[pfSense-discussion] pfSense / Free BSD CPU kern.cp_time Jams in some environments
Hi Is anyone else getting this? It is occurring if you get a either a 1) divide by zero error on the index page for CPU Usage or 2) an indication the CPU is always on 0% use, which it shouldn't be for long! It seems to occur 1.2.2 onwards and on some motherboards and not others. It is particularly relevant because if it is occurring it slows down the whole system, for example I have a Pentium 266 that it does not occur on and a Pentium 400 that it does occur on. When it occurs the Pentium 400 slows down to perform more like about a 486 might in comparative terms for serving up the web pages. I assume it similarly effects traffic. There is also a report of a Pentium III 650 doing the same at http://cvstrac.pfsense.org/tktview?tn=1884 The only (temporary) fix I know is reboot. It can also be demonstrated by executing the command sysctl -n kern.cp_time and getting the same values back. On the susceptible system this happens within about an hour of running. Given it takes an hour typically for the condition to develop may make tracking it down more difficult. It would be good to know how many others are affected on what systems it occurs. Kind regards David - To unsubscribe, e-mail: discussion-unsubscr...@pfsense.com For additional commands, e-mail: discussion-h...@pfsense.com Commercial support available - https://portal.pfsense.org
Re: [pfSense-discussion] pfSense / Free BSD CPU kern.cp_time Jams in some environments
On Sat, Apr 4, 2009 at 4:50 PM, Tortise tort...@paradise.net.nz wrote: Hi Is anyone else getting this? It is occurring if you get a either a 1) divide by zero error on the index page for CPU Usage or 2) an indication the CPU is always on 0% use, which it shouldn't be for long! It seems to occur 1.2.2 onwards and on some motherboards and not others. Should be 1.2.1 onwards, there are no FreeBSD differences from 1.2.1 to 1.2.2. 1.2.3 also exhibits the same behavior on these 440BX systems, though our calculation has changed so you can never get a divide by 0, it just returns 0% when these counters are wrong. I checked a wide range of hardware and I don't have anything that exhibits this, but I don't have any 440BX systems either, which seems to be what this is limited to, and not all of them at that or we would have heard about it quite some time ago I'm sure. - To unsubscribe, e-mail: discussion-unsubscr...@pfsense.com For additional commands, e-mail: discussion-h...@pfsense.com Commercial support available - https://portal.pfsense.org