Date: Fri, 20 Jan 2023 17:08:50 -0800 From: Stavros Filargyropoulos <stav@stav.cc> To: Paul Procacci <pprocacci@gmail.com> Cc: freebsd-questions@freebsd.org Subject: Re: Load 0.20 on a freshly installed idle system Message-ID: <CAGAKwcid8j3C28wiypO87X9HT9Wq8cfTNHWnOQvcQgAcWrFuLw@mail.gmail.com> In-Reply-To: <CAFbbPujMDQgo=zvSc5orcT=v3ftrVPXDjZE-D%2BUxt3%2BdOjFmUA@mail.gmail.com> References: <CAGAKwcjmzz%2B6NMaK2YpjWZPB282RvrKV6zQ%2BSLHsDj3zqXwvGw@mail.gmail.com> <CAFbbPuhX%2BrJFLn8ALVNbCujf1tG3DufVWGSknPkJ%2BKdovFq1xQ@mail.gmail.com> <CAGAKwcgFWukBrB_6SZVAoC4YKyPVhVMGED4aB6Z4E1fwZAO6mg@mail.gmail.com> <CAFbbPuhyJDewvE_a6MQpiGrpUew7EtxZKwt_=RHODNUyCYNduw@mail.gmail.com> <CAGAKwcghHH5t1_dtWPkOCwxPvDra%2BmgNFTQsG65i1_7pzTDkug@mail.gmail.com> <CAFbbPujMDQgo=zvSc5orcT=v3ftrVPXDjZE-D%2BUxt3%2BdOjFmUA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
[-- Attachment #1 --] Well, actually you were not far off. The 5 seconds is the interval we look into the run queue: https://svnweb.freebsd.org/base/head/sys/kern/kern_synch.c?view=markup#l87. The average is still for 1, 5, 15 min. So a load 0.20 means that out of the 12 times we looked in the run queue the past minute, at least 2.5 there was a job. Which I think is consistent with what everyone understands `load` is. And again is not what I see in `top`. Unless we count the `idle` process? Which we shouldn't... On Fri, Jan 20, 2023 at 4:56 PM Paul Procacci <pprocacci@gmail.com> wrote: > I'll concede it probably isn't a 5 second window as what I've provided is > for OpenBSD, but it's still a measurement of processes in the run queue - > that's for sure. > I'm being lazy and don't want to look at the source code. That's > something you could do. > > Thanks, > Paul Procacci > > On Fri, Jan 20, 2023 at 7:28 PM Stavros Filargyropoulos <stav@stav.cc> > wrote: > >> Thanks Paul, but If that was the case, then running 5 `top -SH` with a >> refresh of 1 second, should increase the load to x4. Which is not what I >> see. >> >> This answer seems a bit vague. Is that the best explanation we have? Do >> you mean to say that there is not a single freebsd box out there that has a >> load of less than 0.20? >> >> I'm pretty sure that last time I run FreeBSD that wasn't the case. Also >> haven't seen this load in idle system in any other *BSDs. >> >> Thanks for taking a look anyway. >> >> On Fri, Jan 20, 2023 at 4:19 PM Paul Procacci <pprocacci@gmail.com> >> wrote: >> >>> Stavros, >>> >>> Load averages by themselves mean very little. >>> It's not a figure of cpu usage if that's what you're worried about. >>> >>> Load averages on BSD are simply processes which have wanted to run at >>> least once in the most recent 5-second window, with a degradation over >>> time.[1] >>> >>> Unless you're having a specific problem, I'd guess to say there is no >>> problem here at all. >>> >>> Thanks, >>> Paul Procacci >>> >>> [1] - http://www.undeadly.org/cgi?action=article&sid=20090715034920 >>> >>> >>> On Fri, Jan 20, 2023 at 6:59 PM Stavros Filargyropoulos <stav@stav.cc> >>> wrote: >>> >>>> Thanks Paul. >>>> >>>> Yeah, I used the `-SH` flags. Don't see anything suspicious. >>>> >>>> This is is the output of the `systat`: Not sure I understand much from >>>> it >>>> >>>> 2 users Load 0.20 0.15 0.06 Jan 20 15:54:53 >>>> Mem usage: 1%Phy 0%Kmem VN PAGER SWAP >>>> PAGER >>>> Mem: REAL VIRTUAL in out in >>>> out >>>> Tot Share Tot Share Free count >>>> Act 33476K 10672K 513G 12612K 31367M pages >>>> All 35920K 12928K 513G 47112K ioflt >>>> Interrupts >>>> Proc: cow 34 >>>> total >>>> r p d s w Csw Trp Sys Int Sof Flt zfod >>>> uart0 4 >>>> 25 118 1 21 5 13 1 ozfod 10 >>>> cpu0:timer >>>> %ozfod 9 >>>> cpu1:timer >>>> 0.0%Sys 0.0%Intr 0.0%User 0.0%Nice 100%Idle daefr 6 >>>> cpu2:timer >>>> | | | | | | | | | | | prcfr 4 >>>> cpu3:timer >>>> totfr 4 >>>> xhci0 128 >>>> 229 dtbuf react >>>> ahci0 129 >>>> Namei Name-cache Dir-cache 619780 maxvn pdwak >>>> nvme0:admi >>>> Calls hits % hits % 805 numvn 5 pdpgs >>>> nvme0:io0 >>>> 7 7 100 52 frevn intrn >>>> nvme0:io1 >>>> 355M wire >>>> nvme0:io2 >>>> Disks nvd0 ada0 ada1 ada2 ada3 ada4 da0 17M act >>>> nvme0:io3 >>>> KB/t 16.00 0.00 0.00 0.00 0.00 0.00 0.00 2880K inact >>>> igb0:rxq0 >>>> tps 0 0 0 0 0 0 0 0 laund >>>> igb0:rxq1 >>>> MB/s 0.01 0.00 0.00 0.00 0.00 0.00 0.00 31G free 1 >>>> igb0:rxq2 >>>> %busy 0 0 0 0 0 0 0 142M buf >>>> igb0:rxq3 >>>> >>>> igb0:aq >>>> >>>> And the output of `top -SH` for reference: >>>> >>>> last pid: 1508; load averages: 0.25, 0.22, 0.10 >>>> >>>> up >>>> 0+00:08:13 15:57:54 >>>> 131 threads: 5 running, 106 sleeping, 20 waiting >>>> CPU: 0.1% user, 0.0% nice, 0.0% system, 0.0% interrupt, 99.9% idle >>>> Mem: 16M Active, 3032K Inact, 355M Wired, 142M Buf, 31G Free >>>> Swap: 3962M Total, 3962M Free >>>> >>>> PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU >>>> COMMAND >>>> 11 root 155 ki31 0B 64K RUN 2 8:12 99.99% >>>> idle{idle: cpu2} >>>> 11 root 155 ki31 0B 64K CPU3 3 8:11 99.98% >>>> idle{idle: cpu3} >>>> 11 root 155 ki31 0B 64K CPU0 0 8:12 99.97% >>>> idle{idle: cpu0} >>>> 11 root 155 ki31 0B 64K CPU1 1 8:11 99.96% >>>> idle{idle: cpu1} >>>> 0 root -76 - 0B 640K - 0 0:00 0.04% >>>> kernel{if_config_tqg_0} >>>> 1508 stavros 20 0 14M 3812K CPU2 2 0:00 0.02% top >>>> 1493 stavros 20 0 15M 5080K select 1 0:00 0.02% tmux >>>> 12 root -60 - 0B 320K WAIT 3 0:01 0.01% >>>> intr{swi4: clock (0)} >>>> 5 root -16 - 0B 16K - 0 0:00 0.01% >>>> rand_harvestq >>>> 6 root -16 - 0B 48K psleep 0 0:00 0.00% >>>> pagedaemon{dom0} >>>> 1440 stavros 20 0 21M 9440K select 3 0:00 0.00% sshd >>>> 1404 root 20 0 18M 6988K select 2 0:00 0.00% >>>> sendmail >>>> 15 root -72 - 0B 96K - 3 0:00 0.00% >>>> usb{usbus0} >>>> 12 root -60 - 0B 320K WAIT 2 0:00 0.00% >>>> intr{swi4: clock (2)} >>>> 12 root -88 - 0B 320K WAIT 2 0:00 0.00% >>>> intr{irq128: xhci0} >>>> 8 root 20 - 0B 144K sdflus 3 0:00 0.00% >>>> bufdaemon{/ worker} >>>> 12 root -60 - 0B 320K WAIT 1 0:00 0.00% >>>> intr{swi4: clock (1)} >>>> 0 root -76 - 0B 640K - 2 0:00 0.00% >>>> kernel{if_io_tqg_2} >>>> 8 root -16 - 0B 144K - 1 0:00 0.00% >>>> bufdaemon{bufspacedaemon-0} >>>> 9 root 16 - 0B 16K syncer 3 0:00 0.00% >>>> syncer >>>> 12 root -60 - 0B 320K WAIT 3 0:00 0.00% >>>> intr{swi4: clock (3)} >>>> 8 root -16 - 0B 144K psleep 2 0:00 0.00% >>>> bufdaemon{bufdaemon} >>>> 8 root -16 - 0B 144K - 0 0:00 0.00% >>>> bufdaemon{bufspacedaemon-1} >>>> >>>> Thanks, >>>> Stavros >>>> >>>> >>>> On Fri, Jan 20, 2023 at 2:03 PM Paul Procacci <pprocacci@gmail.com> >>>> wrote: >>>> >>>>> What flags have you included with top? >>>>> >>>>> By default `top` doesn't show any kernel related processes. `top >>>>> -SH` is useful in this respect and others flags may be useful as well. >>>>> >>>>> systat -vmstat >>>>> >>>>> ^^ Another command that I like. Also may give you insight. >>>>> >>>>> Thanks, >>>>> Paul Procacci >>>>> >>>>> On Fri, Jan 20, 2023 at 4:33 PM Stavros Filargyropoulos <stav@stav.cc> >>>>> wrote: >>>>> >>>>>> Hi, >>>>>> >>>>>> I just installed 13.1 on a system, and the load is at 0.20 at all >>>>>> times, although nothing is running and the CPU is 100% free. >>>>>> >>>>>> `ps aux` and `top` don't show anything of interest either. >>>>>> >>>>>> For what's worth it, the motherboard is a Supermicro X12STL-IF and >>>>>> the CPU a Xeon 2314. >>>>>> >>>>>> Any ideas/commands that will help me troubleshoot this further? >>>>>> >>>>>> Thanks, >>>>>> Stavros >>>>>> >>>>> >>>>> >>>>> -- >>>>> __________________ >>>>> >>>>> :(){ :|:& };: >>>>> >>>> >>> >>> -- >>> __________________ >>> >>> :(){ :|:& };: >>> >> > > -- > __________________ > > :(){ :|:& };: > [-- Attachment #2 --] <div dir="ltr"><div>Well, actually you were not far off.</div><div><br></div><div>The 5 seconds is the interval we look into the run queue:</div><div><br></div><div><a href="https://svnweb.freebsd.org/base/head/sys/kern/kern_synch.c?view=markup#l87">https://svnweb.freebsd.org/base/head/sys/kern/kern_synch.c?view=markup#l87</a>.</div><div><br></div><div>The average is still for 1, 5, 15 min. So a load 0.20 means that out of the 12 times we looked in the run queue the past minute, at least 2.5 there was a job. Which I think is consistent with what everyone understands `load` is. And again is not what I see in `top`.</div><div><br></div><div>Unless we count the `idle` process? Which we shouldn't...<br></div><div><br></div><div><br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Jan 20, 2023 at 4:56 PM Paul Procacci <<a href="mailto:pprocacci@gmail.com">pprocacci@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div>I'll concede it probably isn't a 5 second window as what I've provided is for OpenBSD, but it's still a measurement of processes in the run queue - that's for sure.<br></div><div>I'm being lazy and don't want to look at the source code. That's something you could do.</div><div><br></div><div>Thanks,</div><div>Paul Procacci<br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Jan 20, 2023 at 7:28 PM Stavros Filargyropoulos <<a href="mailto:stav@stav.cc" target="_blank">stav@stav.cc</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div></div><div>Thanks Paul, but If that was the case, then running 5 `top -SH` with a refresh of 1 second, should increase the load to x4. Which is not what I see.</div><div><br></div><div>This answer seems a bit vague. Is that the best explanation we have? Do you mean to say that there is not a single freebsd box out there that has a load of less than 0.20?<br></div><div><br></div><div>I'm pretty sure that last time I run FreeBSD that wasn't the case. Also haven't seen this load in idle system in any other *BSDs.<br></div><div><br></div><div>Thanks for taking a look anyway.<br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Jan 20, 2023 at 4:19 PM Paul Procacci <<a href="mailto:pprocacci@gmail.com" target="_blank">pprocacci@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div><div>Stavros,</div><div><br></div>Load averages by themselves mean very little.<br></div><div>It's not a figure of cpu usage if that's what you're worried about.</div><div><br></div><div>Load averages on BSD are simply processes which have wanted to run at least once in the most recent 5-second window, with a degradation over time.[1]<br><br></div><div>Unless you're having a specific problem, I'd guess to say there is no problem here at all.</div><div><br></div><div>Thanks,</div><div>Paul Procacci</div><div><br></div><div>[1] - <a href="http://www.undeadly.org/cgi?action=article&sid=20090715034920" target="_blank">http://www.undeadly.org/cgi?action=article&sid=20090715034920</a></div><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Jan 20, 2023 at 6:59 PM Stavros Filargyropoulos <<a href="mailto:stav@stav.cc" target="_blank">stav@stav.cc</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr">Thanks Paul.<br><br>Yeah, I used the `-SH` flags. Don't see anything suspicious.<br><br>This is is the output of the `systat`: Not sure I understand much from it<br><br> 2 users Load 0.20 0.15 0.06 Jan 20 15:54:53<br> Mem usage: 1%Phy 0%Kmem VN PAGER SWAP PAGER<br>Mem: REAL VIRTUAL in out in out<br> Tot Share Tot Share Free count<br>Act 33476K 10672K 513G 12612K 31367M pages<br>All 35920K 12928K 513G 47112K ioflt Interrupts<br>Proc: cow 34 total<br> r p d s w Csw Trp Sys Int Sof Flt zfod uart0 4<br> 25 118 1 21 5 13 1 ozfod 10 cpu0:timer<br> %ozfod 9 cpu1:timer<br> 0.0%Sys 0.0%Intr 0.0%User 0.0%Nice 100%Idle daefr 6 cpu2:timer<br>| | | | | | | | | | | prcfr 4 cpu3:timer<br> totfr 4 xhci0 128<br> 229 dtbuf react ahci0 129<br>Namei Name-cache Dir-cache 619780 maxvn pdwak nvme0:admi<br> Calls hits % hits % 805 numvn 5 pdpgs nvme0:io0<br> 7 7 100 52 frevn intrn nvme0:io1<br> 355M wire nvme0:io2<br>Disks nvd0 ada0 ada1 ada2 ada3 ada4 da0 17M act nvme0:io3<br>KB/t 16.00 0.00 0.00 0.00 0.00 0.00 0.00 2880K inact igb0:rxq0<br>tps 0 0 0 0 0 0 0 0 laund igb0:rxq1<br>MB/s 0.01 0.00 0.00 0.00 0.00 0.00 0.00 31G free 1 igb0:rxq2<br>%busy 0 0 0 0 0 0 0 142M buf igb0:rxq3<br><div> igb0:aq</div><div><br></div><div>And the output of `top -SH` for reference:<br></div><div><br></div><div>last pid: 1508; load averages: 0.25, 0.22, 0.10 up 0+00:08:13 15:57:54<br>131 threads: 5 running, 106 sleeping, 20 waiting<br>CPU: 0.1% user, 0.0% nice, 0.0% system, 0.0% interrupt, 99.9% idle<br>Mem: 16M Active, 3032K Inact, 355M Wired, 142M Buf, 31G Free<br>Swap: 3962M Total, 3962M Free<br><br> PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU COMMAND<br> 11 root 155 ki31 0B 64K RUN 2 8:12 99.99% idle{idle: cpu2}<br> 11 root 155 ki31 0B 64K CPU3 3 8:11 99.98% idle{idle: cpu3}<br> 11 root 155 ki31 0B 64K CPU0 0 8:12 99.97% idle{idle: cpu0}<br> 11 root 155 ki31 0B 64K CPU1 1 8:11 99.96% idle{idle: cpu1}<br> 0 root -76 - 0B 640K - 0 0:00 0.04% kernel{if_config_tqg_0}<br> 1508 stavros 20 0 14M 3812K CPU2 2 0:00 0.02% top<br> 1493 stavros 20 0 15M 5080K select 1 0:00 0.02% tmux<br> 12 root -60 - 0B 320K WAIT 3 0:01 0.01% intr{swi4: clock (0)}<br> 5 root -16 - 0B 16K - 0 0:00 0.01% rand_harvestq<br> 6 root -16 - 0B 48K psleep 0 0:00 0.00% pagedaemon{dom0}<br> 1440 stavros 20 0 21M 9440K select 3 0:00 0.00% sshd<br> 1404 root 20 0 18M 6988K select 2 0:00 0.00% sendmail<br> 15 root -72 - 0B 96K - 3 0:00 0.00% usb{usbus0}<br> 12 root -60 - 0B 320K WAIT 2 0:00 0.00% intr{swi4: clock (2)}<br> 12 root -88 - 0B 320K WAIT 2 0:00 0.00% intr{irq128: xhci0}<br> 8 root 20 - 0B 144K sdflus 3 0:00 0.00% bufdaemon{/ worker}<br> 12 root -60 - 0B 320K WAIT 1 0:00 0.00% intr{swi4: clock (1)}<br> 0 root -76 - 0B 640K - 2 0:00 0.00% kernel{if_io_tqg_2}<br> 8 root -16 - 0B 144K - 1 0:00 0.00% bufdaemon{bufspacedaemon-0}<br> 9 root 16 - 0B 16K syncer 3 0:00 0.00% syncer<br> 12 root -60 - 0B 320K WAIT 3 0:00 0.00% intr{swi4: clock (3)}<br> 8 root -16 - 0B 144K psleep 2 0:00 0.00% bufdaemon{bufdaemon}<br> 8 root -16 - 0B 144K - 0 0:00 0.00% bufdaemon{bufspacedaemon-1}</div><div><br></div><div>Thanks,</div><div>Stavros<br></div></div><br><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Jan 20, 2023 at 2:03 PM Paul Procacci <<a href="mailto:pprocacci@gmail.com" target="_blank">pprocacci@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div>What flags have you included with top?</div><div><br></div><div>By default `top` doesn't show any kernel related processes. `top -SH` is useful in this respect and others flags may be useful as well.</div><div><br></div><div>systat -vmstat <br><br></div><div>^^ Another command that I like. Also may give you insight.</div><div><br></div><div>Thanks,</div><div>Paul Procacci<br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Jan 20, 2023 at 4:33 PM Stavros Filargyropoulos <<a href="mailto:stav@stav.cc" target="_blank">stav@stav.cc</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div>Hi,</div><div><br></div><div>I just installed 13.1 on a system, and the load is at 0.20 at all times, although nothing is running and the CPU is 100% free.</div><div><br></div><div>`ps aux` and `top` don't show anything of interest either.</div><div><br></div><div>For what's worth it, the motherboard is a Supermicro X12STL-IF and the CPU a Xeon 2314.</div><div><br></div><div>Any ideas/commands that will help me troubleshoot this further?</div><div><br></div><div>Thanks,</div><div>Stavros<br></div></div> </blockquote></div><br clear="all"><br>-- <br><div dir="ltr">__________________<br><br>:(){ :|:& };:</div> </blockquote></div> </blockquote></div><br clear="all"><br>-- <br><div dir="ltr">__________________<br><br>:(){ :|:& };:</div> </blockquote></div> </blockquote></div><br clear="all"><br>-- <br><div dir="ltr">__________________<br><br>:(){ :|:& };:</div> </blockquote></div>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAGAKwcid8j3C28wiypO87X9HT9Wq8cfTNHWnOQvcQgAcWrFuLw>