Date: Thu, 29 Oct 2015 12:47:03 +0100 From: Miroslav Lachman <000.fbsd@quip.cz> To: Jeremy Chadwick <jdc@koitsu.org> Cc: freebsd-stable@freebsd.org Subject: Re: stable/10: high load average when box is idle Message-ID: <56320737.2060901@quip.cz> In-Reply-To: <20151029100910.GA52255@icarus.home.lan> References: <20151027050508.GA7612@icarus.home.lan> <5631EE40.3020404@quip.cz> <20151029100910.GA52255@icarus.home.lan>
next in thread | previous in thread | raw e-mail | index | archive | help
Jeremy Chadwick wrote on 10/29/2015 11:09: > On Thu, Oct 29, 2015 at 11:00:32AM +0100, Miroslav Lachman wrote: >> Jeremy Chadwick wrote on 10/27/2015 06:05: >>> (I am not subscribed to the mailing list, please keep me CC'd) >>> >>> Issue: a stable/10 system that has an abnormally high load average (e.g. >>> 0.15, but may be higher depending on other variables which I can't >>> account for) when the machine is definitely idle (i.e. cannot be traced >>> to high interrupt usage per vmstat -i, cannot be traced to a userland >>> process or kernel thread, etc.). >>> >>> This problem has been discussed many times on the FreeBSD mailing lists >>> and the FreeBSD forum (including some folks seeing it on 9.x, but my >>> complaint here is focused on 10.x so please focus there). >>> >>> I'd politely like to request that anyone experiencing this, or who has >>> experienced it (and if you know when it stopped or why, including what >>> you may have done, include that), to chime in on this ticket from 2012 >>> (made for 9.x but style of issue still applies; c#5 is quite valid): >>> >>> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=173541 >>> >>> For those still experiencing it, I'd suggest reading c#8 and seeing if >>> sysctl kern.eventtimer.periodic=1 relieves the problem for you. (At >>> this time I would not suggest leaving that set indefinitely, as it does >>> seem to increase the interrupt rate in cpuX:timer in vmstat -i. But for >>> me kern.eventtimer.periodic=1 "fixes" the issue) >> >> Is it on real HW server or in some kind of virtualization? I am seeing load >> 0.5 - 1.2 on three virtual machines in VMware. The machines are without any >> traffic. Just fresh instalation of FreeBSD 10.1 and some services without >> any public content. > > I've seen it on both bare-metal and VMs. Please see c#8 in the ticket; > there's an itemised list of where I've seen it, but I'm sure it's not > limited to just those. OK, I have read your c#8 and did some tests on our affected VMs. With sysctl kern.eventtimer.periodic=1 it is better. Where previously load were about 0.40 is 0.15 now. One of these three systems is FreeBSD 10.2 and on this machine the positive effect of kern.eventtimer.periodic=1 is more visible - load is now 0.00 - 0.05. I don't know if this is some coincidence or something is different in 10.2. Settings of kern.eventtimer is the same on all VMs kern.eventtimer.et.LAPIC.flags: 7 kern.eventtimer.et.LAPIC.frequency: 35071418 kern.eventtimer.et.LAPIC.quality: 600 kern.eventtimer.et.i8254.flags: 1 kern.eventtimer.et.i8254.frequency: 1193182 kern.eventtimer.et.i8254.quality: 100 kern.eventtimer.et.RTC.flags: 17 kern.eventtimer.et.RTC.frequency: 32768 kern.eventtimer.et.RTC.quality: 0 kern.eventtimer.choice: LAPIC(600) i8254(100) RTC(0) kern.eventtimer.singlemul: 4 kern.eventtimer.idletick: 0 kern.eventtimer.timer: LAPIC kern.eventtimer.periodic: 1 Miroslav Lachman
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?56320737.2060901>