RE: stress testing

2004-01-05 Thread Poras, Henry R.
Jared, Are you talking about yapppack? I've been using that for a while (nice display. Though like statspack it is system wide so I usually just look for high level stuff and changes). Not aware of a patch though. With most peoplesoft applications I have seen, the bottlenecks aren't

RE: stress testing

2004-01-05 Thread Poras, Henry R.
John, Thanks for the tip. I've used sar and vmstat, but not in enough depth to have any preferences. So far I don't have permissions for sar at this site, but I should be able to get that. Henry -Original Message- John Kanagaraj Sent: Friday, January 02, 2004 2:29 PM To: Multiple

RE: stress testing

2004-01-05 Thread Jared Still
The patch I refer to is one I made that didn't make it into the most recent version of yapppack. YP uses an array as internal storage, and walks through it with a for i in 1..n loop. Since arrays are sparsely populated there is a fair chance of hitting an array element that does not exist. The

RE: stress testing

2004-01-05 Thread Poras, Henry R.
Good catch on the array. I never noticed that. Henry -Original Message- Jared Still Sent: Monday, January 05, 2004 10:59 AM To: Multiple recipients of list ORACLE-L The patch I refer to is one I made that didn't make it into the most recent version of yapppack. YP uses an array as

Re: stress testing

2004-01-05 Thread tjambu_fatcity
Hi Tim Tony Jambu here. Saw your posting to Oracle-l with regards to your sp_vmstat.sh script. I am not sure if know but I write a regular hints tips column for Select Journal. I read your article and would like to mention your script and point people to the script. Do you mind if I mention it

Re: stress testing

2004-01-03 Thread Tim Gorman
Henry, I use the attached shell script to gather and store VMSTAT information in a custom table within the PERFSTAT schema (i.e. schema belonging to STATSPACK). Allows for some nice reporting over time, rather than anecdotal here-and-there observations. Should work OK on Solaris, HP, and Linux.

Re: stress testing

2004-01-02 Thread Jared . Still
As the ultimate indicator of performance is response time, you might like to investigate YAPP at http://www.miracleas.dk/. The data generated gives a good indicator of response time from a database perspective. If you use it, ask me for the patch. Jared Poras, Henry R. [EMAIL PROTECTED]

RE: stress testing

2004-01-02 Thread John Kanagaraj
Henry, Sar is a better tool than vmstat/iostat as it collects a broad range of information. Specifically, sar -q should show up CPU queueing and swapping, and sar -v will show up file/process table overflow issues that may occur during stress testing. IMHO, sar is quite underutilized ( had a

RE: stress testing

2004-01-02 Thread Karniotis, Stephen
recipients of list ORACLE-L Subject:RE: stress testing Henry, Sar is a better tool than vmstat/iostat as it collects a broad range of information. Specifically, sar -q should show up CPU queueing and swapping, and sar -v will show up file/process table overflow issues that may occur during

Re: stress testing

2004-01-02 Thread Tanel Poder
And when you're doing stress testing, make sure you load your system for quite long time. Stress testing shouldn't only verify system performance, but it should also test system's reliability. One of my clients had a problem with Oracle Apps, where they did a stress test for just one hour and