Mailing List Archive

Isn't ^ec-l1 defined by default, i.e. "<Ctrl-e> c l 1" in sequence?

It's been a few years but my fingers seem to remember that break sequence....

Rich
--
Rich Holland (913) 645-1950 SAP Technical Consultant
print unpack("u","92G5S\=\"!A;F]T:&5R(\'!E<FP\@:&%C:V5R\"\@\`\`");

> -----Original Message-----
> From: users-bounces@conserver.com [mailto:users-bounces@conserver.com] On
> Behalf Of Pete Geenhuizen
> Sent: Thursday, April 29, 2004 1:38 PM
> To: users@conserver.com
> Subject: Help with break command
>
> I'm just getting started with conserver. I've pretty much figured out
> what I want to do, but one thing eludes me.
>
> I'm running conserver on a Sun network, and for the life of me I can't
> figure out how to send a break sequence to halt the OS.
>
> There are some examples of break sequences, but I can't get any of them to
> work. I'm used to sending <CTRL>]
> telnet> send break
> or
> sending the alternate break sequence of <RETURN>~<CTRL>b
> So how do I configure conserver's break seqences to do the same thing?
>
> In conserver.cf I currently have "\z" as the sun-std which does nothing
> but print some odd characters and \377\363: not found
> and
> "\r\d~\d^b" which does nothing but print ^b
> BTW this host is just using the standard break sequence.
>
> Any help or insight would be greatly appreciated.
>
> Thx in advance.
> Pete
> --
> "Unencumbered by the thought process"
> --1992-2000 Click and Clack presidential campaign slogan
>
>
> _______________________________________________
> users mailing list
> users@conserver.com
> https://www.conserver.com/mailman/listinfo/users

What sort of sun system?

Was "kbd -a alternate" run or was KEYBOARD_ABORT set in
/etc/default/kbd before the last reboot?

What's in between the conserver process and the sun console? (direct
serial line? terminal server? ALOM/system controller?)

- Bill

Rich,
You're right about ^ec-l1. It's defined in the config file as the
standard sun break sequence, unfortunately it doen't work.

Pete
--
"Unencumbered by the thought process"
--1992-2000 Click and Clack presidential campaign slogan

Rich Holland said:
> Isn't ^ec-l1 defined by default, i.e. "<Ctrl-e> c l 1" in sequence?
>
> It's been a few years but my fingers seem to remember that break
> sequence....
>
> Rich
> --
> Rich Holland (913) 645-1950 SAP Technical Consultant
> print unpack("u","92G5S\=\"!A;F]T:&5R(\'!E<FP\@:&%C:V5R\"\@\`\`");
>
>> -----Original Message-----
>> From: users-bounces@conserver.com [mailto:users-bounces@conserver.com]
>> On
>> Behalf Of Pete Geenhuizen
>> Sent: Thursday, April 29, 2004 1:38 PM
>> To: users@conserver.com
>> Subject: Help with break command
>>
>> I'm just getting started with conserver. I've pretty much figured out
>> what I want to do, but one thing eludes me.
>>
>> I'm running conserver on a Sun network, and for the life of me I can't
>> figure out how to send a break sequence to halt the OS.
>>
>> There are some examples of break sequences, but I can't get any of them
>> to
>> work. I'm used to sending <CTRL>]
>> telnet> send break
>> or
>> sending the alternate break sequence of <RETURN>~<CTRL>b
>> So how do I configure conserver's break seqences to do the same thing?
>>
>> In conserver.cf I currently have "\z" as the sun-std which does nothing
>> but print some odd characters and \377\363: not found
>> and
>> "\r\d~\d^b" which does nothing but print ^b
>> BTW this host is just using the standard break sequence.
>>
>> Any help or insight would be greatly appreciated.
>>
>> Thx in advance.
>> Pete
>> --
>> "Unencumbered by the thought process"
>> --1992-2000 Click and Clack presidential campaign slogan
>>
>>
>> _______________________________________________
>> users mailing list
>> users@conserver.com
>> https://www.conserver.com/mailman/listinfo/users
>

Oops, forgot to mention this happens to be a Sun280R
--
"Unencumbered by the thought process"
--1992-2000 Click and Clack presidential campaign slogan

Bill Sommerfeld said:
> What sort of sun system?
>
> Was "kbd -a alternate" run or was KEYBOARD_ABORT set in
> /etc/default/kbd before the last reboot?
>
> What's in between the conserver process and the sun console? (direct
> serial line? terminal server? ALOM/system controller?)
>
> - Bill
>

Hmm, I wonder what I'm doing wrong then? How is the sequence defined in
the config file? I have break 1 { string "\z"; }

Could the terminal type that I'm using have something to do with it?

Pete
--
"Unencumbered by the thought process"
--1992-2000 Click and Clack presidential campaign slogan

Bill Sommerfeld said:
> The default -- telnet break request -- just worked out of the box for
> me with a couple of the older 64-port annex 3's as well as a digi
> ts-16.
>
> On the client end, I always enter ^E c l 0 to send the default break
> sequence.
>
> - Bill
>
>
>
>
>

> Hmm, I wonder what I'm doing wrong then? How is the sequence defined in
> the config file? I have break 1 { string "\z"; }
>
> Could the terminal type that I'm using have something to do with it?

The conserver.cf I'm using has absolutely no special break
configuration; it's using whatever the startup-time default is
(sequence 0 being "\z").

I'd suggest a divide-and-conquer approach..

Disable the port in conserver, and connect to the TS directly with a
high-port telnet and do a "send break". Whether or not that works
will provide useful information...

- Bill

Pete Geenhuizen wrote:

> Oops, forgot to mention this happens to be a Sun280R

Are you connected to the regular console port, or the RSC console port?

Have you redirected console control to the RSC?

-T.

--
Trevor Fiatal -- trevor@fiatal.net
650.862.3715 (mobile)

Well I've discovered one thing, using conserver I can get the system to
halt by sending ^Ecl0 at the login prompt, but if I log in it fails and
spits out strange characters and ignores the sequence. It looks like the
shell is intercepting conserver's break sequence and munging it up.

I hope that that isn't the expected behaviour.

I have disabled conserver and connected to the box directly with telnet
through the terminal server, and sending a break works just fine whether
I'm at the login prompt or logged in.

FWIW the terminal type on the remote host is set to vt100 and the shell is
either borne or korn.

Pete
--
"Unencumbered by the thought process"
--1992-2000 Click and Clack presidential campaign slogan

Bill Sommerfeld said:
>> Hmm, I wonder what I'm doing wrong then? How is the sequence defined in
>> the config file? I have break 1 { string "\z"; }
>>
>> Could the terminal type that I'm using have something to do with it?
>
> The conserver.cf I'm using has absolutely no special break
> configuration; it's using whatever the startup-time default is
> (sequence 0 being "\z").
>
> I'd suggest a divide-and-conquer approach..
>
> Disable the port in conserver, and connect to the TS directly with a
> high-port telnet and do a "send break". Whether or not that works
> will provide useful information...
>
> - Bill
>

Nope, not using RSC, just connected to ttya.
I can get to the OK prompt doing a telnet to the terminal server and the
appropriate port.
Using conserver I can get to the OK prompt before I log in, not after I
login.


Pete
--
"Unencumbered by the thought process"
--1992-2000 Click and Clack presidential campaign slogan

Trevor Fiatal said:
> Pete Geenhuizen wrote:
>
>> Oops, forgot to mention this happens to be a Sun280R
>
> Are you connected to the regular console port, or the RSC console port?
>
> Have you redirected console control to the RSC?
>
> -T.
>
> --
> Trevor Fiatal -- trevor@fiatal.net
> 650.862.3715 (mobile)
>

On Thu, Apr 29, 2004 at 05:30:16PM -0400, Pete Geenhuizen wrote:
> Nope, not using RSC, just connected to ttya.
> I can get to the OK prompt doing a telnet to the terminal server and the
> appropriate port.
> Using conserver I can get to the OK prompt before I log in, not after I
> login.

here's what i can tell you.

regardless of what state your solaris box is in, conserver should always
send an IAC/BREAK character sequence over the telnet connection when the
"\z" break sequence is used (which is the default for ^Ecl1). nothing
that i know of will cause that to change. it's then up to the terminal
server telnet daemon to process that sequence and actually do the
hardware break on the serial port.

now, there might be something going on internal to conserver to cause it
to get confused, goof up the IAC/BREAK sequence or something else.

ooops...hit the wrong key and off went half of an email...sorry about
that. here's the rest...

what i'd like to know is, what version are you running (certain versions
have had trouble with the encoding of the data on the wire), and can you
run conserver in debug mode (just one -D is good for this info) and send
it to me? there are particular things i'd like to look at to see if the
telnet state was changing (or perhaps a message from the term server
said it's state changed), etc. maybe even two -D options would be
good...that would show all the data getting written to the file
descriptors too...which could verify if conserver is sending the
IAC/BREAK sequence properly.

another option is to enable a debug mode in telnet and see if anything
happens when you log in (a transistion in some telnet state). i believe
you just crank up telnet, do a 'toggle options' and then 'open ...'.
it'll spew a bunch upon connection and the question is, do you get more
when you actually log in. if so, perhaps conserver isn't handling a
telnet option negotiation properly, and the break interpretation is
getting turned off somehow (whereas a regular telnet does work). this
is the only idea i have come up with so far to explain what you've seen.
and if you try this, it would be nice to see all the entire sequence.

those are my current thoughts. if you can get any of this debugging
info (and version info), that would be cool.

Bryan

That's encouraging, any suggestions on what I can do to try to figure out
what is going wrong?

Do you know what the difference is when Solaris is at the login prompt and
when you are actually logged in?

Pete
--
"Unencumbered by the thought process"
--1992-2000 Click and Clack presidential campaign slogan

Bryan Stansell said:
> On Thu, Apr 29, 2004 at 05:30:16PM -0400, Pete Geenhuizen wrote:
>> Nope, not using RSC, just connected to ttya.
>> I can get to the OK prompt doing a telnet to the terminal server and the
>> appropriate port.
>> Using conserver I can get to the OK prompt before I log in, not after I
>> login.
>
> here's what i can tell you.
>
> regardless of what state your solaris box is in, conserver should always
> send an IAC/BREAK character sequence over the telnet connection when the
> "\z" break sequence is used (which is the default for ^Ecl1). nothing
> that i know of will cause that to change. it's then up to the terminal
> server telnet daemon to process that sequence and actually do the
> hardware break on the serial port.
>
> now, there might be something going on internal to conserver to cause it
> to get confused, goof up the IAC/BREAK sequence or something else.
>

I'm running 8.1.4, and I just thought of something that might have a
baring on this, I compiled it on Solaris 8 and copied the binaries over to
another server running Solaris 7. I then run console on a Solaris 8 box
connecting to the conserver binary and configuration on the Solaris 7
server. I wonder if that might be the problem.

I'll look at that and your other suggestion in the morning. I sure hope
that moving everything over to Solaris 8 server solves the problem.

Either way I'll let you know what happens and send you as much debugging
stuff as I can.

Pete
--
"Unencumbered by the thought process"
--1992-2000 Click and Clack presidential campaign slogan

Bryan Stansell said:
> ooops...hit the wrong key and off went half of an email...sorry about
> that. here's the rest...
>
> what i'd like to know is, what version are you running (certain versions
> have had trouble with the encoding of the data on the wire), and can you
> run conserver in debug mode (just one -D is good for this info) and send
> it to me? there are particular things i'd like to look at to see if the
> telnet state was changing (or perhaps a message from the term server
> said it's state changed), etc. maybe even two -D options would be
> good...that would show all the data getting written to the file
> descriptors too...which could verify if conserver is sending the
> IAC/BREAK sequence properly.
>
> another option is to enable a debug mode in telnet and see if anything
> happens when you log in (a transistion in some telnet state). i believe
> you just crank up telnet, do a 'toggle options' and then 'open ...'.
> it'll spew a bunch upon connection and the question is, do you get more
> when you actually log in. if so, perhaps conserver isn't handling a
> telnet option negotiation properly, and the break interpretation is
> getting turned off somehow (whereas a regular telnet does work). this
> is the only idea i have come up with so far to explain what you've seen.
> and if you try this, it would be nice to see all the entire sequence.
>
> those are my current thoughts. if you can get any of this debugging
> info (and version info), that would be cool.
>
> Bryan
>