Profile

unixronin: Galen the technomage, from Babylon 5: Crusade (Default)
Unixronin

December 2012

S M T W T F S
      1
2345678
9101112131415
16171819202122
23242526272829
3031     

Most Popular Tags

Expand Cut Tags

No cut tags
Wednesday, February 18th, 2004 12:04 pm

Don't you just hate it when you've been trying to do something fairly simple that just won't work, and you can't figure out why, and then you realize the problem and want to kick yourself?

I've been trying to get ppp working on a Sparcstation LX running OpenBSD 3.3, with the eventual intention of making it a dial-on-demand PPP/firewall/NAT box.  I'd tried it before on 3.2, but the LX has slightly funky serial ports and OpenBSD versions before 3.3 don't talk to them properly.  Nevertheless, on 3.3, I still couldn't talk to the modem.  The connection would go into limbo as soon as I invoked term.

Then suddenly I figured it out, with a resounding DOH!!!!  The LX didn't have a keyboard connected, so OpenBSD was automatically helpfully redirecting all input and output to the first available terminal.  Yeah, that's right, ttya.  The port with the modem connected to it.  So of COURSE the LX couldn't talk to the modem, it was already using that port as its serial console....

  1. Carefully remove clue bat from packaging.
  2. Apply generously to head and shoulders.
  3. Repeat as necessary.
Wednesday, February 18th, 2004 12:13 pm (UTC)
...fuser and lsof are your two best friends.



Wednesday, February 18th, 2004 04:41 pm (UTC)
Very true, but until I had that sudden "DOH!" moment, I thought it was either a driver issue or a hardware problem. It was only as I rebooted the machine that I happened to catch the "Using tty for input and output" message, and promptly smacked myself upside the head.

"And thus, the student was enlightened."
Wednesday, February 18th, 2004 03:31 pm (UTC)
Twenty years ago I spent hours trying to figure out why a line in a dBASE II program was never being executed (yes, I admit I wrote in dBASE II). It was something like this:

REM Here we test for something
if blah blah blah
REM and if it is we invoke this;
foo(bar)
endif

After more time than I care to admit I remembered that the semicolon was dBASE II's line-continuation character; it was reading "foo(bar)" as part of the preceding comment. Urrgh.
Wednesday, February 18th, 2004 04:56 pm (UTC)
Unnng. Yeah, I can see how that would be a pound-head-on-wall moment.