So to follow along with some of Jim's comments, one of the methods we use is to test if the performance changes in different environments. Describing where you see the problem and where you don't see the problem.
For example, the problem is seen in /var/tmp and not seen in /tmp. Or if you see it in both list both and maybe you see don't see it on another server. Also your environment, how you configured the server, makes a big difference. I worked a problem in which the customer chose a the en_US.UTF-8 language instead of the Posix C language. Using the dtracetoolkit I was able to see the issue and it did not matter what type of file system they were using. See my blog on the issue. http://blogs.sun.com/blaze/entry/dtrace_for_troubleshooting Dave Sean Liu wrote: > Hi Jim, > > Thanks for your reply, and that's good information. > I actually did most of what you suggested. > I started with dtrace and seems every system call takes longer to run on the > filesystem when it is shared than not. So that's why I needed to understand > what actually happens in the kernel. > > Anyway I found color.d from Brendan's blog, which is quite helpful, it traces > into all the way into every module. > > Thanks again. > > Sean > > > This message posted from opensolaris.org > _______________________________________________ > perf-discuss mailing list > perf-discuss@opensolaris.org > _______________________________________________ perf-discuss mailing list perf-discuss@opensolaris.org
