home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!elroy.jpl.nasa.gov!usc!news!benton.prepress.com!crash!gpitcher
- From: gpitcher@crash.cts.com (Glenn Pitcher)
- Newsgroups: comp.unix.sysv386
- Subject: Getty problem with Interactivee 3.0?
- Keywords: getty init interactive
- Message-ID: <1992Aug31.171421.21396@crash>
- Date: 1 Sep 92 00:14:21 GMT
- Lines: 61
-
- I just finished fighting my system for most of the day and it's working now,
- but I sure would like to know what happened...
-
- First off, here's our configuration...
-
- Dell 433SE running Interactive SVR3.2
- 64MB memory
- (3) 1.4GB drives (3.8BG formatted)
- Exabyte 8200 8mm drive
- Equinox Megaplex-96 board with 4 cluster multiplexors.
-
- What seems to have happened is init wouldn't spawn *any* gettys. Not even on
- the console. Everything else would work fine except for the gettys.
- My printers even were printing jobs that were left over from the weekend due
- to no paper and these are hooked to the same muxes as the terminals!
- I managed to get the system up into init 1 and checked everything. I checked
- /etc/inittab, I checked getty and init, I checked /etc/gettydefs and almost
- everything else I can think of. Everything checked out fine.
-
- We did, as a last resort, check our EISA configuration. The only problem we
- found was that the 'fast video' setting was 'on' and not 'cached'. I changed
- it to 'cached', reset the system and behold! the system came up. I went back
- into the configuration and found out that I really hadn't changed this EISA
- after all! I did go ahead and make sure it was changed it but it made no
- difference... the system still came up fine.
-
- Here's a couple of interesting tidbits that may help...
-
- 1) Not only could I not get getty running on my dumb terminals hooked to the
- Equinox board, I couldn't get any running on my virtual terminals on the
- PC console (/dev/vt01 thru /dev/vt08)
- 2) When the system was coming up, I would receive normal output from jobs
- on /dev/console (/dev/console is the same as /dev/ttyaA on our system and
- is hooked up to the Equinox board). Even when fsck would ask questions,
- I could respond. Only when getty should have taken over would I have
- problems.
- 3) Like I mentioned above, I had one printer still producing output even
- though I was connected to the Equinox board so I think I can pretty safely
- assume that this is not the problem.
- 4) Once I got the system into init 1, we messed around with the inittab file;
- trying to get one bank of terminals to come up in init 1, along with my
- virtual terminals. After I did 'init q', nothing happened... but when I
- issued 'who -a', I found a bunch of dead LOGIN and getty processes for
- these terminals. This output told me they all ended with SIGTERM and a
- rc of 0. Hmmm.
- 5) The system was fine when I was last on yesterday at 1400. I recycle the
- system at 0145 and all the output was fine except for a couple items.
- 1) There was no login prompt after this init 6 so that means that this
- problem happened sometime between 1400 on Sunday and 0145 on Monday.
- 2) The initdefault entry in the /etc/inittab file is set to 3 because I
- run a SLIP line locally. Saturday morning when the system was recycled,
- I'm sure it came up fine but the console output from the last init 6
- did not show the jobs which are run when it goes into init 3. It only
- showed the output that would be produced with init 2. And yes, I did make
- 100% sure my inittab file was ok :-)
-
- I don't know... The only time I've seen something like this was when I would
- do something stupid like wipe out the inittab file but I know I didn't
- do anything this time!! Honest! :-)
-
- Does anyone have any ideas?
-