTIFIED SOLUTION
Avatar of pkutter
pkutter

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of RJLemon

ASKER

Not right now but it is still driving me nuts.  Now at least I know it is not issolated.  Good or bad I am not sure.