steps to determine the system performance?

[Jagadesh_2]

This question seems to be blank.. but still let me know what are the recomended process to find out where exactly the performace problem occur's (CPU / disk / network.. etc) and how to resolve the same. Hope this thread will be useful for others to know more about system performance tuning. All answers are welcome.

[Joseph Loo]

hi,

refer to this docID, UPERFKBAN00000726:

http://www1.itrc.hp.com/service/cki/search.do?category=c0&mode=id&searchString=UPERFKBAN00000726&searchCrit=exactphrase&docType=Security&docType=Patch&docType=EngineerNotes&docType=BugReports&docType=Hardware&docType=ReferenceMaterials&docType=ThirdParty&search.x=24&search.y=10

or docID, KBRC00000947:

http://www1.itrc.hp.com/service/cki/search.do?category=c0&mode=id&searchString=KBRC00000947&searchCrit=exactphrase&docType=Security&docType=Patch&docType=EngineerNotes&docType=BugReports&docType=Hardware&docType=ReferenceMaterials&docType=ThirdParty&search.x=24&search.y=9

regards.

[MarkSyder]

HP-UX comes with several useful monitoring tools built in: sar, top, vmstat, iostat.

sar = system activity report. Format:

sar x y where x = interval in seconds and y = number of reports.

E.g. sar 2 5 will produce 5 reports at 2 second intervals.

vmstat = virtual memory performance. Format is the same as sar.

iostat = input output statistics. Format is the same as sar.

top - show the processes that are taking up most of the processor time.

If you want to spend money on something more sophisticated, glance has a good reputation.

Mark Syder (like the drink but spelt different)

[Pete Randall]

The first step is to make sure you have Glance. No system should be without it. With Glance, it becomes a relatively simple matter to isolate the problem as to whether the bottleneck is CPU or disk or memory or network.


Pete

[Jagadesh_2]

Hi Joseph,
Thanks for the link.. but i am not able to open that link.
error:
technical knowledge base - no results found
Note: We're sorry but your search produced 0 results

[Jagadesh_2]

Hi Mark Syder,

Thanks for the information.. can you please provide more information, the limit values for sar/vmstat/iostat

for example
*************
In solaris

sar â g 2 10
â SunOS host1 5.8 Generic_108528-22 sun4u 09/01/03
â 12:32:04 pgout/s ppgout/s pgfree/s pgscan/s %ufs_ipf
12:32:15 307.42 679.68 674.30 78.29 0.00
12:32:25 393.52 788.89 777.87 1.76 0.00
12:32:36 344.87 875.09 843.66 17.63 0.00
â Average 348.64 781.05 765.14 32.57 0.00

If page scan = 0 : ALL Ok
If page scan > 0 & < 100 : memory is use â no problem
If page scan > 100 & < 1000 : more me

[Joseph Loo]

hi,

i have no problem accessing it from the Americas / Asia Pacific forum.

or u may like to search it at the technical knowledge base:

http://www1.itrc.hp.com/service/cki/enterService.do

based on the 2 docID i sent. if not search by the title, Introduction to Performance Tuning and System Performance Tuning Guide.


regards.

[MarkSyder]

sar produces idle time on the processor - the higher the better. I always like to see something in the region of 95%.

vmstat - the final column shows idle vm. 100 is the ideal figure. It's been said many times on this forum - your system shouldn't be swapping at all. swapinfo -tam is another useful tool for monitoring vm.

iostat - disc usage statistics. bps = kilobytes transferred per second, sps = number of seeks per second, msps = milliseconds per average seek. This will help to identify an overloaded disc.

Mark