FreeBSD 10 under XenServer 6.2(SP1) - Higher load average?
Jeroen van der Ham
jeroen at dckd.nl
Thu Jan 30 16:02:56 UTC 2014
Hi,
I’ve noticed this already back in June: http://lists.freebsd.org/pipermail/freebsd-xen/2013-June/001639.html
Nothing really came out of the discussion there, but it doesn’t really seem worrying or have much an impact on performance though.
Jeroen.
On 30 Jan 2014, at 15:10, Karl Pielorz <kpielorz_lst at tdx.co.uk> wrote:
>
> I've just installed a couple of FreeBSD 10-R instances on our Xen pool. The load averages on these machines seems to run higher for an idle box, than FreeBSD 9.x did
>
> e.g. 10.0-R (amd64 GENERIC):
>
> last pid: 4219; load averages: 0.31, 0.23, 0.12 up 0+00:07:45 14:04:08
> 15 processes: 1 running, 14 sleeping
> CPU: 0.0% user, 0.0% nice, 0.0% system, 0.0% interrupt, 100% idle
> Mem: 16M Active, 15M Inact, 44M Wired, 20M Buf, 1893M Free
> Swap: 2046M Total, 2046M Free
>
>
> A 9.2-STABLE (amd64 XENHVM) instance on the same XenServer:
>
> last pid: 76440; load averages: 0.00, 0.00, 0.00 up 2+15:07:27 14:05:10
> 22 processes: 1 running, 21 sleeping
> CPU: 0.0% user, 0.0% nice, 0.0% system, 0.0% interrupt, 100% idle
> Mem: 13M Active, 128M Inact, 91M Wired, 59M Buf, 237M Free
> Swap: 494M Total, 494M Free
>
>
> Both have xe-guest-utilities installed.
>
> The second box is actually technically busier than the first (as it's routing traffic between it's interfaces - admittedly, not much).
>
> But the load average on 10.0-R never settles to zero (like it did for 9.x)
>
> Just a bit confused as to if the user, nice, system and interrupt times are zero - how can the LA be >0?
>
> Anyone else noticed this? - I know an LA of 0.31 isn't the end of the world - but it's a bit of a jump on 0.00...
>
> -Karl
> _______________________________________________
> freebsd-xen at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-xen
> To unsubscribe, send any mail to "freebsd-xen-unsubscribe at freebsd.org"
More information about the freebsd-xen
mailing list