Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 24 Sep 1998 16:56:39 -0500
From:      Tom Jackson <toj@silverback.gorilla.net>
To:        freebsd-current@FreeBSD.ORG
Subject:   Re: Weired top display at smp current kernel from today
Message-ID:  <19980924165639.A408@TOJ.org>
In-Reply-To: <19980924015300.A807@TOJ.org>; from Tom Jackson on Thu, Sep 24, 1998 at 01:53:00AM -0500
References:  <711DCB8FB391D111B9DD00805F8BDDBAF7ED39@exips0019.itops.ldn.bzwint.com> <199809221644.AAA08290@spinner.netplex.com.au> <19980924015300.A807@TOJ.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Rebuiding world didNot work and manually starting the 2 rc5des clients
put both on one cpu. Using an aout init *did* fix the problem...  Ugh..

On Thu, Sep 24, 1998 at 01:53:00AM -0500, Tom Jackson wrote:
> Hi Peter,
> 
> I was really constipated about this on smp until I saw a posting from you
> and one by Mike Smith that told me others were having this problem.
> 
> As discussed by Mike, people using a script in rc.d to fire up 2 rc5des's
> see this regular. X is not involved in this case. When I boot, and it happens
> about every other day including today, both rc5des's are running at 50/50
> on one cpu and init is running 100 on the other. There is no login prompt
> so I have to login from another machine, comment out the rc5des starts and
> reboot. Then start the rc5des's manually. What's funny is this happens almost
> invariably every other day. Thinking about rebuilding world for today and see
> if it corrects itself.
> 
> btw, I'm running ELF, no softupdates, on a asus p2l97-ds w 128 MB ecc RAM.
> 
> I'll keep lurking and checking for yours or Mike's fix :)
> 
> On Wed, Sep 23, 1998 at 12:44:19AM +0800, Peter Wemm wrote:
> > Pierre.Dampure@barclayscapital.com wrote:
> > > 
> > > Frank,
> > > 
> > > Are you by any chance running an ELF version of X? I experienced pretty
> > > weird results over the past few days with and ELF X (built as per the port
> > > mods posted to the list): a make world running under X would basically bring
> > > the box to a halt, massive delays when typing characters on any of the other
> > > VTs, etc... this is on a box with 384Mb RAM, SMP w/PPro 200/512 and more
> > > than 256Mb free / no swap used, doing nothing else than the build.
> > 
> > Two things..
> > 
> > 1: There have been some clock tweaks in the last few days that might 
> > affect this.
> > 
> > 2: Try (humor me :-) booting up into single user, running a few commands
> > (eg: sync; sync; sync; fsck -p; ls) and then doing a ctrl-D to go into
> > multi-user. 
> > 
> > If this (#2) is the key to making your system work properly, then you are 
> > seeing the same problem I've been unable to put a finger on for ages.  
> > None of the other SMP folks saw it, but only I was running ELF at the time.
> > 
> > Somebody else (John Hay) mentioned that this fixes his problem, but it's 
> > not clear whether (or when) he's using elf.
> > 
> > The annoying thing was that it appeared to be the combination of ELF
> > executables and SMP that caused the problem.  Also, having init spawn 12
> > /usr/libexec/getty processes would kill the system too.
> > 
> > I had decided that it was something peculiar with my system that was 
> > causing this.  If other people are seeing it now, I'll be greatly relieved 
> > to know that I wasn't imagining it. :-)  (and also worried, because this 
> > is a show-stopper)
> > 
> > Cheers,
> > -Peter
> > 
> > > Best Regards,
> > > 
> > > 
> > > PY
> > > 
> > > -----Original Message-----
> > > From: Frank Nobis [mailto:fn@radio-do.de]
> > > Sent: Monday, September 21, 1998 2:01 PM
> > > To: freebsd-smp@FreeBSD.ORG
> > > Cc: freebsd-current@FreeBSD.ORG
> > > Subject: Weired top display at smp current kernel from today
> > > 
> > > 
> > > Running a mkae world -j4 I noticed only 50% idle time during the make.
> > > 
> > > Running just two programs whit a busy loop gives that output fron top
> > > 
> > > last pid:   314;  load averages:  1.84,  0.75,  0.30
> > > 14:55:17
> > > 45 processes:  3 running, 42 sleeping
> > > CPU states: 50.0% user,  0.0% nice,  0.0% system,  0.0% interrupt, 50.0%
> > > idle
> > > Mem: 7080K Active, 6332K Inact, 16M Wired, 100K Cache, 3457K Buf, 473M Free
> > > Swap: 199M Total, 199M Free
> > > 
> > >   PID USERNAME PRI NICE  SIZE    RES STATE  C   TIME   WCPU    CPU COMMAND
> > >   308 fn        90   0   748K   236K RUN    1   1:36 50.00% 49.97% pig
> > >   310 fn        91   0   748K   236K RUN    1   0:57 49.18% 49.02% pig
> > >   311 fn        28   0  1568K   908K CPU1   1   0:00  0.00%  0.00% top
> > > 
> > > Is this a bug in top? Can someone reproduce that?
> > > 
> > > pig is simply 'for(;;) i++;'
> > > 
> > > Regards
> > > 	Frank
> > > 
> > > -- 
> > >  Frank Nobis                            Email: PGP AVAILABLE
> > >  Landgrafenstr. 130                     dg3dcn   http://www.radio-do.de/~fn/
> > >  44139 Dortmund				Powered by FreeBSD
> > 
> > 
> > To Unsubscribe: send mail to majordomo@FreeBSD.org
> > with "unsubscribe freebsd-smp" in the body of the message
> > 
> 
> -- 
> Tom
> 
> --
> IMail Server for Windows NT. Evaluation version.
> Copyright (c) 1995-98 Ipswitch, Inc. http://www.ipswitch.com/
> 
> 
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-smp" in the body of the message
> 

-- 
Tom

--
IMail Server for Windows NT. Evaluation version.
Copyright (c) 1995-98 Ipswitch, Inc. http://www.ipswitch.com/


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980924165639.A408>