[THIN] Re: DOS App Performance Issue
- From: "Steve Greenberg" <steveg@xxxxxxxxxxxxxx>
- To: <thin@xxxxxxxxxxxxx>
- Date: Wed, 2 Jul 2003 13:13:11 -0700
Is there any difference in the size/nature of the database that this app is
using between the sites?
Steve Greenberg
Thin Client Computing
34522 N. Scottsdale Rd. suite D8453
Scottsdale, AZ 85262
(602) 432-8649
(602) 296-0411 fax
steveg@xxxxxxxxxxxxxx
-----Original Message-----
From: thin-bounce@xxxxxxxxxxxxx [mailto:thin-bounce@xxxxxxxxxxxxx] On Behalf
Of george.wasgatt@xxxxxxxxxxxx
Sent: Wednesday, July 02, 2003 1:07 PM
To: thin@xxxxxxxxxxxxx
Subject: [THIN] DOS App Performance Issue
One of my customers is having what appears to be a rather mysterious
performance issue with our application running on Citrix boxes. Note that
this application is at multiple customer sites running under Citrix and this
is the first one that reports that they can reach less than half of the
number of users per server we told them they could get.
Hardware is multiple Compaq DL380 G# with dual 1.6GB processor, Raid array
and 2.5GB of RAM. They software is WinK Server with Citrix XP FR2.
This is a DOS application and the customer is running TAME 4.4. TAME-MON is
in the AUTOEXEC.NT and when I run TAMEVIEW I see the multiple DOS
applications running under TAME. They go in and out of sleep mode as you
would expect. CPU utilization for the sessions drop as expected when TAME
puts them to sleep. Overall CPU utilization can peak up to 80% but it tends
to hover around 30% or so even with 7 simultaneous users. Neither the
processors nor the CPUs look to be overloaded.
This customer reports that when simultaneous users goes up to about 10 that
screen to screen response time starts to suffer. So much so that the users
start abandoning their sessions. We find this inexplicable. At our site
and at multiple other sites we have seen equivalent servers handling 25 or
more users before anyone starts to complain.
So where is the bottleneck? The customer reports that the complaints are
from multiple branch offices - the slow downs cannot be blamed on bandwidth
to the offices. I asked if people on the local LAN and WAN networks report
slowdowns in other applications and got told no, the issue is with our
application on the Citrix boxes.
Where do I go from here? I could run PERFMON on the Citrix servers and file
servers and throw a load at them to see what I get for stats.
Suggesting are welcome, thanks.
Other related posts:
