4285ch02.fm Draft Document for Review May 4, 2007 11:35 am
60 Linux Performance and Tuning Guidelines
Example 2-25 Output of strace counting for system time
[root@lnxsu4 ~]# strace -c find /etc -name httpd.conf
/etc/httpd/conf/httpd.conf
Process 3563 detached
% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
25.12 0.026714 12 2203 getdents64
25.09 0.026689 8 3302 lstat64
17.20 0.018296 8 2199 chdir
9.05 0.009623 9 1109 open
8.06 0.008577 8 1108 close
7.50 0.007979 7 1108 fstat64
7.36 0.007829 7 1100 fcntl64
0.19 0.000205 205 1 execve
0.13 0.000143 24 6 read
0.08 0.000084 11 8 old_mmap
0.05 0.000048 10 5 mmap2
0.04 0.000040 13 3 munmap
0.03 0.000035 35 1 write
0.02 0.000024 12 2 1 access
0.02 0.000020 10 2 mprotect
0.02 0.000019 6 3 brk
0.01 0.000014 7 2 fchdir
0.01 0.000009 9 1 time
0.01 0.000007 7 1 uname
0.01 0.000007 7 1 set_thread_area
------ ----------- ----------- --------- --------- ----------------
100.00 0.106362 12165 1 total
For the complete syntax of the strace command, issue:
strace -?
2.3.16 Proc file system
The proc file system is not a real file system, but nevertheless is extremely useful. It is not
intended to store data; rather, it provides an interface to the running kernel. The proc file
system enables an administrator to monitor and change the kernel on the fly. Figure2-5
depicts a sample proc file system. Most Linux tools for performance measurement rely on the
information provided by /proc.