RE: [Fwd: the Richmond saga continues]

[steven james <pyro@linuxlabs.com>]

Greetings,

Node 8 looks like it may have lost it's boot order configuration, and is
trying the HD before PXE. The Tyan BIOS frequently tries to boot the MBR
blindly (leaving the serial screen scrape ISR in place) resulting in a
hang.

The BIOS can be reset through minicom (hit reset on node, then hit F2 on
minicom. In the boot order menu, the UNDI should be above hard drvie.

The other node's boot issue likely is a transient net condition. This sort
of thing is fairly common in booting since all nodes demand the master at
once and even w/ gig on the master, max out it's bandwidth 5 times over.

I found the sec. master running. Just in case, I re-made it's swap space
and rebooted it. It is up now.

On the master, it sounds like the LED is actually the problem. The LEDs
refer to the internal case fans. Those fans are also connected to a beeper
that will emit a piercing alarm if any fan's RPM goes too low.

Really, the LEDs on those cases are a sort of one size fits all thing. One
might expect the T leds to refer to temperature, but really they just
indicate that they have power in this case. 

Just to be sure, I checked CPU temp and fan RPM on the nodes, and they
show normal (in /proc/sys/dev/sensors/w83782d-i2c-0-2d )

I will have a look at the remaining library issues.

G'day,
sjames


On Tue, 12 Nov 2002, Stefanovski, Sasko wrote:

> Hi,
> 
> I rebooted the pscm1 and all the nodes. I was bringing them up 10 at at
> time.
> The current status is:
> 1. node 8 is down. It hangs during POST just after HDD is detected. The
> null-modem cable is currently attached to that node.
> 2. some of the nodes would not boot right away. They ususaly hang just after
> bpslave is started during recv: /lib/libc-2.2.4.so
>    Probably some transient network problems, since they boot up after couple
> of retries.
> 3. Secondary master (node 48) hangs during boot, at "Adding swap space"
> 4. The FAN led is off on the Primary Master (does it mean that internal
> processor's fan does not work, since the main power fan was blowing air?
> 5. T1 led on node 45 is off. What this led is saying?
> 
> Sasko
> 
> -----Original Message-----
> From: gilfoyle [mailto:ggilfoyl@richmond.edu]
> Sent: Tuesday, November 12, 2002 9:11 AM
> To: Sasko Stafanovski
> Subject: [Fwd: the Richmond saga continues]
> 
> 
> hi sasko,
> 
>    see the attached message from steven james. when you get a 
> chance, please reboot the entire cluster and send him an email.
> 
> thanks-in-advance,
> 
> jerry
> 
> 

-- 
-------------------------steven james, director of research, linux labs
... ........ ..... ....                     230 peachtree st nw ste 701
the original linux labs                             atlanta.ga.us 30303
      -since 1995                              http://www.linuxlabs.com
                                   office 404.577.7747 fax 404.577.7743
-----------------------------------------------------------------------