slow machine, swap in use, but more than 5GB of RAM inactive
Erich Dollansky
erichsfreebsdlist at alogt.com
Wed Mar 8 07:10:24 UTC 2017
Hi,
On Tue, 7 Mar 2017 16:46:21 -0800
Kevin Oberman <rkoberman at gmail.com> wrote:
> On Tue, Mar 7, 2017 at 3:14 PM, Erich Dollansky
> <erichsfreebsdlist at alogt.com
> > wrote:
>
> > On Tue, 7 Mar 2017 23:30:58 +1100 (EST)
> > Ian Smith <smithi at nimnet.asn.au> wrote:
> >
> > > On Tue, 7 Mar 2017 10:19:35 +0800, Erich Dollansky wrote:
> > > >
> > > > I wonder about the slow speed of my machine while top shows
> > > > ample inactive memory:
> > >
> > > ( quoting from this top output because it's neater :)
> > >
> > > > last pid: 85287; load averages: 2.56, 2.44, 1.68
> > > > up 6+10:24:45 10:13:36 191 processes: 5 running, 186 sleeping
> > > > CPU 0: 47.1% user, 0.0% nice, 51.4% system, 0.0% interrupt,
> > > > 1.6% idle CPU 1: 38.4% user, 0.0% nice, 60.4% system, 0.0%
> > > > interrupt, 1.2% idle CPU 2: 38.8% user, 0.0% nice, 59.2%
> > > > system, 0.0% interrupt, 2.0% idle CPU 3: 45.5% user, 0.0%
> > > > nice, 51.0% system, 0.4% interrupt, 3.1% idle Mem: 677M
> > > > Active, 5600M Inact, 1083M Wired, 178M Cache, 816M Buf,301M
> > > > Free Swap: 16G Total, 1352M Used, 15G Free, 8% Inuse
> > >
> > > Others have covered the swap / inactive memory issue.
> > >
> > > But I'd expect this to be slow, for any new work anyway .. there's
> > > next to no idle on any CPU. I'd be asking, what's all of that
> > > system usage?
> > >
> > this is building ports in the background. Still, used doing this
> > ones a month, I know the feeling when the ports are updated. This
> > one was really slow. Hopefully, it was just an unlucky coincidence.
> >
> > I rebooted meanwhile the machine. It is faster now, I would say, it
> > is back to normal now. It did not come to its limits since the new
> > start. It is now on:
> >
> > FreeBSD 10.3-STABLE #3 r314363
> >
> > Erich
> >
>
> Well, looks like over half of the CPU is running in system space and
> that seems rather high for what I would assume is compilation. I
> thinnk you will need to poke around with things like systat, and the
> like to see just what the system is doing for 55% or so of all CPUs.
> Since there doe snot seem to be a lot of IO or memory at issue, the
> various command for those are probably not very interesting. Probably
> not lock stats, either.
>
> This reminds me of when some operation (IIRC NFS related) was calling
> system time routines that are fairly expensive on FreeBSD almost
> continually.
> --
There were one or two NFS clients connected but should have been idle.
Both are on an older FreeBSD 12.
Could it be caused by a Seagate 2TB 2 1/2" HD with
8GB SSD (ST2000LX001-1RG174)?
The system behaved strangely when the disk was new. It seemed that the
flash part was used at the beginning also for writing. It took then
some times up to a minute before data could have been read again after
prolonged write operations. The disk has now a transfer volume of some
5TB and did not show this behaviour for a few weeks. It is only hard for
me to see why this happens. I use this machine since years. The only
change was the disk. I know, things can happen.
As I have rebooted the machine, I will keep this in mind and check also
this direction when this happens again.
Thanks!
Erich
More information about the freebsd-stable
mailing list