tsadmin taking up cpu usage

[Guest John R]

Hello,

 

I hope maybe someone here can point me in the right direction.

 

I have about 120 terminal servers in my datacenter, anywhere from 4 to 12

dedicated to each customer. My support people here (a group of 6 analysts)

run tsadmin on a single management server. They are in the habbit of

connecting to all of the customer TSs in tsadmin and staying in all day.

This has never really been an issue.

 

All of the TSs are on the same lan subnet, and are connected via cisco

managed switches. All hardware is identical, and all ports on the switches

are configured the same.

 

Recently however, the management server has been over-utilizing cpu usage,

remaining at 95-100% almost constantly. Investigation has shown that

tsadmin for each of the support people is using almost all of it. Further

investigation has shown that if I have them disconnect from a particular

group of 12 TSs, that usage returns to normal. If I connect only to that

group of 12, then my usage begins to hog the cpu. So, it would appear the

issue is with a particular group of 12 TSs, although I can't seem to narrow

it down any further. (Yes, the 12 are all related to the same customer).

Change control notes show no activity on this group of TSs lately, yet the

problem only appears to have started within the last week or two.

 

So far, the only tool I have been using trying to diagnose this issue is

task manager. I really don't like playing with the customer environment

without going through our change control process which involves group

decisions for things such as software and driver installs, nic

configurations, even cat-5 wire replacement.

 

If anyone could point me towards how I might try to diagnose this issue

further, I would really appreciate it. I will admit up front that I am

still learning the terminal services environment.

 

Thanks in advance for any advice you might offer.

 

John R

[Guest John R]

Re: tsadmin taking up cpu usage

 

 

"John R" <jsr^^^813@zoom^^^internet.net> wrote in message

news:%23ugEcrLlIHA.5280@TK2MSFTNGP02.phx.gbl...

> Hello,

>

> I hope maybe someone here can point me in the right direction.

>

> I have about 120 terminal servers in my datacenter, anywhere from 4 to 12

> dedicated to each customer. My support people here (a group of 6

> analysts) run tsadmin on a single management server. They are in the

> habbit of connecting to all of the customer TSs in tsadmin and staying in

> all day. This has never really been an issue.

>

> All of the TSs are on the same lan subnet, and are connected via cisco

> managed switches. All hardware is identical, and all ports on the

> switches are configured the same.

>

> Recently however, the management server has been over-utilizing cpu usage,

> remaining at 95-100% almost constantly. Investigation has shown that

> tsadmin for each of the support people is using almost all of it. Further

> investigation has shown that if I have them disconnect from a particular

> group of 12 TSs, that usage returns to normal. If I connect only to that

> group of 12, then my usage begins to hog the cpu. So, it would appear the

> issue is with a particular group of 12 TSs, although I can't seem to

> narrow it down any further. (Yes, the 12 are all related to the same

> customer). Change control notes show no activity on this group of TSs

> lately, yet the problem only appears to have started within the last week

> or two.

>

> So far, the only tool I have been using trying to diagnose this issue is

> task manager. I really don't like playing with the customer environment

> without going through our change control process which involves group

> decisions for things such as software and driver installs, nic

> configurations, even cat-5 wire replacement.

>

> If anyone could point me towards how I might try to diagnose this issue

> further, I would really appreciate it. I will admit up front that I am

> still learning the terminal services environment.

>

> Thanks in advance for any advice you might offer.

>

 

I got it. It seems that the users were setting their view to the domain so

that all users on all terminal servers were listed in the right pane.

Although this made finding a particular user really easy, it was causing

tsadmin to go nuts querying 120 servers and sorting the info. When they

selected a single terminal server to display, the cpu usage returned to

normal.

 

It seems one of them 'discovered' this the other day and told everyone else

"Hey, you can see ALL users".

 

A little bit of warning was issued to them to keep that to a minimum and

problem solved.

 

Thanks to anyone that read this and gave it some thought anyway.

 

John R