sparc 10 hanging

Hi all,

we have a sparc 10 as an embedded device ,and it runs 2.4 (sunos
4.1.3).

I get a
WARNING: TOD clock not initialized -- CHECK AND RESET THE DATE!

however the MAC is still there so I dont thinkg the vram battery is
dead

and the last line I see before the hang is

dump on sdb0 fstype spec size 66628k

is that signinficant? the next line I think should be the le0 ethernet
connection ?

any advice on how to proceed?

On 2009-07-06, merrittr <> wrote:
> Hi all,
>
> we have a sparc 10 as an embedded device ,and it runs 2.4 (sunos
> 4.1.3).

Hmm ... SunOs 4.1.3 is *not* the same as Solaris 2.4. The latter
is SunOs 5.4, FWIW. The former was called "Solaris 1.1.1" IIRC, with
Solaris 1.1.2 being the last BSD based one -- SunOs 4.1.4.

> I get a
> WARNING: TOD clock not initialized -- CHECK AND RESET THE DATE!

Is this during install? SunOs 4.1.3 (and others of the SunOs
4.1 family) are BSD based, and compare the TOD clock to the time stored
in the filesytem during boot. Since the time stored in the filesystem
is on CD-ROM, it *can't* be updated (it is stuck somewhen in the early
1990s), so every time you boot from the CD-ROM you will get this message
(usually worded a bit differently, what you are seeing suggests that it
is instead actual problems getting the TOD clock chip to run. What *is*
the MAC address shown, and does the HostId match what a SS-10 should be?
According to the FEH, a SS10 should start with 72xxxxxx, and should have
a 48T08 NVRAM. And if the NVRAM (TOD clock chip) has been replaced with
a more modern one, there is an incompatibility -- Sun used an
undocumented feature in the chip which went away when the manufacturer
upgraded it.

The boot message was also common with booting and installing
from tape.

> however the MAC is still there so I dont thinkg the vram battery is
> dead

make sure the hostid (shown with the MAC address before booting
starts) starts with 72 as above. Otherwise, it can confuse the OS as to
what configuration to look for.

> and the last line I see before the hang is
>
> dump on sdb0 fstype spec size 66628k

Hmm ... is it actually saying that it is dumping, or just that
that sdb0 (hmm shouldn't that be sd0b?) is the device used for dumps?

> is that signinficant? the next line I think should be the le0 ethernet
> connection ?

Is there an ethernet connected? Is the system configured to
boot via the net instead of from its own disk drive?

> any advice on how to proceed?

More information?

Try pulling the disk and enabling diag mode in the NVRAM. See
what it complains about as it goes through the tests. (I forget whether
the SS-10 has an option to set the level of the diag or just the
"diag-mode?=true" (or false). I don't have one in a position to check
these days. You might have some flakey memory. Do you have two CPUs or
only one installed?

Do you have a Sun keyboard plugged in? If not, then it will be
sending subsequent diagnostics information to TTYA instead of the
screen.

Of course, I don't know what "embedded" may have done to change
this beastie.

Good Luck,
DoN.

--
Email: <> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

On Jul 6, 6:59*pm, "DoN. Nichols" <dnich...@d-and-d.com> wrote:
> On 2009-07-06, merrittr <merri...@gmail.com> wrote:
>
> > Hi all,
>
> > we have a sparc 10 as an embedded device ,and it runs 2.4 (sunos
> > 4.1.3).
>
> * * * * Hmm ... SunOs 4.1.3 is *not* the same as Solaris 2.4. Thelatter
> is SunOs 5.4, FWIW. *The former was called "Solaris 1.1.1" IIRC, with
> Solaris 1.1.2 being the last BSD based one -- SunOs 4.1.4.
>
> > I get a
> > WARNING: TOD clock not initialized * -- CHECK AND RESET THE DATE!
>
> * * * * Is this during install? *SunOs 4.1.3 (and others of theSunOs
> 4.1 family) are BSD based, and compare the TOD clock to the time stored
> in the filesytem during boot. *Since the time stored in the filesystem
> is on CD-ROM, it *can't* be updated (it is stuck somewhen in the early
> 1990s), so every time you boot from the CD-ROM you will get this message
> (usually worded a bit differently, what you are seeing suggests that it
> is instead actual problems getting the TOD clock chip to run. *What *is*
> the MAC address shown, and does the HostId match what a SS-10 should be?
> According to the FEH, a SS10 should start with 72xxxxxx, and should have
> a 48T08 NVRAM. *And if the NVRAM (TOD clock chip) has been replaced with
> a more modern one, there is an incompatibility -- Sun used an
> undocumented feature in the chip which went away when the manufacturer
> upgraded it.
>
> * * * * The boot message was also common with booting and installing
> from tape.
>
> > however the MAC is still there so I dont thinkg the vram battery is
> > dead
>
> * * * * make sure the hostid (shown with the MAC address before booting
> starts) starts with 72 as above. *Otherwise, it can confuse the OS as to
> what configuration to look for.
>
> > and the last line I see before the hang is
>
> > dump on sdb0 fstype spec size 66628k
>
> * * * * Hmm ... is it actually saying that it is dumping, or justthat
> that sdb0 (hmm shouldn't that be sd0b?) is the device used for dumps?
>
> > is that signinficant? the next line I think should be the le0 ethernet
> > connection ?
>
> * * * * Is there an ethernet connected? *Is the system configured to
> boot via the net instead of from its own disk drive?
>
> > any advice on how to proceed?
>
> * * * * More information?
>
> * * * * Try pulling the disk and enabling diag mode in the NVRAM.*See
> what it complains about as it goes through the tests. *(I forget whether
> the SS-10 has an option to set the level of the diag or just the
> "diag-mode?=true" (or false). *I don't have one in a position to check
> these days. *You might have some flakey memory. *Do you have two CPUsor
> only one installed?
>
> * * * * Do you have a Sun keyboard plugged in? *If not, then itwill be
> sending subsequent diagnostics information to TTYA instead of the
> screen.
>
> * * * * Of course, I don't know what "embedded" may have done to change
> this beastie.
>
> * * * * Good Luck,
> * * * * * * * * DoN.
>
> --
> *Email: * <dnich...@d-and-d.com> * | Voice (all times): (703) 938-4564
> * * * * (too) near Washington D.C. |http://www.d-and-d.com/dnichols/DoN.html
> * * * * * *--- Black Holes are where God is dividing by zero ---

thanks don,

Nope not installing this ss-10 was running last week fine, it controls
a robot arm
run by hydraulics. It boots off its own drive (I have an image of the
drive I made last year too on external disk).
the ethernet on the thing is crossed over to a motorola controler card
so that shouldn't be the problem either?
I can stop-a it and tried boot -s but got the same result of it
hanging. Also the MAC is correct and before it was rebooted the
lab guy looking after the robot did a date <not sure what parameter>
then rebooted thats when this all started.
Any advice?

whoops the MAC is actually

00:00:3b:80:07:a0

merrittr schrieb:
> whoops the MAC is actually
>
> 00:00:3b:80:07:a0
the mac is wrong....

sun mac adresses start with 08:00:20 or 8:0:20

Best regards, David

David Reichelt <Freud-> writes:

>merrittr schrieb:
>> whoops the MAC is actually
>>
>> 00:00:3b:80:07:a0
>the mac is wrong....

>sun mac adresses start with 08:00:20 or 8:0:20

>Best regards, David

Many more:

0:3:ba:0:0:0
0:14:4f:0:0:0 (x86 and SPARC, bge)
0:23:8b:0:0:0
0:b:5d:0:0:0 (Ok, this one is a Fujitsu)

Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.

On Jul 7, 2:13*pm, v_borch...@despammed.com (Volker Borchert) wrote:
> David Reichelt wrote:
> > merrittr schrieb:
> > > whoops the MAC is actually
>
> > > 00:00:3b:80:07:a0
> > the mac is wrong....
>
> > sun mac adresses start with 08:00:20 or 8:0:20
>
> According to my list of vendor prefixes, sun also has 0:0:7d and 0:20:f2.
> 0:0:3b would be i Controls, Inc., KOREA; given that this is an embedded
> system, it might be correct.
>
> --
>
> "I'm a doctor, not a mechanic." Dr Leonard McCoy <mc...@ncc1701.starfleet..fed>
> "I'm a mechanic, not a doctor." Volker Borchert *<v_borch...@despammed.com>

actually its labeled

AXIL HWS310.

so these guys might be using a non sun NIC? it is a valid NIC.


But the thing that changed before the hang is the Robot guy whos lab
its in
did a date <somthing>
and now it hanges is ther a way to reset the date and see if it boots
all the way?

On 2009-07-07, merrittr <> wrote:
> On Jul 6, 6:59*pm, "DoN. Nichols" <dnich...@d-and-d.com> wrote:
>> On 2009-07-06, merrittr <merri...@gmail.com> wrote:
>>
>> > Hi all,
>>
>> > we have a sparc 10 as an embedded device ,and it runs 2.4 (sunos
>> > 4.1.3).
>>
>> * * * * Hmm ... SunOs 4.1.3 is *not* the same as Solaris 2.4. The latter
>> is SunOs 5.4, FWIW. *The former was called "Solaris 1.1.1" IIRC, with
>> Solaris 1.1.2 being the last BSD based one -- SunOs 4.1.4.
>>
>> > I get a
>> > WARNING: TOD clock not initialized * -- CHECK AND RESET THE DATE!

[ ... lots of my text snipped ... ]

> thanks don,
>
> Nope not installing this ss-10 was running last week fine, it controls
> a robot arm
> run by hydraulics. It boots off its own drive (I have an image of the
> drive I made last year too on external disk).

Hmm ... does the image drive have a bootblock installed?
Without that, you can't boot it. And it will either have to have things
like the /etc/vfstab edited to change the drive being accessed, or the
bootblock installed for it to boot from the drive placed where the
original one was.

> the ethernet on the thing is crossed over to a motorola controler card
> so that shouldn't be the problem either?

O.K. As long as it sees something so it does not hang trying to
get a connection for NIS or whatever.

> I can stop-a it and tried boot -s but got the same result of it
> hanging.

So -- if you can stop-a it, that means that you *do* have a Sun
keyboard connected. Have you tried hooking a serial port to TTYA
unplugging the keyboard, and trying to boot? That should give you more
diagnostic information if things are getting confused.

> Also the MAC is correct and before it was rebooted the
> lab guy looking after the robot did a date <not sure what parameter>
> then rebooted thats when this all started.

Hmm ... "date" while the system was booted, I presume? But if
the battery in the NVRAM is getting tired, the clock chip might get
confused before the rest. Get the barcode number from the NVRAM chip
and talk to Sun about getting a replacement. Even if the battery is
still good, it probably won't be for long. This is a long time for
those to live even if the system is powered up 24/7/365, and if it is
turned off overnight, it is almost certainly reaching the end of its
life. I don't know whether the entirity of the hostid is critical for
you. It depends on whether you have any software on the machine which
depends on a software key. If you do, then you *will* need to talk to
Sun about the replacement giving them the barcode number so they can
look up and program the Hostid and the MAC address into the new
NVRAM/clock chip.

What about the hostid from the screen during boot? Does it
start with "72" as it should? I don't know whether the clock or the
NVRAM is the most sensitive to battery voltage.

> Any advice?

Mostly what I already gave before.

Do you have the CD-ROM for the OS, a bootable CD-ROM drive, and
a spare disk to swap in for testing? See whether you can boot the
CD-ROM and start installing on a spare disk. If not, then there is
something going wrong with the system board or the CPU modules. (Does
it have just one or two CPUs? If it has two, try booting with only one
installed at a time. If the first one works, still pull it and swap in
the second one.

The mention about "bootable CD-ROM drive" is because the drive
needs firmware which starts it up as 512 bytes per sector, not the 2048
bytes per sectors which is more common. If it is a Sun supplied CD-ROM
drive -- especially one built into the system, it should be right. If it
is a Sun drive in an external Sandwichbox housing, you still should be
O.K. You'll need an external terminator in that case.

And do you know for sure *which* OS you want? There was a bit
of confusion in your first post -- and it either needs to be SunOs 4.1.3
(Solaris 1.1.1) or SunOs 5.x, (Solaris 2.x).

And -- do you have a separate copy (other than the duplicate
disk) of the software which runs the robotic arm? (If anything, this is
what is likely to be software key locked, and you'll need to run FlexLM
(License Manager) to serve the keys.

Minimum setup is one bank of RAM, one CPU, and a serial terminal
on TTYA. If everything works there, start adding more things until
something goes wrong, and look closely at what you just re-installed as
the probable cause of the problem.

Good Luck,
DoN.

--
Email: <> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

On 2009-07-07, merrittr <> wrote:
> whoops the MAC is actually
>
> 00:00:3b:80:07:a0

Hmm ... I think that from that period, you should have "8:0:20:" as
the leading characters. At least that is what my two SS-5 systems have,
along with an Ultra5 and two Ultra-10s. Another reason to suspect the
battery in the NVRAM/clock chip.

The later machines (which I am running (SF-280R, SB-2000,
SB-1000) start with "0:3:ba:" and are running 100BaseT instead of the
10BaseT or AUI of the SS-10. Also, a QFE card (Quad Fast Ethernet) in
the SS-5 from Sun also starts with "0:3:ba:" I guess that sun used up
the full run of the "8:0:20:" series and acquired another block to work
with.

Good Luck,
DoN.

--
Email: <> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

DoN. Nichols wrote:

> Hmm ... does the image drive have a bootblock installed?

> > and the last line I see before the hang is
> >
> > dump on sdb0 fstype spec size 66628k

Looking at the dmesg of this here diskless 10 running 4.1.4, this
seems to be output by the kernel after probing the devices, so the
bootblock should be OK.

--

"I'm a doctor, not a mechanic." Dr Leonard McCoy <>
"I'm a mechanic, not a doctor." Volker Borchert <>