Virtually impossible to keep load under 0.5 on idle desktop with F14+kms
dennisml at conversis.de
Tue Nov 9 01:47:00 UTC 2010
On 11/08/2010 11:46 PM, Adam Williamson wrote:
> On Mon, 2010-11-08 at 00:08 -0600, Jeff Raber wrote:
>> Seem similar to the issue described here:
>> I'm not sure if anything ever came of that discussion.
> Not much, but I did ask kernel team if there was any reason to consider
> it a release blocker, and they didn't seem particularly bothered.
It probably shouldn't be a release blocker because it doesn't prevent the
system from functioning correctly but it's certainly a problem and it will
probably debugging other issues harder because nobody can tell if elevated
load is the result of the reported problem or this bug.
FWIW "nohz=off" seems to fix the high load problem for me although I still
see the load balancer accounting for most of the wakup calls in powertop
(which might or might no be an actual problem).
More information about the test