headers
Stephen M. Jones | 4 Aug 2005 00:03
Favicon

occassional hangs (non-smp) 2.0.2

I've been experiencing occassional hangs on an NFS server running 
NetBSD 2.0.2 (Jul 24th sys.tar.gz source and prior up to the 2.0.2 release).
The NFS clients also run the same kernel as the NFS server but do not
experience hangs.

The machine hangs under normal load and can break to the debugger on
the console.  

I can not switch to single user mode by killing init, nor can I continue.
I can provide further details upon request.

A trace and a portion of the process table follow:

cpu_Debugger() at netbsd:cpu_Debugger+0x4
comintr() at netbsd:comintr+0x768
alpha_shared_intr_dispatch() at netbsd:alpha_shared_intr_dispatch+0x6c
sio_iointr() at netbsd:sio_iointr+0x40
interrupt() at netbsd:interrupt+0x278
XentInt() at netbsd:XentInt+0x1c
--- interrupt (from ipl 0) ---
ltsleep() at netbsd:ltsleep+0xe0
uvm_scheduler() at netbsd:uvm_scheduler+0xd4
check_console() at netbsd:check_console

 PID           PPID     PGRP        UID S   FLAGS LWPS          COMMAND    WAIT
 11548         8769     3013          0 2  0x6000    1              cut
 14530          780       11         92 2  0x4000    1            sleep
 2440         15827    15827         99 2  0x4002    1             more   ttyin
 8769         24297     3013          0 2  0x4000    1              ksh
 28533            1    28533          0 2   0x400    3            named       *
(Continue reading)

Permalink | Reply | 1 comment |
headers
Hernani Marques Madeira | 4 Aug 2005 00:31

Re: occassional hangs (non-smp) 2.0.2

On Wed, Aug 03, 2005 at 05:03:45PM -0500, Stephen M. Jones wrote:

> I've been experiencing occassional hangs on an NFS server running 
> NetBSD 2.0.2 (Jul 24th sys.tar.gz source and prior up to the 2.0.2 release).
> The NFS clients also run the same kernel as the NFS server but do not
> experience hangs.
> 
> The machine hangs under normal load and can break to the debugger on
> the console.  

I've had similar problems (currently that box is running 1.6.2_RC4)
reported in PR #30148, but while the box was running as NFS client, although 
I'v never set up that box as NFS server:
http://www.netbsd.org/cgi-bin/query-pr-single.pl?number=30148

--

-- 
hernani
Permalink | Reply | 0 comments |
headers
sp4rc | 9 Aug 2005 02:36
Picon

raided1 root / installboot drives me mad

Hello list,

i am getting mad trying to installing netbsd 2.0.2 using a raided1 root
on as1200. So far I came to point 16.3.9
(http://www.netbsd.org/guide/en/chap-rf.html#chap-rf-ex-raid1root)

the quide says to install the boot loader on sda1 using the following
command:

# /usr/sbin/installboot -v /dev/rsd1a /usr/mdec/bootblk

as i am installing on an alpha system and not sparc64, bootblk doesn't
exist, so i tried /usr/mdec/boot.

# installboot -v /dev/rsd1a /usr/mdec/boot
File system:         /dev/rsd1a
File system type:    raw (blocksize 8192, needswap 0)
Primary bootstrap:   /usr/mdec/boot
Bootstrap start sector:  1
Bootstrap sector count:  103
New boot block checksum: 0xf4f4a3e360232b13
Writing bootstrap
Writing boot block
installboot: Writing `/dev/rsd1a': Read-only file system
installboot: Set bootstrap operation failed  <-- aaarggh

but the bootstrap process always fails. Why does installboot complains
about a read-only file system? 
Althoug i don't understand why to take a instead of c, i tried using c
instead of a, which seemed to work
(Continue reading)

Permalink | Reply | 3 comments |
headers
Simon Burge | 9 Aug 2005 03:56

Re: raided1 root / installboot drives me mad

sp4rc wrote:

> Hello list,
> 
> i am getting mad trying to installing netbsd 2.0.2 using a raided1 root
> on as1200. So far I came to point 16.3.9
> (http://www.netbsd.org/guide/en/chap-rf.html#chap-rf-ex-raid1root)
> 
> the quide says to install the boot loader on sda1 using the following
> command:
> 
> # /usr/sbin/installboot -v /dev/rsd1a /usr/mdec/bootblk
> 
> as i am installing on an alpha system and not sparc64, bootblk doesn't
> exist, so i tried /usr/mdec/boot.

For alpha, you want to use /usr/mdec/bootxx_ffs with installboot, and
also copy /usr/mdec/boot to /boot.  bootxx_ffs will be loaded by SRM,
then that loads /boot and /boot finally loads the kernel itself.

Simon.
--
Simon Burge                            <simonb <at> wasabisystems.com>
NetBSD Support and Service:         http://www.wasabisystems.com/
Permalink | Reply | 0 comments |
headers
Miles Nordin | 9 Aug 2005 14:08

Re: raided1 root / installboot drives me mad

>>>>> "s" == sp4rc  <sammyshome <at> gmx.net> writes:

     s> Why does installboot complains about a read-only file system?

Three things:

1. securelevel 1 forbids writing to the raw device that corresponds to
   a mounted partition.  If you can unmount your root filesystem, just
   copy boot to /boot, unmount it, and then run installboot.  The new
   installboot can work on unmounted filesystems.  

   Or, you can run installboot in single-user mode where securelevel
   is 0.  I think it is:

>>>boot -flags -s

   in the console, but that might not be exactly right.  Or, use '-d'
   to enter ddb on boot, and before the system mounts the root
   filesystem or configures most drivers, you'll be able to enter:

   ddb> write/l securelevel -1
   ddb> c

   it's untested, but I think that might work to stay insecure in
   multi-user, just for one boot-up.  Then you can reboot and get
   normal securelevel back.

2. sometimes you have to do:

   # sleep 90 < /dev/rsd1c &
(Continue reading)

Permalink | Reply | 1 comment |
headers
sp4rc | 9 Aug 2005 18:41
Picon

Re: raided1 root / installboot drives me mad

>    # sleep 90 < /dev/rsd1c &
>    # disklabel -W sd1
>    # installboot ...
>    # disklabel -N sd1

[...]

>    # cd /usr/mdec
>    # cp boot /
>    # sync
>    # installboot -v /dev/rsd1a bootxx_ffs /boot

Thank you guys so much. "disklabel -NW" and 
"installboot -v /dev/rsd1a bootxx_ffs /boot" did the trick. Now i am
proud owner of a raided 1 root netbsd alpha system xD

--

-- 
Regards
	/sp4rc
Permalink | Reply | 0 comments |
headers
Einar Karttunen | 7 Aug 2005 19:22
Picon
Picon
Favicon

X11 with VGA?

Hello

Is it possible to get VGA cards to work under NetBSD/alpha? I have a 
PCI Matrox millenium II card I would like to use in my alphastation,
but the X server provided in the X sets seems to be tga only.

At least on the netbsd-3 branch compiling x11/XFree86-* does not seem
to work (e.g. the libs build results in many errors, but installs
lacking most libraries from the PLIST file and being unusable).
Is there a solution to make things work?

- Einar Karttunen
Permalink | Reply | 3 comments |
headers
Torsten Harenberg | 10 Aug 2005 10:15
Picon
Favicon

Re: X11 with VGA?

Dear Einar,

have you allready looked here:

http://www.caseydoodle.com/index1.html

Best regards,

    Torsten

--

-- 
<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
<>                                                              <>
<> Torsten Harenberg         harenberg <at> physik.uni-wuppertal.de  <>
<> Bergische Universitaet                                       <>
<> FB C - Physik             Tel.: +49 (0)202 439-3521          <>
<> Gaussstr. 20              Fax : +49 (0)202 439-2811          <>
<> 42097 Wuppertal                                              <>
<>                                                              <>
<><><><><><><>< Of course it runs NetBSD http://www.netbsd.org ><>
Permalink | Reply | 0 comments |
headers
David Hopper | 10 Aug 2005 19:16

Re: X11 with VGA?

On Aug 7, 2005, at 10:22 AM, Einar Karttunen wrote:

> Hello
>
> Is it possible to get VGA cards to work under NetBSD/alpha? I have a
> PCI Matrox millenium II card I would like to use in my alphastation,
> but the X server provided in the X sets seems to be tga only.
>
> At least on the netbsd-3 branch compiling x11/XFree86-* does not seem
> to work (e.g. the libs build results in many errors, but installs
> lacking most libraries from the PLIST file and being unusable).
> Is there a solution to make things work?

Last I checked (which was, admittedly, some time ago), in-tree builds  
of x11 using build.sh do not work properly.  However, doing a 'make  
World' and 'make install' on a clean src/x11 tree (i.e. do not use  
build.sh) worked just fine for my alphastation.

The key is to build a static server, which build.sh -x will not do.

See http://mail-index.netbsd.org/tech-x11/2005/04/04/0001.html, and  
PR 29882.

>
> - Einar Karttunen
>

Thanks,
Dave
(Continue reading)

Permalink | Reply | 1 comment |
headers
sp4rc | 11 Aug 2005 08:54
Picon

AS1200 freeze

Hello list,

while trying to reinitialize the parity status of a raid1 set and doing
a "newfs -O 1 /dev/sd?a" on NetBSD 2.0.2 at the same time freezes my
AS1200 as soon as the newfs process has finished. 
1) Can anyone confirm this behaviour on his machine? 
2) Maybee this is a bug within the kernel, raidctl or newfs? 
3) If so how can i provide debug information?
--

-- 
Regards
	/sp4rc
Permalink | Reply | 0 comments |

Gmane