on Fri, Nov 16, 2001 at 02:05:18PM +0200, Tuomas Pellonpera (tp58494@uta.fi) wrote: > On 15 Nov 2001, John Hasler wrote: > > > Dave Sherohman <esper@sherohman.org> wrote on 15/11/2001 (16:53) : > > > At one time, leaving a machine on overnight consumed less power than > > > rebooting it (i.e., shutting down and powering back up wasted more > > > resources than just leaving it on). Is that still the case? > > > > It never was. > > Does (daily) shutting down and rebooting wear out the hardware more than > leaving the computer on for days/weeks/months does? I've seen one reasonably detailed study of this, possibly by IBM, and probably carried at Slashdot, in the past couple-three years. The consensus seemed to be that there was little real distinction between leaving a system running, or shutting it down, with the exception of the monitor (assuming a CRT w/o powersaving features). Thermal cycling does impact computer hardware, but the degredation over the effective life of most hardware (say, 3-4 years) is not significant. Power consumption is a slight factor, though again, excepting a CRT monitor, it's relatively low. My own balance point: I leave my systems up, and typically get uptimes of 2-3 months, 94 days is the top recorded (I think I've been higher, but haven't broken 100). The benefits of always-on beat the minimal power savings and mornign boot wait. For laptops, sleep and suspend states are a nice proxy. OTOH, I also believe that, particularly for systems that are being dynamically modified, periodic reboots to confirm that system startup proceeds as expected, is a good idea. So, say, quarterly boots aren't all bad. Continuous uptimes of over a year likely indicate a system which could stand for a kernel update. For those who're looking for on-the-fly kernel updates, the two-kernel Monty hack allows booting a new kernel from an existing one. Uptime isn't preserved by default, though it might be possible to populate this data by passing appropriate kernel parameters. While useful, TKM also introduces certain security risks, recognizing that by changing kernel properties, virtually any aspect of system function may be modified. Peace. -- Karsten M. Self <kmself@ix.netcom.com> http://kmself.home.netcom.com/ What part of "Gestalt" don't you understand? Home of the brave http://gestalt-system.sourceforge.net/ Land of the free Free Dmitry! Boycott Adobe! Repeal the DMCA! http://www.freesklyarov.org Geek for Hire http://kmself.home.netcom.com/resume.html
Attachment:
pgp66oqv_Jaf1.pgp
Description: PGP signature