Re: [RFC] Framework for easy distribution of SystemTap scripts (V4)

["Frank Ch. Eigler" <fche at redhat dot com>]

Hi -

On Wed, Apr 14, 2010 at 09:48:21PM +0530, Anithra P Janakiraman wrote:
> [...]
> >Well, if you are targeting this to such a specific subset of users,
> >then perhaps there is little need to include the script within the
> >systemtap source tree.  It could be as easily downloaded from the wiki
> >for example, no?
> 
> That should be ok too..

OK, in that case, feel free to post it to the wiki as/when you like.


> >>With the -start option, staprun executes in the background while the
> >>script exits. To stop users need to run the package with -stop, it would
> >>not be possible to store the child pids in this scenario.
> >
> >(If you used the normal flight-recorder mode startup/shutdown options,
> >the pids wouldn't be needed, since the module name is used as the
> >key.)
> 
> We are assuming that SystemTap might not be installed on the machine 
> where the package will be executed. We package the bare minimal 
> (staprun/stapio) needed to run the module. The initscripts might not be 
> available.

I wasn't talking about initscripts, but about stap's -F and staprun's
-L/-A options.


> >>[...]  The config file and parameters should have been passed to
> the post-processing script. I had completely overlooked that.  [...]

OK.  It almost sounds like what you'd need is another post-processing
option, where the user specifies a command line (rather than a file
name).  Then the 'ipaddress' sorts of filtering would be simply:

    stap-makepkg --filter-output 'grep 127.0.0.1' .....

with no post-processing script/file.


- FChE