Mailing List Archive: problems with console ports

Hello,

I'm a first time user of "conserver". Been using "rtty" for some time
now, but since it's not actively developed anymore, I've decided to try
conserver. I've got it installed fine (I think). I have a couple
hundred Cyclades serial ports connected via PCI cards. I can connect to
a Cisco switch and manage it fine with 'minicom'. When I reboot the
switch, the 'conserver' program properly captures the serial console
output, and dumps it into the logfile (/var/log/console/ttyC48)
properly. ...so it seems to be reading off of the serial line properly.

However, when I try and log in via "console ttyC48", I get no reply back
from the device. I tried "^eco" - and it replies back "[up]". I also
have a couple of Solaris boxes connected via ttyC94 and ttyC95.
'minicom' connects and interacts with them without any problems. However
'console ttyC94' does the same thing, nothing.

I'm running 9600 8n1 via minicom, and set the conserver.cf config to use
9600 baud, even parity. Here are the relavent conserver.cf entries:

ttyC48:/dev/ttyC48:9600e:&:
ttyC94:/dev/ttyC94:9600e:&:
ttyC95:/dev/ttyC95:9600e:&:

When running 'conserver -D -v' - here's the relavent output for ttyC48
(cisco switch):

[Mon May 5 12:40:25 2003] conserver (9266): DEBUG: ReadLine():
returning <ttyC48:/dev/ttyC48:9600e:&:>
[Mon May 5 12:40:25 2003] conserver (9266): INFO: [ttyC48] /dev/ttyC48
(9600e) logged to /var/log/console/ttyC48
[Mon May 5 12:40:25 2003] conserver (9267): DEBUG: DestroyConsent():
destroying `ttyC48'
... snip of duplicates ...
[Mon May 5 12:40:39 2003] conserver (9282): DEBUG: DestroyConsent():
destroying `ttyC48'
[Mon May 5 12:40:42 2003] conserver (9270): INFO: [ttyC48] user
root@goober2 authenticated
[Mon May 5 12:40:42 2003] conserver (9270): [ttyC48] login root@goober2
[Mon May 5 12:40:51 2003] conserver (9270): [ttyC48] logout root@goober2
[Mon May 5 12:40:54 2003] conserver (9266): DEBUG:
DumpDataStructures(): server=ttyC48, dfile=/dev/ttyC48,
lfile=/var/log/console/ttyC48
[Mon May 5 12:40:54 2003] conserver (9266): DEBUG: DestroyConsent():
destroying `ttyC48'
[Mon May 5 12:40:55 2003] conserver (9270): DEBUG:
DumpDataStructures(): server=ttyC48, dfile=/dev/ttyC48,
lfile=/var/log/console/ttyC48
[Mon May 5 12:40:55 2003] conserver (9270): DEBUG: DestroyConsent():
destroying `ttyC48'

I'm running all of this on a RedHat Linux 8.0 platform, stock kernel, on
a Dell 2650 server, and Cyclades Z series terminal server boards.

Any ideas? Thoughts?

v/r
Shane
sgibson@digitalimpact.com

On Fri, May 09, 2003 at 01:26:58PM -0700, Shane Y. Gibson wrote:
> I'm running 9600 8n1 via minicom, and set the conserver.cf config to use
> 9600 baud, even parity. Here are the relavent conserver.cf entries:
>
> ttyC48:/dev/ttyC48:9600e:&:

well, i'm not sure why it would matter *that* much, but it's serial
weirdness so it's worth a shot. if you change '9600e' to '9600n',
you'll get 8n1 with conserver as well. it *might* help.

also, when you connect with the console client and try and interact
with the device, do you get a 'console is read-only' type message? if
so, it's a permissions problem on /dev/ttyC48 and i supposed it's
always possible that minicom is just overriding it. unlikely, but
perhaps there's something here.

otherwise, i'm not sure. i seem to remember others having difficulty
with the cyclades z boards. or maybe it was another type of board. if
there is any sort of flow control settings you can adjust, maybe that
would help. or if you can view the serial settings with stty when
minicom and conserver are connected and compare them, maybe we'll see
an extra setting somewhere that would make a difference.

i'm kinda shooting in the dark, but maybe the muzzle blast light up
something interesting.

Bryan