EFRTech
asked on
Citrix Presentation Server 4.5 %Interrupt Time Metric Error
I have a Windows 2003 box running Citrix Presentation Server 4.5 with all hot-fixes installed. It is running on a HP DL370 G4 server that has a X3.6 800Mhz 1mb processor with 4GB of RAM and 3 72GB hard drives in a RAID 5 configuration. I keep getting a %Interrupt Time Metric Error. When I look at the graph, it shows peaks of 44%, 32%, and 25%. My metric thresholds are at the default level of 0.3%(yellow) and 0.4%(red). These peaks happen during the normal business day. I support 40 citrix users that access our application set which includes the desktop, SQL app, and Office apps. I have the server on a reboot schedule of 48hrs because of a printer auto-create problem.
Do I need to add the 2nd Xeon processor to overcome this metric error? Or do I need to adjust my metrics? What would be the correct metric settings for 40+ users? Would this metric error cause printer auto-create problems?
Do I need to add the 2nd Xeon processor to overcome this metric error? Or do I need to adjust my metrics? What would be the correct metric settings for 40+ users? Would this metric error cause printer auto-create problems?
ASKER
I checked the utilization of the processor, with both the Task Manager and with the HP System Manager, it averages out to 25%. I also ran diagnostics from the HP System Manager and everything checked out to be normal. Since adding 2GB of ram, we only had 2GB to start, the Interrupt Time has gone down 20% at all of the spikes. Do I need to consider getting the 2nd processor? Or should I be looking at a 2nd server to distribute some of the load?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
The definition of this metric:
"The processor is spending a large amount of time responding to input and output rather than user processing. A large value for interrupt time usually indicates a hardware problem or a very busy server."
0.3-0.4 is an incorrect default setting for this Metric.
Suggested values for Interrupt Time are 1.5 - 2.0. Anything consistently over 2 is considered a problem. Spikes above 2 but right back down aren't. Terminal Servers by nature will have much higher interrupting than normal servers or workstations. I had a look at Microsoft's reccommendation for this metric and they say not to worry about it unless if hits 15%. And then, only if it is around 90%.
Some things to look at:
Baseline your server to ascertain the correct metric settings...
http://www.johnkellett.co.uk/index.php/2006/07/28/configuring-the-citrix-load-evaluators/
Hope this helps!