netra x1 serial console multi user login

[Hoyte Swager]

hi there
on my netra x1 configure on the boot prompt to use seriala for in/out
is can succesfully login in single user mode and acces the machine
but if i switch to multiuser mode (exit/ctrl-d) apparently the tty settings
change in such a way that a lf or cr gives a new login prompt but
trying to login with a username (eg.root) gives an emptyline and the
login prompt again.

i tried every possible option in /etc/ttydefs and /etc/inittab and
basically am stuck getting this machine accessed after single user mode

any suggestions?
regards
hoyte swager
[email]news@hoyte.nl[/email]

[Darren Dunham]

Hoyte Swager <h.swager@chello.nl> wrote:

> hi there
> on my netra x1 configure on the boot prompt to use seriala for in/out

How did you do that? Normally seriala will be the console, and there's
already a login on the console. You don't want to add a second one via
ttym

> is can succesfully login in single user mode and acces the machine
> but if i switch to multiuser mode (exit/ctrl-d) apparently the tty settings
> change in such a way that a lf or cr gives a new login prompt but
> trying to login with a username (eg.root) gives an emptyline and the
> login prompt again.

> i tried every possible option in /etc/ttydefs and /etc/inittab and
> basically am stuck getting this machine accessed after single user mode

> any suggestions?

What is the output from the following commands after you reach
multiuser? (I'm assuming you can still log in on the network?)

pmadm -l
ps -ef | grep console

--
Darren Dunham [email]ddunham@taos.com[/email]
Unix System Administrator Taos - The SysAdmin Company
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >

[Hoyte Swager]

well, basically it is the "out of the box" solaris config
the only installations after that where fw1 and ace which
did not seem to have any influence on the x1 serial config.
pmadm -l gives a zsmon/ttymon on ttyb
(i disabled the one monitoring ttya to check if that would make any
difference).
the terminal settings in the command line options have tvi925 at the end
(?is that off?)
(zsmon ttymon ttyb u root /dev/term/b I - /usr/bin/login - 9600
ldterm,ttcompat ttyb login: - tvi925 y #)

# grep ttymon /etc/inittab
co:234:respawn:/usr/lib/saf/ttymon -g -h -p "`uname -n` console login: " -T
vt100 -d /dev/console -l console -m ldterm,ttcompat

# grep '^console:" /etc/ttydefs
console:9600 opost onlcr:9600::console

what makes it switch behaviour after single user mode?
not the settings in /etc/ttydefs because i tried those...
any help appreciated
regards
hoyte


"Darren Dunham" <ddunham@redwood.taos.com> wrote in message
news:KyTWa.257$_g3.41@newssvr24.news.prodigy.com.. .

> Hoyte Swager <h.swager@chello.nl> wrote:

> > hi there
> > on my netra x1 configure on the boot prompt to use seriala for in/out

>
> How did you do that? Normally seriala will be the console, and there's
> already a login on the console. You don't want to add a second one via
> ttym
>
>

> > is can succesfully login in single user mode and acces the machine
> > but if i switch to multiuser mode (exit/ctrl-d) apparently the tty

settings

> > change in such a way that a lf or cr gives a new login prompt but
> > trying to login with a username (eg.root) gives an emptyline and the
> > login prompt again.

>

> > i tried every possible option in /etc/ttydefs and /etc/inittab and
> > basically am stuck getting this machine accessed after single user mode

>

> > any suggestions?

>
> What is the output from the following commands after you reach
> multiuser? (I'm assuming you can still log in on the network?)
>
> pmadm -l
> ps -ef | grep console
>
> --
> Darren Dunham [email]ddunham@taos.com[/email]
> Unix System Administrator Taos - The SysAdmin Company
> Got some Dr Pepper? San Francisco, CA bay area
> < This line left intentionally blank to confuse you. >

[Greg Andrews]

"Hoyte Swager" <h.swager@chello.nl> writes:

>
>what makes it switch behaviour after single user mode?
>

The usual problem is that init doesn't think the machine has
finished booting into multi-user mode (actually run level 3).
Init waits for the /sbin/rc2 and /sbin/rc3 scripts to finish
before running the console ttymon.

This can happen when one of the boot scripts in the /etc/rc2.d
or /etc/rc3.d has hung rather then finished.

Telnet into the machine and see if the /sbin/rc2 or /sbin/rc3
script is still running. If it is, find out which script has
hung by runing ptree on the numeric process id of the rc2 or
rc3 script.

I.e., if the rc2 script is running and has a pid of 105, run
the command "ptree 105". The output will tell you which script
hasn't finished yet, and even which command in that script has
hung.

-Greg
--
Do NOT reply via e-mail.
Reply in the newsgroup.

[Darren Dunham]

Hoyte Swager <h.swager@chello.nl> wrote:

> well, basically it is the "out of the box" solaris config
> the only installations after that where fw1 and ace which
> did not seem to have any influence on the x1 serial config.
> pmadm -l gives a zsmon/ttymon on ttyb
> (i disabled the one monitoring ttya to check if that would make any
> difference).

Normally they're both inactive anyway. I just wanted the output to
verify.

> the terminal settings in the command line options have tvi925 at the end
> (?is that off?)

For pmadm output, it's probably irrelvant.

> (zsmon ttymon ttyb u root /dev/term/b I - /usr/bin/login - 9600
> ldterm,ttcompat ttyb login: - tvi925 y #)

Right. The 'I' after the /dev/term/b means 'Initialize only'. There's
not actually a service here, which is normal. If you hadn't disabled
it, I would have expected a similar line for /dev/term/a, which would
still be okay. Likewise, sinced it's not running, the tvi925 isn't
really used by anything.

> # grep ttymon /etc/inittab
> co:234:respawn:/usr/lib/saf/ttymon -g -h -p "`uname -n` console login: " -T
> vt100 -d /dev/console -l console -m ldterm,ttcompat

Looks okay.

> # grep '^console:" /etc/ttydefs
> console:9600 opost onlcr:9600::console

I have a 'hupcl' option also, but that shouldn't cause this behavior.

> what makes it switch behaviour after single user mode?

My assumption is that something else running since single-user started
running and grabbed the port.

Anything else with the device open? I had hoped you would run a 'ps'
and see if anything has 'console' as a controlling tty.

ps -ef | grep console
fuser /dev/console

--
Darren Dunham [email]ddunham@taos.com[/email]
Unix System Administrator Taos - The SysAdmin Company
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >

[Hoyte Swager]

# ps -ef | grep console
root 302 1 0 11:07:45 console 0:00 /usr/lib/saf/ttymon -g -h -p x1 console
login: -T vt100 -d /dev/console -l con
(chopped of at -l con)

# fuser /dev/console
/dev/console: 302oy

so no other processes accessing the console port

if i compare stty -a in both singleuser mode and
multiuser mode the only difference is hupcl but changing
that in ttydefs make no difference

if checked the voltages on the serial cable (i am becoming desperate ;-)
and only noticed that the terminal has no level on dtr but joining dtr with
dsr
did not help either.

what is happening here?
why is everything ok in singleuser mode and not in multiuser mode?
why does a LF or CR when given without a username correctly gives back
the prompt again and if i type a username and LF or CR it first gives back
an empty line and then the prompt again?

any help appreciated
getting more and more frustrated....
(all this time spent just to get the serial console keeping working
properly...;-(
hoyte swager

[Darren Dunham]

Hoyte Swager <h.swager@chello.nl> wrote:

> if i then do the same with the ttymon
> /usr/lib/saf/ttymon -g -d /dev/console -l console -T vt100
> i get the login prompt, but as soon as i enter the username
> and hit return is says in the serial console:

> Warning -- ttymon cannot allocate controlling tty on "/dev/console"
> there may be another session active on this port

> i am quite sure there is no other process having a lock on it
> (checked with ps/fuser/ptree and lsof |grep devices.pseudo.cn)

I'd start 'truss'ing the ttymon process and see what it does. Do you
have another functioning machine to compare the truss output against?

--
Darren Dunham [email]ddunham@taos.com[/email]
Unix System Administrator Taos - The SysAdmin Company
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >

[Darren Dunham]

Hoyte Swager <h.swager@chello.nl> wrote:

> well, after doing that and comparing output it seemed to look similar.
> i ended up copying sulogin to login which gave a functioning multiuser login
> again
> and after that i got fresh copies of login and ttymon and that solved the
> problem.
> what corrupted the login exec remains a mystery...

If login has been corrupted/modified, I tend to think "rootkit" or
"breakin" first. Better run a rootkit checker or do some binary
analysis.

--
Darren Dunham [email]ddunham@taos.com[/email]
Unix System Administrator Taos - The SysAdmin Company
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >