Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

2200M2 ELOM serial port

115 views
Skip to first unread message

David Lesher

unread,
Sep 27, 2009, 9:53:22 PM9/27/09
to
The 2200M2 has a 9600 bps serial port that talks to the Lights Out
Mgmt. If you farkle up the IP address; it's the suggested recovery
method.

I wasted several hours trying to accomplish this, to no avail. Of course,
I was fighting with a com program I didn't know, & a USB-serial adapter;
and all the things I've not missed about RS-232.

Does anyone have first-hand experience with this? AFICTell, the
only line the 2200 was asserting was TD!

(There is a back door via the console; but it's hard to find:
F2, then <RET> after "AMD" appears, then Advanced, IPMI, LAN..)

--
A host is a host from coast to coast.................wb8foz@nrk.com
& no one will talk to a host that's close........[v].(301) 56-LINUX
Unless the host (that isn't close).........................pob 1433
is busy, hung or dead....................................20915-1433

Tristram Scott

unread,
Sep 28, 2009, 6:54:22 AM9/28/09
to
In comp.sys.sun.admin David Lesher <wb8...@panix.com> wrote:
> The 2200M2 has a 9600 bps serial port that talks to the Lights Out
> Mgmt. If you farkle up the IP address; it's the suggested recovery
> method.
>

Default setting is 9600/8N1/no CTS/no XON XOFF.

> I wasted several hours trying to accomplish this, to no avail. Of course,
> I was fighting with a com program I didn't know, & a USB-serial adapter;
> and all the things I've not missed about RS-232.

That never helps. Practise against another server first.

>
> Does anyone have first-hand experience with this? AFICTell, the
> only line the 2200 was asserting was TD!

I have used the serial connection to the LOM quite a few times on mine, no
troubles at all.

I think that as with the x2100, there is a method in the BIOS for changing
the mode of the serial port between working for the ELOM and working as
COM1. The notes I have for the x2100 go like this:

To make use of an on-board serial port you need to set the following
setting in the BIOS.

BIOS--> Advanced--> IPMI2.0 Configuration --> External COM port

Set "External COM port" to "system" instead of the defaut "BMC"

Please note, that in case of any system failure you will have to attach a
keyboard and a monitor to gain access to the system.

Is it possible that your x2200 had been configured to use the port as COM1
rather than ELOM?


>
> (There is a back door via the console; but it's hard to find:
> F2, then <RET> after "AMD" appears, then Advanced, IPMI, LAN..)
>

Didn't know that one. Thanks.


--
Dr Tristram J. Scott
Energy Consultant
| England
tristra...@quantmodels.co.uk | ph (+44 1223) 526 255

David Lesher

unread,
Sep 28, 2009, 12:21:32 PM9/28/09
to
tristra...@ntlworld.com (Tristram Scott) writes:

>> I wasted several hours trying to accomplish this, to no avail. Of course,
>> I was fighting with a com program I didn't know, & a USB-serial adapter;
>> and all the things I've not missed about RS-232.

>That never helps. Practise against another server first.

I hauled out a modem and used that as a test object... It worked fine..


>BIOS--> Advanced--> IPMI2.0 Configuration --> External COM port

>Set "External COM port" to "system" instead of the defaut "BMC"

>Please note, that in case of any system failure you will have to attach a
>keyboard and a monitor to gain access to the system.

>Is it possible that your x2200 had been configured to use the port as COM1
>rather than ELOM?

I'm not seeing exactly that in the docs.

Part No. 819-6588-13 October 2007, Revision A
Part No. 819-6596-12 February 2007, Revision A


Changing Serial Output to Display the System Information The BIOS has
two options for serial redirection after the Embedded LOM SP has been
configured:

SP (default). Serial output displays the SP output (SP booting and CLI)
and the system console can be accessed by issuing a CLI command.

System. Serial output displays system information (that is. POST information)

By default, the BIOS is set to redirect its serial output to the SP. If
you would rather not have to issue an SP CLI command to access the BIOS
redirected serial stream, you can change the BIOS configuration option to
make the default serial redirection be from the BIOS rather than the SP.

....

3. Press F2 to enter the BIOS on the console. 4. Change the following
BIOS option under IPMI 2.0 configuration from SP to System. 5. Press F10
to save


>> (There is a back door via the console; but it's hard to find:
>> F2, then <RET> after "AMD" appears, then Advanced, IPMI, LAN..)
>>
>Didn't know that one. Thanks.

It's mentioned in the ELOM Fine Manual BUT they don't describe the <RET> needed
while the BOIS message is scrolling by; without that, it jumps into grub and runs.


I *think* what's going on is this. The serial method to reach the ELOM
works ONLY when the server is off; when it is on, the main process owns
the serial port. The server was on [fans running].

Is that possible?

Tristram Scott

unread,
Sep 28, 2009, 2:51:50 PM9/28/09
to
In comp.sys.sun.admin David Lesher <wb8...@panix.com> wrote:
> tristra...@ntlworld.com (Tristram Scott) writes:
>
>
> I *think* what's going on is this. The serial method to reach the ELOM
> works ONLY when the server is off; when it is on, the main process owns
> the serial port. The server was on [fans running].
>
> Is that possible?
>

It is certainly possible on the x2100 M2, but I am not sure for the X2200
M2. I had a need to use the serial port on our x2100 M2, and had to put in
a call to sun to find out how to configure the machine so this would work.
The default setting in the BIOS is to have the COM port be for the
exclusive use of the ELOM. I have not ever needed to do the same for the
x2200 M2 I have, so I don't have any notes on that.

I am aware that the ELOM is not the same on the x2100 and the x2200
machines, though, so my experience from the x2100 might not be on target
here. There doesn't seem to be any relavent information on SunSolve
either.

By the way, did you also try the other option of attaching a keyboard,
mouse and monitor? Obviously not much use if you are not located near the
machine.

David Lesher

unread,
Sep 28, 2009, 5:38:47 PM9/28/09
to
>> I *think* what's going on is this. The serial method to reach the ELOM
>> works ONLY when the server is off; when it is on, the main process owns
>> the serial port. The server was on [fans running].
>>
>> Is that possible?
>>

>It is certainly possible on the x2100 M2, but I am not sure for the X2200
>M2. I had a need to use the serial port on our x2100 M2, and had to put in
>a call to sun to find out how to configure the machine so this would work.
>The default setting in the BIOS is to have the COM port be for the
>exclusive use of the ELOM. I have not ever needed to do the same for the
>x2200 M2 I have, so I don't have any notes on that.

I failed on talking to it in "no fans" mode; but I'm still testing...

>There doesn't seem to be any relavent information on SunSolve
>either.

I think I need a keyboard macro for that phrase.... No one seems to
have a clue re: our ELOM failures.

>By the way, did you also try the other option of attaching a keyboard,
>mouse and monitor? Obviously not much use if you are not located near the
>machine.

That's how I solved the immediate issue of resetting the IP, once I found
the secret to get to the setup menu.

David Lesher

unread,
Sep 30, 2009, 2:55:13 PM9/30/09
to

(I'm giving up for now on the serial port; I can talk to other
devices OK so I think my serial term works OK.)

But I went to upgrade to the latest flash and ended up with more confusion.

I grabbed the tools_driver 2.2.0 ISO, opened it, found two images and no
useful help; finally found a mention elsewhere that you need one of them,
depending on if you have a .5M or a 1M flash chip.

I can't see how you can tell; the Download page say you need Motherboard
PN:375-3560-01 or later for the 3D15; but the ELOM says the server board
is PN 26S39MA0001; with current BIOS being S39_3B26.

I guessed that 323_3B27 was the right one; installed it with the ELOM GUI
tool, and the ELOM showed it was now upgraded to 3.23. Great.

But the Server Board BIOS is still S39_3B26. Huh?

(Then we get into the ELOM JAVA swamp... The ELOM needs JAVA too new for
my laptop, but I have temporary use of an IntelMac, and thought I'd
try it. Alas, when I try and start the pushed applet; it opens a text
edit box with XML instead of Java Web Start. Now to figure out how to
straighten THAT out.)

David Lesher

unread,
Sep 30, 2009, 8:22:05 PM9/30/09
to

I'm going downhill here. I flashed the BIOS to 3B27, and while ELOM is
still working, the server is dead. Note that per the ELOM, the BIOS is
still at 3B26...

When you power up, the fans spin, lights blink, but the USB and video
ports are dead, and the machine does nothing.

Flash older BIOS versions -- no change.

John D Groenveld

unread,
Sep 30, 2009, 10:11:52 PM9/30/09
to
In article <ha0snd$4p8$1...@reader1.panix.com>,

David Lesher <wb8...@panix.com> wrote:
>When you power up, the fans spin, lights blink, but the USB and video
>ports are dead, and the machine does nothing.

WAG, pull the power cables from your power supplies and try a
cold boot.

Also open the enclosure and confirm the P/N on the mainboard
if you don't have Sun's yellow/gold system spec sheets handy
which break out the parts shipped in your system.

John
groe...@acm.org

David Lesher

unread,
Sep 30, 2009, 11:21:31 PM9/30/09
to
groe...@cse.psu.edu (John D Groenveld) writes:

>In article <ha0snd$4p8$1...@reader1.panix.com>,
>David Lesher <wb8...@panix.com> wrote:
>>When you power up, the fans spin, lights blink, but the USB and video
>>ports are dead, and the machine does nothing.

>WAG, pull the power cables from your power supplies and try a
>cold boot.

Tried that 3-4 times...

>Also open the enclosure and confirm the P/N on the mainboard
>if you don't have Sun's yellow/gold system spec sheets handy
>which break out the parts shipped in your system.

How will I know what I'm looking for?

DoN. Nichols

unread,
Oct 1, 2009, 1:22:40 AM10/1/09
to
On 2009-10-01, David Lesher <wb8...@panix.com> wrote:
> groe...@cse.psu.edu (John D Groenveld) writes:
>
>>In article <ha0snd$4p8$1...@reader1.panix.com>,
>>David Lesher <wb8...@panix.com> wrote:
>>>When you power up, the fans spin, lights blink, but the USB and video
>>>ports are dead, and the machine does nothing.
>
>>WAG, pull the power cables from your power supplies and try a
>>cold boot.
>
> Tried that 3-4 times...
>
>>Also open the enclosure and confirm the P/N on the mainboard
>>if you don't have Sun's yellow/gold system spec sheets handy
>>which break out the parts shipped in your system.
>
> How will I know what I'm looking for?

Well ... based on the Sun Blade 1000 and Sun Blade 2000, there
should be a barcode label somewhere inconvenient on the system board (PC
people call it the Motherboard.) (On these two SB systems, it is hidden
by PCI cards which may or may not be installed). It will include both
the bar code and a human-readable number. The first seven digits of the
number will match one of the two numbers you are looking for -- but
there will be more digits as well, which turn the part number into a
serial number for the board.

Be prepared with a small LED flashlight to help in the search.
On the CPU modules for the SB-[12]000, the barcode label is gold
colored, but on the system boards, it is white, so I don't know what to
expect in your system.

Hopefully, you have not totally locked yourself out from an
ability to change the firmware in flash.

Good Luck,
DoN.

--
Email: <dnic...@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 ---

Tristram Scott

unread,
Oct 1, 2009, 3:43:54 AM10/1/09
to
In comp.sys.sun.admin David Lesher <wb8...@panix.com> wrote:
>
> I'm going downhill here. I flashed the BIOS to 3B27, and while ELOM is
> still working, the server is dead. Note that per the ELOM, the BIOS is
> still at 3B26...
>
> When you power up, the fans spin, lights blink, but the USB and video
> ports are dead, and the machine does nothing.
>
> Flash older BIOS versions -- no change.
>
>

Do you have a copy of the Sun Fire X2200 M2 Server Service Manual,
819-6597? Chapter 1 has a section on clearing the CMOS.

Switch off power, remove the top cover, remove the air duct, remove the
system battery (cr2032 or similar), short the contacts in the battery
holder for 60 seconds. Install battery and air duct and power on again.

David Lesher

unread,
Oct 1, 2009, 5:12:40 PM10/1/09
to
tristra...@ntlworld.com (Tristram Scott) writes:


>Do you have a copy of the Sun Fire X2200 M2 Server Service Manual,
>819-6597? Chapter 1 has a section on clearing the CMOS.

>Switch off power, remove the top cover, remove the air duct, remove the
>system battery (cr2032 or similar), short the contacts in the battery
>holder for 60 seconds. Install battery and air duct and power on again.

I see that but my fear is that also erases the ELOM setup. Does
anyone know if it does?

David Lesher

unread,
Oct 1, 2009, 5:50:03 PM10/1/09
to
"DoN. Nichols" <dnic...@d-and-d.com> writes:


> Well ... based on the Sun Blade 1000 and Sun Blade 2000, there
>should be a barcode label somewhere inconvenient on the system board (PC
>people call it the Motherboard.) (On these two SB systems, it is hidden
>by PCI cards which may or may not be installed). It will include both
>the bar code and a human-readable number. The first seven digits of the
>number will match one of the two numbers you are looking for -- but
>there will be more digits as well, which turn the part number into a
>serial number for the board.

I see a stenciled label 375-4561, near the CPU's, but I don't know if
that is the correct part. [It could be the bare board, not the stuffed
one, etc.]

The download page says "375-3560-01 or greater" which is little help..

John D Groenveld

unread,
Oct 1, 2009, 6:04:37 PM10/1/09
to
In article <ha386a$25h$1...@reader1.panix.com>,

David Lesher <wb8...@panix.com> wrote:
>I see a stenciled label 375-4561, near the CPU's, but I don't know if

375-4561 doesn't exist in the Sun System Handbook.

Assuming its 375-3461, try reflashing with the BIOS on 1.6 CD.
<URL:https://cds.sun.com/is-bin/INTERSHOP.enfinity/WFS/CDS-CDS_SMI-Site/en_US/-/USD/ViewProductDetail-Start?ProductRef=SunFire-X2200-M2-1.6-A-G-F@CDS-CDS_SMI>

John
groe...@acm.org

David Lesher

unread,
Oct 1, 2009, 6:50:45 PM10/1/09
to
groe...@cse.psu.edu (John D Groenveld) writes:


>375-4561 doesn't exist in the Sun System Handbook.

>Assuming its 375-3461, try reflashing with the BIOS on 1.6 CD.

><https://cds.sun.com/is-bin/INTERSHOP.enfinity/WFS/CDS-CDS_SMI-Site/en_US/-/USD/ViewProductDetail-Start?ProductRef=SunFire-X2200-M2-1.6-A-G-F@CDS-CDS_SMI>


Yes, I got it wrong; once I got nose to nose with, I see it is 3461-01 with
Sun FRU-2028QTF--0633MB800583.

DoN. Nichols

unread,
Oct 2, 2009, 1:47:01 AM10/2/09
to
On 2009-10-01, David Lesher <wb8...@panix.com> wrote:
> "DoN. Nichols" <dnic...@d-and-d.com> writes:
>
>
>> Well ... based on the Sun Blade 1000 and Sun Blade 2000, there
>>should be a barcode label somewhere inconvenient on the system board (PC
>>people call it the Motherboard.) (On these two SB systems, it is hidden
>>by PCI cards which may or may not be installed). It will include both
>>the bar code and a human-readable number. The first seven digits of the
>>number will match one of the two numbers you are looking for -- but
>>there will be more digits as well, which turn the part number into a
>>serial number for the board.
>
> I see a stenciled label 375-4561, near the CPU's, but I don't know if
> that is the correct part. [It could be the bare board, not the stuffed
> one, etc.]

Stenciled? You mean etched into the conductor layer? Or Part
of the stuffing labels silkscreen? The labels on the boards which I
have experience with are a white (or possibly orange) plastic film with
the barcodes and the human readable number printed on it -- so it can
change when some of the components used to stuff the board change. Of
course, this is not a SPARC system, and those are the only ones which I
have experience with, so they could have changed their standard
practices. Generally, I don't trust etched or silkscreened numbers,
only the film labels.

> The download page says "375-3560-01 or greater" which is little help..

Since you could get lots of different numbers with the same etch
as the stuffing components change.

Casper H.S. Dik

unread,
Oct 2, 2009, 4:40:50 AM10/2/09
to
David Lesher <wb8...@panix.com> writes:

>tristra...@ntlworld.com (Tristram Scott) writes:


>>Do you have a copy of the Sun Fire X2200 M2 Server Service Manual,
>>819-6597? Chapter 1 has a section on clearing the CMOS.

>>Switch off power, remove the top cover, remove the air duct, remove the
>>system battery (cr2032 or similar), short the contacts in the battery
>>holder for 60 seconds. Install battery and air duct and power on again.

>I see that but my fear is that also erases the ELOM setup. Does
>anyone know if it does?


I think so.

Did you try to power-on the system using the ELOM commands?

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.

David Lesher

unread,
Oct 2, 2009, 2:00:10 PM10/2/09
to
Casper H.S. Dik <Caspe...@Sun.COM> writes:


>>I see that but my fear is that also erases the ELOM setup. Does
>>anyone know if it does?


>I think so.

>Did you try to power-on the system using the ELOM commands?


Many times. No joy. I can flash the BIOS via the ELOM, but it appears
to have zero effect. I think I'll flashing back to a far older version
before I try pulling the battery.

It does not appear it reverts to loading from the CD/DVD drive.

David Lesher

unread,
Oct 2, 2009, 10:23:20 PM10/2/09
to
update:

I pulled the battery, shorted the pins, and when I powered up, it booted
from the CD/DVD. I loaded a 1.6 image, and flashed the machine.

It booted and ran.....It still had ELOM 2.32 but the BIOS was 3B25.

Then I booted a 2.2.0 DVD and it flashed the BIOS to 3B27.
I guess a 375-3461 is the 0.5MB RAM version.

Now, if only I could figure out the original ELOM problem....


David Lesher

unread,
Oct 2, 2009, 10:24:36 PM10/2/09
to

And lesson-learned; pulling the BIOS battery does NOT wipe the
ELOM settings.

David Lesher

unread,
Oct 4, 2009, 11:57:51 PM10/4/09
to

So last night after I got the box running; a friend and I used WireShark
to examine the session between the laptop and the 2200M2.

So we immediately saw that the Java-based Remote Console uses port 8890
as well as 443. Googling THAT gets me references to other folks with the
same issues; one reports 9000-9003 are needed as well.

I just want to gripe that nowhere in the Sun docs did I see any mention
of needed ports for ILOM or ELOM.

If that had, it would have same me a lot of wasted time; and I can't
be alone...

Tristram Scott

unread,
Oct 5, 2009, 4:43:49 AM10/5/09
to
In comp.sys.sun.admin David Lesher <wb8...@panix.com> wrote:
>
>
> So last night after I got the box running; a friend and I used WireShark
> to examine the session between the laptop and the 2200M2.
>
> So we immediately saw that the Java-based Remote Console uses port 8890
> as well as 443. Googling THAT gets me references to other folks with the
> same issues; one reports 9000-9003 are needed as well.
>
> I just want to gripe that nowhere in the Sun docs did I see any mention
> of needed ports for ILOM or ELOM.
>
> If that had, it would have same me a lot of wasted time; and I can't
> be alone...
>

I'm glad you have the machione running again.

I thought that your difficulty was with communicating via the serial port,
not through the network interface. Your gripe is valid, though. I
remember taking some time to track down that information myself.

David Lesher

unread,
Oct 5, 2009, 10:36:41 AM10/5/09
to
tristra...@ntlworld.com (Tristram Scott) writes:


>I'm glad you have the machione running again.

>I thought that your difficulty was with communicating via the serial port,
>not through the network interface. Your gripe is valid, though. I
>remember taking some time to track down that information myself.

You REALLY did not want to ask:

Saga:

A) Build box with ELOM over CAT5.

B) Deploy in colo

....


C) Have issue, can't get ELOM Remote Console to work

D) WTF?!?! It did at A)

E) Waste weeks looking for answers, building SSH tunnels, trying to get
an adjacent box &/or one of the 2200's own VM to run the Java so we can
sniff. As I recall, OpenSol lacks the Java needed. [Someone else tried
this; don't recall actual failure..]

F) Give up, pull box. [Takes coordination; it's not really a colo; it's
donated space in $BigCompany's room with theirs.]

G) Friend/Wireshark guru and I try and set up same to test.

H) Discover not even 64Bit Apple Java 6 will run ELOM; also fails - %&*%

I) Try and set ELOM to have temporary outside access so that a Sun
worshipper 60 miles away can look at it from there.

J) Wireshark guru/Friend botches address change [we were tired...this was
an after-dinner attempt] and lose access...

K) Try monitor, see F2 prompt but ZERO mention of the secret "hit a key
while BIOS scrolls by" clue.

L) Give up; bring back, waste many more hours on serial per Fine Manuals.

M) Stumble across K) backdoor. Eureka!! Fix ELOM IP's.

N) Flash via ELOM

O) Brick box when it fails.

P) Ask help.

Q) Get battery+CD image suggestion

R) Pull batt, burn CD's, rescue box.

S) Carry back to G's and run Wireshark with a Linux and XP VM's;

T) Futz with their Java until it works...

U) Run Wireshark; see ports. Took 5 min max.

V) [Yet to do: ask colo to open new ACL holes. Want to be SURE we
know all the ports needed..]

W) [Reinstall box.]

Y) Try and grok ORIGINAL problem; Xen keeps dying. (Meanwhile, XVM Server
has been abandoned.)

ALL the C through W because Sun did not say: "You need these ports" and
"You need this JAVA version..."

For the record; from what I've read: ILOM uses different ports, and has
different JAVA requirements.

0 new messages