[comp.unix.xenix] moby bignum tty accesses in /usr/adm/wtmp

mc68020@gilsys.UUCP (Thomas J Keller) (10/12/87)

Environment:   Tandy 6000, XENIX 3.1.2 (Tandy number, MIcroSlush # 3.0),
	       user accounting turned on, disk write verification on.
	       One phoneline, 300/1200 Hayes modem, one Wyse 75 terminal
	       on /dev/tty05, 12Mbyte primary, 1 10Mbyte secondary, 1
	       15Mbyte secondaryhard disks, 2 8" DSDD floppies, 1Mbyte RAM,
	       1 three-port serial expansion card.

Problem:  /usr/adm/wtpm growing at alarming rate (>500K/day).  Manual
	  checking revealed a *LARGE* number (on the order of 9000) accesses
	  to /dev/tty02 beginning at 03:00 and continuing through 04:29,
	  at the rate of approximately 95-120 accesses per MINUTE! (rate
	  is somewhat variable, depending on what else is happening on the
	  system.  In particular, the rate drops below 100 after 04:00, as
	  that is when I unbatch received news for the day)  Checking
	  the accounting data reveals no useful information,  as there are no
	  indiciations of any related processes at the times these accesses
	  are being logged by wtmp.  

Query:  what in the name of GHU is going on!????  Could the problem be in the
	Hayes modem?  If so, why would it start and stop the way it does?

	Any assistance greatly appreciated.


-- 
Tom Keller 
VOICE  : + 1 707 575 9493
UUCP   : {ihnp4,ames,sun,amdahl,lll-crg,pyramid}!ptsfa!gilsys!mc68020

wolfgang@mgm.mit.edu (Wolfgang Rupprecht) (10/15/87)

In article <1134@gilsys.UUCP> mc68020@gilsys.UUCP (Thomas J Keller) writes:
>Problem:  /usr/adm/wtpm growing at alarming rate (>500K/day).  Manual
>	  checking revealed a *LARGE* number (on the order of 9000) accesses
>	  to /dev/tty02 beginning at 03:00 and continuing through 04:29,
>	  at the rate of approximately 95-120 accesses per MINUTE! (rate
>	  is somewhat variable, depending on what else is happening on the
>	  system.  In particular, the rate drops below 100 after 04:00, as
>	  that is when I unbatch received news for the day)  Checking
>	  the accounting data reveals no useful information,  as there are no
>	  indiciations of any related processes at the times these accesses
>	  are being logged by wtmp.  

You didn't mentions if /dev/tty02 was a dial-in dial-out or both.  I
assume that it is the modem line. I can think of two things that might
do this.  1) both getty and some other program are both sucking down
chars from tty02, ie. if you try to dial out on a line that still has
getty enabled. Neither one would do too well here. 2) if your modem is
echoing stuff back to getty (which echoes something to the modem ....
<repeat forever>).  In either case login wouldn't be too happy with
what it got.

Why not solve the mystery and put a terminal on the line and watch 
the chars fly? (just connect up the terminal's rs232 signal-ground 
and recieve lines to the tty as a monitor.)
--
Wolfgang Rupprecht	       UUCP: mirror!mit-mgm!wolfgang
<add us to your host table>    ARPA: wolfgang@mgm.mit.edu (IP addr 18.82.0.114)

jfh@killer.UUCP (The Beach Bum) (10/16/87)

In article <1616@bloom-beacon.MIT.EDU>, wolfgang@mgm.mit.edu (Wolfgang Rupprecht) writes:
> 
> In article <1134@gilsys.UUCP> mc68020@gilsys.UUCP (Thomas J Keller) writes:
> >Problem:  /usr/adm/wtpm growing at alarming rate (>500K/day).  Manual
> >	  checking revealed a *LARGE* number (on the order of 9000) accesses
> >	  to /dev/tty02 beginning at 03:00 and continuing through 04:29,
> >	  at the rate of approximately 95-120 accesses per MINUTE!
> 
> Why not solve the mystery and put a terminal on the line and watch 
> the chars fly? (just connect up the terminal's rs232 signal-ground 
> and recieve lines to the tty as a monitor.)

A problem I saw once on an unconnected line was that character transmitted
by the system were being recieved back, possibly changed alittle ;-), even
though nothing was one the line.  Hooking up a terminal made the problem go
away - BUT - the guy on the other end of the wire had an AB switch and didn't
like having to use only one system.

The Fix?  I added a 10K resistor between RxD and ground on the System end of
the cable.  Didn't do anything while he was using the line, and when he
wasn't, the resistor pulled the line down to ground rather than letting it
float.

The Moral? Engineers (he was an engineer) have their heads up their butts
too much of the time ... (he also happened to have wired the systems ;-)

- John.
-- 
John F. Haugh II		HECI Exploration Co. Inc.
UUCP:	...!ihnp4!killer!jfh	11910 Greenville Ave, Suite 600
"Don't Have an Oil Well?"	Dallas, TX. 75243
" ... Then Buy One!"		(214) 231-0993

rob@array.UUCP (Rob Marchand) (10/16/87)

In article <1134@gilsys.UUCP> mc68020@gilsys.UUCP (Thomas J Keller) writes:
 
>Problem:  /usr/adm/wtpm growing at alarming rate (>500K/day).  Manual
>	  checking revealed a *LARGE* number (on the order of 9000) accesses
>	  to /dev/tty02 beginning at 03:00 and continuing through 04:29,
>	  at the rate of approximately 95-120 accesses per MINUTE! (rate
>	  is somewhat variable, depending on what else is happening on the

	   It is likely caused by getty processes being spawned and 
	   quickly dying when no carrier signal is detected.  I've seen
	   this on our system when the cable was unplugged from the modem
	   (also a hayes), which provides a Carrier Detect signal whether
	   or not a carrier is present.  Same symptoms; wtmp grows like
	   hell, and performance drops like hell (although its no prize
	   here anyways :-)

>	  system.  In particular, the rate drops below 100 after 04:00, as
>	  that is when I unbatch received news for the day)  Checking
	
	   This probably is just a result of the load that unbatching
	   places on the system;  I would guess (I've stuck my foot
	   in it now, why quit) that this is what is slowing the
	   respawning of the gettys down (maybe because of news
	   decompression?)
>
>Query:  what in the name of GHU is going on!????  Could the problem be in the
>	Hayes modem?  If so, why would it start and stop the way it does?
>
	    Two things; check the dip switch on the modem to ensure that
	    it provides a carrier detect (or get a getty program that
	    has some smarts; uugetty seems to work for us), and check
	    to see that the serial line provides the CD signal....
>
>
>-- 
>Tom Keller 

		Good Luck
		  Rob M.
		  rob@array.UUCP
-- 
Rob Marchand                   UUCP: {mnetor,utzoo}!lsuc!array!rob
Array Systems Computing        ARPA: rob%array.UUCP@uunet.UU.NET
200-5000 Dufferin Street       Phone : +1(416)736-0900   Fax: (416) 736-4715
Downsview, Ont CANADA M3H 5T5  Telex : 063666 (CNCP EOS TOR) .TO 21:ARY001