On Mon, 2014-04-28 at 17:33 +0200, Radovan Garabik wrote: > On Mon, Apr 28, 2014 at 04:06:48PM +0100, Ben Hutchings wrote: > > On Mon, 2014-04-28 at 16:21 +0200, Radovan Garabík wrote: > > > > > > * Package name : verynice > ... > > > An utility to throttle long running processes beyond what can be > > > achieved by nice(1). This is done by periodically suspending and > > > resuming the process and its child proceses. > > > > This really doesn't seem like a good idea. You can use scheduler > > cgroups to do this sort of bandwidth limitation instead. > > Not quite - verynice is simple, easy to use (verynice 'command' or > verynice -p PID), and aims to be somewhat multiplatform (in particular, > runs on kFreeBSD). > > We are using it mostly to limit processes that start to hog NFS server > at inconvenient times thus slowing down other nfs clients (that was the > primary reason I wrote it) and that do not need to finish soon. So you are using CPU throttling to solve I/O contention? This is an even worse idea. > It turned out to be incredibly useful. That's nice, but please don't present this as more than a site-specific hack. Ben. -- Ben Hutchings Q. Which is the greater problem in the world today, ignorance or apathy? A. I don't know and I couldn't care less.
Attachment:
signature.asc
Description: This is a digitally signed message part