stap-merge man page

stap-merge — systemtap per-cpu binary merger

Synopsis

stap-merge [ Options ] [ INPUT FILENAMES ]

Description

The stap-merge executable applies when the -b option has been used  while running a stap script.  The -b option will generate files  per-cpu, based on the timestamp field. Then stap-merge will  merge and sort through the per-cpu files based on the timestamp field.

Options

The systemtap merge executable supports the following options.

-v

Verbose mode, displays three extra fields per set of collected data.

[cpunumber,sequencenumberofdata,thelengthofthedataset]
-o OUTPUT_FILENAME

Specify the name of the file you would like the output to be  redirected into.  If this option is not specified than the output will be pushed to standard out.

Examples

$ stap -v -b -e 'probe syscall.open { printf("%s(%d) open\n",
execname(), pid()) }' 

This should result in several stpd_cpu files (each labled with a number  representing which cpu the file was produced from).

$ stap-merge -v stpd_cpu0 stpd_cpu1

Running the stap-merge program in the same directory as the stap  script earlier in the example, will produce an ordered sequence of  packets with the three part label for each set of data.  This result will be pushed through the standard output.  An output file  could have been specified using the "-o" option.

Files

Important files and their corresponding paths can be located in the

stappaths (7) manual page.

See Also

stapprobes(3stap),
stappaths(7),
staprun(8),
stapvars(3stap),
stapex(3stap),
stap-server(8),
gdb(1)

Bugs

Use the Bugzilla link of the project web page or our mailing list. http://sourceware.org/systemtap/,<systemtap@sourceware.org>.