This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: systemtap memory usage
- From: fche at redhat dot com (Frank Ch. Eigler)
- To: Martin Hunt <hunt at redhat dot com>
- Cc: systemtap at sources dot redhat dot com
- Date: 16 Feb 2007 16:39:38 -0500
- Subject: Re: systemtap memory usage
- References: <1171640857.5679.27.camel@dragon>
hunt wrote:
>[...] But wait. We also have "stap" hanging around while the probe
>runs. That's a huge 71Mbytes.
Heck, it could be even bigger if one's script invokes considerable
line record searching.
> [...] Clearly we should rethink having "stap" hang around.
Well, user-space memory is a much less scarce commodity than kernel
memory. Plus staprun alone is a fully supported way of re-executing
probe modules, in which case stap memory consumption does not arise.
It also turns out that stap does very little work after it launches
staprun. It could release most of its dynamic memory by that time.
(Hm, this is crying out for a mudflap memory-usage analysis!)
Actually, that could result in the sudden provision of free kernel
memory, just at the right moment of need.
If it weren't for desiring to print last-gasp messages like the pass-5
times, and to clean up temporary directories, it would need not stay
around at all and could just exec sudo et al.
- FChE