Stability?

classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

Stability?

jim smith-3
What's the stability like on hppa?  I just acquired a c3600 [552 MHz
processor, 1024 Mb RAM, 18GB and 9Gb HD, CDROM], and loaded OpenBSD on
it [I have experience with i386, sparc and sparc64 with various BSD's,
so hppa is new to me].

I know from reading the mail that X is problematic, but for other
things [mailserver, ftp server] should it be OK?

I thought I might give it a stress test by building a couple of simple
ports [bash, dxpc] and it went well.

I thought I'd stress it more by compiling a kernel.  It puked on
untarring/ungzipping the source files.  Guess I won't do that.  

Any other 'gotchas' to watch out for?

--
Jim Smith <[hidden email]>

Reply | Threaded
Open this post in threaded view
|

Re: Stability?

Rob Sciuk
On Tue, 12 Feb 2008, Jim Smith wrote:

> What's the stability like on hppa?  I just acquired a c3600 [552 MHz
> processor, 1024 Mb RAM, 18GB and 9Gb HD, CDROM], and loaded OpenBSD on
> it [I have experience with i386, sparc and sparc64 with various BSD's,
> so hppa is new to me].
>
> I know from reading the mail that X is problematic, but for other
> things [mailserver, ftp server] should it be OK?
>
> I thought I might give it a stress test by building a couple of simple
> ports [bash, dxpc] and it went well.
>
> I thought I'd stress it more by compiling a kernel.  It puked on
> untarring/ungzipping the source files.  Guess I won't do that.
>
> Any other 'gotchas' to watch out for?


I'm running on HP9K K200 2cpu's, 512MBRam.  Tcl/Tk, SQLite, PostgreSQL and
some ISC software (DHCP/BIND) all compiled to completion ... and while I
was futzing around a couple of weeks back, trying to get some LVD
drives to work using a bus adapter, I've compiled the kernel, and all
seems well with a 4.2 snapshot ... so far, so good, but YMMV ...

[hppa.controlq.com:161>>% uname -a
OpenBSD hppa.controlq.com 4.2 xyzzy#0 hppa
[hppa.controlq.com:162>>% uptime
  3:48PM  up 25 days, 23 mins, 1 user, load averages: 0.07, 0.08, 0.08

HTH,
Rob Sciuk

Reply | Threaded
Open this post in threaded view
|

Re: Stability?

jim smith-3
In reply to this post by jim smith-3
On Wed, 13 Feb 2008 07:56:36 +1030
"Graham Gower" <[hidden email]> wrote:

> Perhaps you should report your bugs and then you will have stabiity.
> OpenBSD runs well on my 735/125.

Indeed--but let me be clear.  I was not saying 'OpenBSD 4.2 puked doing
XX and is therefore unstable and I hate it', but rather "I noticed it
puked on X, anything else to watch out for?".

I checked into patches for 4.2, but they require building a kernel.
Since the machine failed on untarring the source files, um, ah, I'm
short some resources to do that.  ;-)

I'll download the 11 Jan snapshot and give that a go.

--
Jim Smith <[hidden email]>

Reply | Threaded
Open this post in threaded view
|

Re: Stability?

Mark Kettenis
In reply to this post by jim smith-3
> Date: Tue, 12 Feb 2008 15:16:35 -0600
> From: Jim Smith <[hidden email]>
>
> What's the stability like on hppa?  I just acquired a c3600 [552 MHz
> processor, 1024 Mb RAM, 18GB and 9Gb HD, CDROM], and loaded OpenBSD on
> it [I have experience with i386, sparc and sparc64 with various BSD's,
> so hppa is new to me].

These machines are fairly reliable and amazingly fast for a 550 MHz
machine.

> I know from reading the mail that X is problematic, but for other
> things [mailserver, ftp server] should it be OK?

Should be mostly fine, although there is a freaky bug that sometimes
makes programs crash if you spawn a lot of them in a short time.  It
sometimes hits me when doing a "make obj" in preparation of doing a
"make build", and almost never under any other circumstances.  And the
problem seems to disappear completely if the machine has been up for a
while.

If you need 24/7 availability, you're probably better of using
something else, but otherwise I'd say give it a shot.

> I thought I'd stress it more by compiling a kernel.  It puked on
> untarring/ungzipping the source files.  Guess I won't do that.  

Puked in what sense?

> Any other 'gotchas' to watch out for?

Not really.

Reply | Threaded
Open this post in threaded view
|

Re: Stability?

Alastair Boyanich-2
In reply to this post by Rob Sciuk
Greetings,

I'm running on a A180C (staccato). It'll happily build the kernel
and a lot of things from pkgsrc.

The gotchas on the A180C I have is the date+time, it gets amnesia
each boot and greets you with:

[..snip!..]
bootpath: 8/16/5.6 class=1 flags=c0<autoboot,autosearch> hpa=0xffd06000 spa=0x0
io=0x12000
root on sd0a swap on sd0b dump on sd0b
WARNING: clock lost 14083 days -- CHECK AND RESET THE DATE!

The other semi-gotcha is that for disk IO it's very slow compared to
hockey-pucks on the same machine (dual boot off two disks). obsd gets
around 620kb/sec and hockey-pucks around 1.2mb/sec. Given that it's
narrow se, that's not "that" bad and it's still highly usable for
tooling around and building random things. Also interrupt servicing
during heavy IO is a killer. Don't forget to include softdep for
your file system. Every speed-bump can count. C3600's potentially
have a hell of a lot more IO bandwidth. Haven't tried obsd on
my c3750's so I can't speak authoratively. I do seem to recall they
only run in 32bit mode though. Not a problem for the A180C though
because it's only pa1.1 7300lc.

Also, this may not effect you but I've had problems trying to get
the ethernet card in the GSC slot working. Your allegro doesn't
have GSC so that should be a moot point.

--
Alastair Boyanich <[hidden email]>


On Tue, 12 Feb 2008 16:51:18 -0500
Rob Sciuk <[hidden email]> wrote:

> On Tue, 12 Feb 2008, Jim Smith wrote:
> > What's the stability like on hppa?  I just acquired a c3600 [552 MHz
> > processor, 1024 Mb RAM, 18GB and 9Gb HD, CDROM], and loaded OpenBSD on
> > it [I have experience with i386, sparc and sparc64 with various BSD's,
> > so hppa is new to me].
> >
> > I know from reading the mail that X is problematic, but for other
> > things [mailserver, ftp server] should it be OK?
> >
> > I thought I might give it a stress test by building a couple of simple
> > ports [bash, dxpc] and it went well.
> >
> > I thought I'd stress it more by compiling a kernel.  It puked on
> > untarring/ungzipping the source files.  Guess I won't do that.
> >
> > Any other 'gotchas' to watch out for?
>
>
> I'm running on HP9K K200 2cpu's, 512MBRam.  Tcl/Tk, SQLite, PostgreSQL and
> some ISC software (DHCP/BIND) all compiled to completion ... and while I
> was futzing around a couple of weeks back, trying to get some LVD
> drives to work using a bus adapter, I've compiled the kernel, and all
> seems well with a 4.2 snapshot ... so far, so good, but YMMV ...
>
> [hppa.controlq.com:161>>% uname -a
> OpenBSD hppa.controlq.com 4.2 xyzzy#0 hppa
> [hppa.controlq.com:162>>% uptime
>   3:48PM  up 25 days, 23 mins, 1 user, load averages: 0.07, 0.08, 0.08
>
> HTH,
> Rob Sciuk
>
>

This is an email from Fujitsu Australia Software Technology Pty Ltd, ABN 27 003 693 481. It is confidential to the ordinary user of the email address to which it was addressed and may contain copyright and/or legally privileged information. No one else may read, print, store, copy or forward all or any of it or its attachments. If you receive this email in error, please return to sender. Thank you.

If you do not wish to receive commercial email messages from Fujitsu Australia Software Technology Pty Ltd, please email [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Stability?

jim smith-3
In reply to this post by Mark Kettenis
On Tue, 12 Feb 2008 23:24:21 +0100 (CET)
Mark Kettenis <[hidden email]> wrote:

> > I thought I'd stress it more by compiling a kernel.  It puked on
> > untarring/ungzipping the source files.  Guess I won't do that.  
>
> Puked in what sense?

In doing "sudo tar -xzf src.tar.gz" in /usr, the machine ground away
for awhile and then became unuseable.  I was ssh'd into it from another
machine, and had several xterms running simultaneously.  If I tried
"l" [which I have aliased as ls -laFT], it would output 'abort trap'.

If I entered the complete command "ls -laFT" I could list the contents
of a directory.  "top" produced an abort trap, "df" didn't.  Trying any
additional ssh sessions were unsuccessful.  I tried to log in at the
console [serial connected laptop], and that was unsuccessful.

On reboot, as it began to start the network daemons, there were a bunch
of abort traps and some interesting, weird outputs [along the lines
of "/etc/rc   drwxrwxrwx" blah blah (not in front of the machine right
now)] that indicates to me that somehow a directory listing was written
over a file in /etc/rc?

I thought at first it was a bad HD, but I've reproduced it on both the
9GB and the 18GB HD--it does exactly the same thing on both.

I'll install the snapshot from 11 Jan tonight and see how that goes.

--
Jim Smith <[hidden email]>

Reply | Threaded
Open this post in threaded view
|

Re: Stability?

Maurice Janssen-2
In reply to this post by Mark Kettenis
On Tuesday, February 12, 2008 at 23:24:21 +0100, Mark Kettenis wrote:

>> Date: Tue, 12 Feb 2008 15:16:35 -0600
>> From: Jim Smith <[hidden email]>
>>
>> What's the stability like on hppa?  I just acquired a c3600 [552 MHz
>> processor, 1024 Mb RAM, 18GB and 9Gb HD, CDROM], and loaded OpenBSD on
>> it [I have experience with i386, sparc and sparc64 with various BSD's,
>> so hppa is new to me].
>
>These machines are fairly reliable and amazingly fast for a 550 MHz
>machine.
>
>> I know from reading the mail that X is problematic, but for other
>> things [mailserver, ftp server] should it be OK?
>
>Should be mostly fine, although there is a freaky bug that sometimes
>makes programs crash if you spawn a lot of them in a short time.  It
>sometimes hits me when doing a "make obj" in preparation of doing a
>"make build", and almost never under any other circumstances.  And the
>problem seems to disappear completely if the machine has been up for a
>while.

I noticed this as well on a c3650.  'make obj' usually fails when
/usr/src is NFS mounted but I haven't seen any problems when the source
is stored on a local disk.

Maurice

Reply | Threaded
Open this post in threaded view
|

Re: Stability?

Travers Buda-2
In reply to this post by jim smith-3
* Jim Smith <[hidden email]> [2008-02-12 15:16:35]:

> What's the stability like on hppa?  I just acquired a c3600 [552 MHz
> processor, 1024 Mb RAM, 18GB and 9Gb HD, CDROM], and loaded OpenBSD on
> it [I have experience with i386, sparc and sparc64 with various BSD's,
> so hppa is new to me].
>
> I know from reading the mail that X is problematic, but for other
> things [mailserver, ftp server] should it be OK?
>
> I thought I might give it a stress test by building a couple of simple
> ports [bash, dxpc] and it went well.
>
> I thought I'd stress it more by compiling a kernel.  It puked on
> untarring/ungzipping the source files.  Guess I won't do that.  
>
> Any other 'gotchas' to watch out for?
>
> --
> Jim Smith <[hidden email]>
>
>

I've not had any stability problems on several C-class machines
(C240, C360,) several B180L+'es and a 712/60.  These all had local
disks though, they were not using NFS.  It you're having issues,
I'd suspect the ram.  Mickey says there's some reboot bug on the
C360, but I have not seen it.  

--
Travers Buda

Reply | Threaded
Open this post in threaded view
|

Re: Stability?

Marco Peereboom
In reply to this post by Mark Kettenis
So am I the only one that ran hppa for years as a firewall??

On Tue, Feb 12, 2008 at 11:24:21PM +0100, Mark Kettenis wrote:

> > Date: Tue, 12 Feb 2008 15:16:35 -0600
> > From: Jim Smith <[hidden email]>
> >
> > What's the stability like on hppa?  I just acquired a c3600 [552 MHz
> > processor, 1024 Mb RAM, 18GB and 9Gb HD, CDROM], and loaded OpenBSD on
> > it [I have experience with i386, sparc and sparc64 with various BSD's,
> > so hppa is new to me].
>
> These machines are fairly reliable and amazingly fast for a 550 MHz
> machine.
>
> > I know from reading the mail that X is problematic, but for other
> > things [mailserver, ftp server] should it be OK?
>
> Should be mostly fine, although there is a freaky bug that sometimes
> makes programs crash if you spawn a lot of them in a short time.  It
> sometimes hits me when doing a "make obj" in preparation of doing a
> "make build", and almost never under any other circumstances.  And the
> problem seems to disappear completely if the machine has been up for a
> while.
>
> If you need 24/7 availability, you're probably better of using
> something else, but otherwise I'd say give it a shot.
>
> > I thought I'd stress it more by compiling a kernel.  It puked on
> > untarring/ungzipping the source files.  Guess I won't do that.  
>
> Puked in what sense?
>
> > Any other 'gotchas' to watch out for?
>
> Not really.

Reply | Threaded
Open this post in threaded view
|

Re: Stability?

jim smith-3
In reply to this post by jim smith-3
On Tue, 12 Feb 2008 23:46:17 -0500
"Jeff Quast" <[hidden email]> wrote:

> On Feb 12, 2008 5:36 PM, Jim Smith <[hidden email]> wrote:
> >  that indicates to me that somehow a directory listing was written
> > over a file in /etc/rc?
>
> I often see this problem with bad memory on any architecture.
>
> I saw this exact same problem with an hppa that came with bad memory
> (abort traps and disk corruption).
>
> I replaced the ram, and it fixed it.
>
> jq

That is good to know.  I did pull the RAM and reseat it when I received
this machine.  I'll check eBay to see what's available.

--
jim smith <[hidden email]>

Reply | Threaded
Open this post in threaded view
|

Re: Stability?

jim smith-3
On Wed, 13 Feb 2008 20:57:33 -0500
"Jeff Quast" <[hidden email]> wrote:
>
> I purchased replacement ram for my 712/80 (hard to come by HP Apollo
> ram) through https://www.memoryx.net/
>
> I was able to find it by searching for the p/n, I suspect they may
> have the ram you need as well (and provide reliable return service and
> warranty, something ebay can't help you with)

Thanks--good to know!  I did pull one stick of RAM tonight and it's
building the kernel OK [as of 3 min ago].  One thing to note,
however--I had placed src.tar.gz in /usr and was untarring it to /usr
rather than /usr/src [up too late].  That may have been what the
problem was.

--
Jim Smith <[hidden email]>