PF usage

[Keith]

General Info:
Windows Server ENT 2003 SP2 Current with updates.
Page File 4095 on E: and 800 on C:
Clients are Windows XP SP2 and SP3
RDP is 6.0 and 6.1.
4 CPU
14 Gig Ram
This is a VMware guest on ESX 3.5
The image sits on a SAN with 4 Gig fiber connecting it to the ESX server.

I am trying to find a way to see how much memory is being used by each user.
I can go two weeks with people using the system with no problems. Anywhere
between 1 user to 30 users and never get over 4 gigs commit charge total.
Then it will climb to 9 gigs and hang all day. I check the server at the end
of the day and I am the only person on the system and it is still at 9 gigs.
It has not BS yet. I end up restarting it. Also when all the users logoff I
check the processes to see if any apps are hung. I am not seeing anything.
I am thinking a print problem but not sure.

 

[Cláudio Rodrigues [MVP]]

Before we try to find the culprit, please note that it is a best practice to
reboot a TS frequently and in some cases, daily.
The reason for that is not TS itself. Most of the time applications leak
memory and when running on a regular PC you do not notice as users usually
reboot their machines and more than that it is a single user using the app.
Once you bring TS to the picture now you may have 50 people using the app at
the same time. This will make it leak 50 times faster at least and the fact
you usually do not reboot the box will make that worse.
That is why we highly recommend as part of any TS deployment, a scheduled
reboot.

Cláudio Rodrigues
Citrix CTP

"Keith" wrote:

> General Info:
> Windows Server ENT 2003 SP2 Current with updates.
> Page File 4095 on E: and 800 on C:
> Clients are Windows XP SP2 and SP3
> RDP is 6.0 and 6.1.
> 4 CPU
> 14 Gig Ram
> This is a VMware guest on ESX 3.5
> The image sits on a SAN with 4 Gig fiber connecting it to the ESX server.
>
> I am trying to find a way to see how much memory is being used by each user.
> I can go two weeks with people using the system with no problems. Anywhere
> between 1 user to 30 users and never get over 4 gigs commit charge total.
> Then it will climb to 9 gigs and hang all day. I check the server at the end
> of the day and I am the only person on the system and it is still at 9 gigs.
> It has not BS yet. I end up restarting it. Also when all the users logoff I
> check the processes to see if any apps are hung. I am not seeing anything.
> I am thinking a print problem but not sure.
>

 

[Keith]

We just replace one of our old TS with this. The old one would reboot every
night at midnight. We never got around to setting this up on the new system
as we were looking to see why we had this problem. I will set it up tonight
but I still need to track this down. The first few weeks we had no problems.
Now I can go two weeks or one day? We plan on adding more users so this
could be a problem. Do you have ideas on how to track this down? I am
thinking print drivers but not sure. The apps are the same as the old one
and we never had this problem.
Thanks for the help


"Cláudio Rodrigues [MVP]" wrote:

> Before we try to find the culprit, please note that it is a best practice to
> reboot a TS frequently and in some cases, daily.
> The reason for that is not TS itself. Most of the time applications leak
> memory and when running on a regular PC you do not notice as users usually
> reboot their machines and more than that it is a single user using the app.
> Once you bring TS to the picture now you may have 50 people using the app at
> the same time. This will make it leak 50 times faster at least and the fact
> you usually do not reboot the box will make that worse.
> That is why we highly recommend as part of any TS deployment, a scheduled
> reboot.
>
> Cláudio Rodrigues
> Citrix CTP
>
> "Keith" wrote:
>
> > General Info:
> > Windows Server ENT 2003 SP2 Current with updates.
> > Page File 4095 on E: and 800 on C:
> > Clients are Windows XP SP2 and SP3
> > RDP is 6.0 and 6.1.
> > 4 CPU
> > 14 Gig Ram
> > This is a VMware guest on ESX 3.5
> > The image sits on a SAN with 4 Gig fiber connecting it to the ESX server.
> >
> > I am trying to find a way to see how much memory is being used by each user.
> > I can go two weeks with people using the system with no problems. Anywhere
> > between 1 user to 30 users and never get over 4 gigs commit charge total.
> > Then it will climb to 9 gigs and hang all day. I check the server at the end
> > of the day and I am the only person on the system and it is still at 9 gigs.
> > It has not BS yet. I end up restarting it. Also when all the users logoff I
> > check the processes to see if any apps are hung. I am not seeing anything.
> > I am thinking a print problem but not sure.
> >

 

[Cláudio Rodrigues [MVP]]

As a start you could track down which printer drivers you have on it and look
for more updated ones or using out-of-the-box Windows 2003 drivers for these
printers. You may lose some functionality on the printing but they are for
sure 100% compatible with TS.

Claudio Rodrigues
CEO, WTSLabs Inc.

http://www.wtslabs.com

Citrix CTP
Provision Networks VIP


"Keith" wrote:

> We just replace one of our old TS with this. The old one would reboot every
> night at midnight. We never got around to setting this up on the new system
> as we were looking to see why we had this problem. I will set it up tonight
> but I still need to track this down. The first few weeks we had no problems.
> Now I can go two weeks or one day? We plan on adding more users so this
> could be a problem. Do you have ideas on how to track this down? I am
> thinking print drivers but not sure. The apps are the same as the old one
> and we never had this problem.
> Thanks for the help
>
>
> "Cláudio Rodrigues [MVP]" wrote:
>
> > Before we try to find the culprit, please note that it is a best practice to
> > reboot a TS frequently and in some cases, daily.
> > The reason for that is not TS itself. Most of the time applications leak
> > memory and when running on a regular PC you do not notice as users usually
> > reboot their machines and more than that it is a single user using the app.
> > Once you bring TS to the picture now you may have 50 people using the app at
> > the same time. This will make it leak 50 times faster at least and the fact
> > you usually do not reboot the box will make that worse.
> > That is why we highly recommend as part of any TS deployment, a scheduled
> > reboot.
> >
> > Cláudio Rodrigues
> > Citrix CTP
> >
> > "Keith" wrote:
> >
> > > General Info:
> > > Windows Server ENT 2003 SP2 Current with updates.
> > > Page File 4095 on E: and 800 on C:
> > > Clients are Windows XP SP2 and SP3
> > > RDP is 6.0 and 6.1.
> > > 4 CPU
> > > 14 Gig Ram
> > > This is a VMware guest on ESX 3.5
> > > The image sits on a SAN with 4 Gig fiber connecting it to the ESX server.
> > >
> > > I am trying to find a way to see how much memory is being used by each user.
> > > I can go two weeks with people using the system with no problems. Anywhere
> > > between 1 user to 30 users and never get over 4 gigs commit charge total.
> > > Then it will climb to 9 gigs and hang all day. I check the server at the end
> > > of the day and I am the only person on the system and it is still at 9 gigs.
> > > It has not BS yet. I end up restarting it. Also when all the users logoff I
> > > check the processes to see if any apps are hung. I am not seeing anything.
> > > I am thinking a print problem but not sure.
> > >

 

[Keith]

We just replace alot of printer so I am going in and remove the drivers from
all the old printers and will also review what we have installed.
Thanks

"Cláudio Rodrigues [MVP]" wrote:

> As a start you could track down which printer drivers you have on it and look
> for more updated ones or using out-of-the-box Windows 2003 drivers for these
> printers. You may lose some functionality on the printing but they are for
> sure 100% compatible with TS.
>
> Claudio Rodrigues
> CEO, WTSLabs Inc.
>
> http://www.wtslabs.com
>
> Citrix CTP
> Provision Networks VIP
>
>
> "Keith" wrote:
>
> > We just replace one of our old TS with this. The old one would reboot every
> > night at midnight. We never got around to setting this up on the new system
> > as we were looking to see why we had this problem. I will set it up tonight
> > but I still need to track this down. The first few weeks we had no problems.
> > Now I can go two weeks or one day? We plan on adding more users so this
> > could be a problem. Do you have ideas on how to track this down? I am
> > thinking print drivers but not sure. The apps are the same as the old one
> > and we never had this problem.
> > Thanks for the help
> >
> >
> > "Cláudio Rodrigues [MVP]" wrote:
> >
> > > Before we try to find the culprit, please note that it is a best practice to
> > > reboot a TS frequently and in some cases, daily.
> > > The reason for that is not TS itself. Most of the time applications leak
> > > memory and when running on a regular PC you do not notice as users usually
> > > reboot their machines and more than that it is a single user using the app.
> > > Once you bring TS to the picture now you may have 50 people using the app at
> > > the same time. This will make it leak 50 times faster at least and the fact
> > > you usually do not reboot the box will make that worse.
> > > That is why we highly recommend as part of any TS deployment, a scheduled
> > > reboot.
> > >
> > > Cláudio Rodrigues
> > > Citrix CTP
> > >
> > > "Keith" wrote:
> > >
> > > > General Info:
> > > > Windows Server ENT 2003 SP2 Current with updates.
> > > > Page File 4095 on E: and 800 on C:
> > > > Clients are Windows XP SP2 and SP3
> > > > RDP is 6.0 and 6.1.
> > > > 4 CPU
> > > > 14 Gig Ram
> > > > This is a VMware guest on ESX 3.5
> > > > The image sits on a SAN with 4 Gig fiber connecting it to the ESX server.
> > > >
> > > > I am trying to find a way to see how much memory is being used by each user.
> > > > I can go two weeks with people using the system with no problems. Anywhere
> > > > between 1 user to 30 users and never get over 4 gigs commit charge total.
> > > > Then it will climb to 9 gigs and hang all day. I check the server at the end
> > > > of the day and I am the only person on the system and it is still at 9 gigs.
> > > > It has not BS yet. I end up restarting it. Also when all the users logoff I
> > > > check the processes to see if any apps are hung. I am not seeing anything.
> > > > I am thinking a print problem but not sure.
> > > >

 

[TP]

Hi,

Is this 32-bit or 64-bit 2003?

Have you used Task Manager to see which process or processes
have a high amount of committed memory? You can do this by
turning on the VM Size column, View-->Select Columns-->Virtual
Memory Size. There are certain cases where you will not be able
to see a process that is using a large chunk of commited, one example,
if the process creates a section object backed by pagefile, however,
most of the time the VM Size column will be all you need to find the
hogs.

Just for your reference, the PF Usage number is not the amount
of the page file that is in use. You did not say that, but I thought I
would mention it since many people get confused.

Thanks.

-TP

Keith wrote:
> General Info:
> Windows Server ENT 2003 SP2 Current with updates.
> Page File 4095 on E: and 800 on C:
> Clients are Windows XP SP2 and SP3
> RDP is 6.0 and 6.1.
> 4 CPU
> 14 Gig Ram
> This is a VMware guest on ESX 3.5
> The image sits on a SAN with 4 Gig fiber connecting it to the ESX
> server.
>
> I am trying to find a way to see how much memory is being used by
> each user. I can go two weeks with people using the system with no
> problems. Anywhere between 1 user to 30 users and never get over 4
> gigs commit charge total. Then it will climb to 9 gigs and hang all
> day. I check the server at the end of the day and I am the only
> person on the system and it is still at 9 gigs. It has not BS yet. I
> end up restarting it. Also when all the users logoff I check the
> processes to see if any apps are hung. I am not seeing anything. I
> am thinking a print problem but not sure.