Date: Fri, 7 Oct 2016 16:52:36 -0700 From: Mark Johnston <markj@freebsd.org> To: Bruce Evans <brde@optusnet.com.au> Cc: src-committers@freebsd.org, svn-src-user@freebsd.org Subject: Re: svn commit: r306831 - in user/alc/PQ_LAUNDRY/usr.bin: systat vmstat Message-ID: <20161007235236.GC65501@wkstn-mjohnston.west.isilon.com> In-Reply-To: <20161008093219.O2989@besplex.bde.org> References: <201610072217.u97MHh08042582@repo.freebsd.org> <20161008093219.O2989@besplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Oct 08, 2016 at 10:35:59AM +1100, Bruce Evans wrote: > On Fri, 7 Oct 2016, Mark Johnston wrote: > > > Log: > > Teach vmstat and systat about v_pdshortfalls. > > systat has a negative amount of space available for expansion. Hm, I see. I'll revert the systat modifications then. It isn't particularly handy to have the rate of shortfall wakeups given what's already available in the vmstat display. > > > Modified: user/alc/PQ_LAUNDRY/usr.bin/systat/vmstat.c > > ============================================================================== > > --- user/alc/PQ_LAUNDRY/usr.bin/systat/vmstat.c Fri Oct 7 22:17:25 2016 (r306830) > > +++ user/alc/PQ_LAUNDRY/usr.bin/systat/vmstat.c Fri Oct 7 22:17:43 2016 (r306831) > > @@ -339,14 +340,15 @@ labelkre(void) > > mvprintw(VMSTATROW + 8, VMSTATCOL + 9, "react"); > > mvprintw(VMSTATROW + 9, VMSTATCOL + 9, "pdwak"); > > mvprintw(VMSTATROW + 10, VMSTATCOL + 9, "pdpgs"); > > - mvprintw(VMSTATROW + 11, VMSTATCOL + 9, "intrn"); > > - mvprintw(VMSTATROW + 12, VMSTATCOL + 9, "wire"); > > - mvprintw(VMSTATROW + 13, VMSTATCOL + 9, "act"); > > - mvprintw(VMSTATROW + 14, VMSTATCOL + 9, "inact"); > > - mvprintw(VMSTATROW + 15, VMSTATCOL + 9, "laund"); > > - mvprintw(VMSTATROW + 16, VMSTATCOL + 9, "free"); > > - if (LINES - 1 > VMSTATROW + 17) > > - mvprintw(VMSTATROW + 17, VMSTATCOL + 9, "buf"); > > + mvprintw(VMSTATROW + 11, VMSTATCOL + 9, "pdshort"); > > + mvprintw(VMSTATROW + 12, VMSTATCOL + 9, "intrn"); > > + mvprintw(VMSTATROW + 13, VMSTATCOL + 9, "wire"); > > + mvprintw(VMSTATROW + 14, VMSTATCOL + 9, "act"); > > + mvprintw(VMSTATROW + 15, VMSTATCOL + 9, "inact"); > > + mvprintw(VMSTATROW + 16, VMSTATCOL + 9, "laund"); > > + mvprintw(VMSTATROW + 17, VMSTATCOL + 9, "free"); > > + if (LINES - 1 > VMSTATROW + 18) > > + mvprintw(VMSTATROW + 18, VMSTATCOL + 9, "buf"); > > > > mvprintw(GENSTATROW, GENSTATCOL, " Csw Trp Sys Int Sof Flt"); > > Already the display of 'buf' was broken by expansion: on freefall now: > > 30 users Load 0.21 0.18 0.13 Oct 7 22:32 > Mem usage: 82%Phy 10%Kmem > Mem: KB REAL VIRTUAL VN PAGER SWAP PAGER > Tot Share Tot Share Free in out in out > Act 784512 23764 7319516 42640 4515744 count > All 893752 129672 7995532 624636 pages > Proc: Interrupts > r p d s w Csw Trp Sys Int Sof Flt ioflt 569 total > 422 3216 49 362 232 33 5 1 cow uart0 4 > 2 zfod 29 cpu0:timer > 0.3%Sys 0.0%Intr 0.0%User 0.0%Nice 99.7%Idle ozfod 4 igb0:que 0 > | | | | | | | | | | %ozfod 1 igb0:que 1 > daefr 12 igb0:que 2 > 1 dtbuf 93 prcfr 1 igb0:que 3 > Namei Name-cache Dir-cache 485980 desvn 117 totfr 10 igb0:que 4 > Calls hits % hits % 485400 numvn react 1 igb0:que 5 > 17 17 100 179227 frevn pdwak 1 igb0:que 6 > 147 pdpgs 4 igb0:que 7 > Disks da0 da1 da2 da3 da4 da5 da6 intrn igb0:link > KB/t 20.12 19.75 19.75 20.00 19.62 20.00 0.00 5096864 wire 196 ciss0 274 > tps 31 31 31 31 31 31 0 330628 act 27 cpu1:timer > MB/s 0.61 0.60 0.60 0.60 0.59 0.60 0.00 14524764 inact 6 cpu18:time > %busy 15 13 16 16 18 15 0 cache 7 cpu8:timer > 4515744 free 3 cpu10:time > 2 cpu16:time > > 'buf' is not important, but 'free' is. > > The last line is not available for the main display, but is used for the > cpu16:time<content lost to verbose naming>. This line is used for input, > so and everything displayed on it is cleared for input and then messed up > in an uglier way by redisplaying only parts (for the cpu16: field, the > count is re-displayed on screen refresh but the description is not > re-displayed without a switch to another screen and back). > > The interrupts column and the input line are generally broken. With > 24 CPUs and also some other per-cpu interrupt counts, it takes about > a 200-row terminal to display all the interrupts and then 1 more row > to stop the input line running into the interrupts. The rest of the > display is tuned for 25 rows and looks strange when it is 7/8 unused. > Per-CPU interrupts are also miscounted in totals... > > Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20161007235236.GC65501>