[news.admin] Trn help needed

abrams@cass.amd.bnl.gov (The Ancient Programmer) (02/09/91)

	I have built trn on my Sun3 (OS4.1) and it reads news beautifully.
However, I have encountered one problem, and have been unable to locate
the cause of it. 
	My machines name is dan.amd.bnl.gov, but when I post an
article that machine the From: line says "From abrams@dan.tmc.edu".

	Grepping the trn source for "dan.amd" produces:
Pnews:		'') sitename="dan.amd" ;;
Pnews:		sitename="dan.amd"
Rnmail:		 '') sitename="dan.amd" ;;
Rnmail:		undef) sitename="dan.amd" ;;
config.h:	#define SITENAME "dan.amd"
config.sh:	sitename='dan.amd'

	and a grep for "bnl.gov":
Pnews:          sitename=${sitename}.bnl.gov
Rnmail:         sitename=${sitename}.bnl.gov
config.h:	#define OURDOMAIN "bnl.gov"
config.sh:	domain='bnl.gov'
And in NNTP/common/conf.h:	#define   DOMAIN  "bnl.gov"

I've defined HOSTFILE in config.h as dan.amd.bnl.gov
#define HOSTFILE "/usr/local/lib/usenetname" 

I then thought that I could find tmc.edu in either trn's or nntp's source.
All that I turned up were comments!
grep tmc.edu Nntp/*/*
Nntp/common/README.SYSV:	from archi...
Nntp/common/README.SYSV:	If you hav...
Nntp/doc/nntpd.dst:	Stan Barber (In...
Nntp/inews/inews.c:	 * Send bug rep...
Nntp/server/main.c:	 *   (Internet:	...
Nntp/server/misc.c:	 * by Stan Barb...
Nntp/server/misc.c:	 * written by S...
Nntp/support/nntp_awk:	# Stan Barbe...
Nntp/xfer/nntpxfer.c:	 * Stan Barbe...
#dan<su 11> grep tmc.edu Trn/*
Trn/trn.1:	and is now under the...
Trn/trn.1.02:	and is now under ...

Question:  Where is the tmc.edu coming from in my From: line?

I had to login on this machine in order to post an article with the
correct return address.

Any clues as to what to do or what to look for would be greatly appreciated. 

			Thanks in advance.

abrams@dan.amd.bnl.gov (The Ancient Programmer) (02/13/91)

	Thanks to a clue from Don "Truck" Lewis of  Harris Semiconductor, 
I was lead to the solution to the problem.

> Check the definition of DOMAIN in Nntp/common/conf.h.  Be sure that it
> was correct when you built inews.
> -- 
> Don "Truck" Lewis                      Harris Semiconductor

Domain was defined correctly, however a close
reading of the docs concerning DOMAIN showed me what the solutions was.
By changing my hostname.le0 file to contain my fully qualified name, and
undefining DOMAIN in common/conf.h, the From: line was corrected.

I still do not know how tmc.edu got into my From: line.

	What the hell, it's working.



--
INTERNET:  abrams@bnl.gov               |/ |    ^
BITNET:    abrams@bnlux0.BITNET         |\ |__ /-\

abrams@dan.ccd.bnl.gov (The Ancient Programmer) (04/30/91)

	I am running remote TRN on a Sun4 OS4.1.1. My NNTP server is an 
DecStation 5000. Everything was working well until this weekend when mthreads 
suddenly went berserk and began filling the mt.log file with thousands of 
messages:
Apr 28 01:35 ** interrupt 11 **
		:
Apr 29 12:43 ** interrupt 11 **
Apr 29 12:43 ** interrupt 11 **
Apr 29 12:43 ** interrupt 11 **
Apr 29 12:43 ** interrupt 11 **

	This fills the partition and causes the system to go around the bend.
Normally, mthreads is run from cron, so I tried running it manually and got
the same results.  It started to print its regular output and then went wild.
----------------------------------------------------------------------------
#dan<1>mthreads -v
X..................................................#####..######.#####..#############.##..#..#########################interrupt 11!
interrupt 11!                         ^- it always happens about here. Is there
interrupt 11!                            a way to tell on  which group it is 
	:                                working?
------------------------------------------------------------------------------

In /usr/include/signal.h, SIGSEGV is defined to be 11 (segmentation violation)
but I'm not at all sure that mthreads' interrupt 11 is the same.

Has anyone any idea what's happening to my mthreads?


			thanx in advance.
--
INTERNET:  abrams@bnl.gov               |/ |    ^
BITNET:    abrams@bnlux0.BITNET         |\ |__ /-\

technews@iitmax.iit.edu (Kevin Kadow) (05/01/91)

I'm a user here (we run RN on an ENCORE unix) and was wondering what would
be involved in upgrading from RN to TRN?

(in other words, could I pester the sysadmin to put in the upgrade or would
it be too expensive in terms of fees/downtime/overall hassle ??

-- 

technews@iitmax.iit.edu                           kadokev@iitvax (bitnet)
                         My Employer Disagrees.

gerry@jts.com (G. Roderick Singleton ) (05/01/91)

In article <1991Apr29.204515.11126@bnlux1.bnl.gov> abrams@dan.ccd.bnl.gov (The Ancient Programmer) writes:
>
>	I am running remote TRN on a Sun4 OS4.1.1. My NNTP server is an 
>DecStation 5000. Everything was working well until this weekend when mthreads 
>suddenly went berserk and began filling the mt.log file with thousands of 
>messages:
>Apr 28 01:35 ** interrupt 11 **
>		:
	[stuff deleted]
>

You forgot to mention the Revision and patchlevel of your TRN.

I ask 'cause I saw similar behaviour with an early version of mthreads
that I used with standalone TRN.  The current version (Trn v1.0.2)
corrects this problem at my site.

Uunet, and possibly other archive sites, has the current version
so you should be able to grab a copy easily.  I am unaware of any
further patches but stand to be corrected.

Cheers,
ger
-- 
G. Roderick Singleton, System and Network Manager, JTS Computers 
{yunexus | uunet | geac | torsqnt}!gerry@jtsv16.jts.com

abrams@dan.ccd.bnl.gov (The Ancient Programmer) (05/04/91)

	I was running mthreads (trn 1.0.1), when it began be run amock,
filling up mt.log. I have since upgraded to trn v1.0.2, and mthreads now 
dumps core or produces the output shown below.
----------------------------------------------------------------------------
mthreads -z -vvvv
X...................#..#..#....#.#.............E!eeee###eeee####e######eeee
e####e#eee#eee##eeeeeeeeeeeeeee#..eeeeeeeeeeeeeeeeeeeee#eeeeeeeeeeeeeeee#ee
eeeeeeee#eeee#eeeee#eeeeeeee#eeeeeeeeeeeeeeeeeeeeeeeeee#eee#e##e#ee#ee##eee
eeeeeeeeeee########eeeeeeeeeeeeeeeeeeeeeeeeee#eeeeeeeeeeeeeeeeeeeeeeeeeeeee
eeeeeeeeeeee#e#eeeeeeeeeeeeeeeeeeeeeeee#eee#eeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
eeeeeeeeeeee#eeeeeeeeeeeeeeeeeeeeeeeeeeeee#eeeeeeeeeeeeeeeeeeeeee#eeeeeeeee
eeeeeeeeeeeeeeeeeeeee##eeeeeeee#eeeeeeeeeeeeeeeeeeeeee#eeeee#eeeee#eeeeeeee
eeeeeeeeeeeee#e#eeee#eeeeeee#eeeee#eee#eee#e#eeeeee#ee##dan<44>
                                                        ^^^^^^
Note the displaced prompt.

	Mt.log is filled with messages such as:
May  3 11:35 comp.graphics: Invalid author in data file: -18648 [0]
			:
                        :
May  3 13:28 comp.sys.atari.st: Missing reference line!  [36361]
	(many, MANY of these)
			:
May  3 13:29 misc.consumers: Reference ended in '*'.


	and so on and on .....

	Can anyone tell me what is happening here? Is Trn screwing up
or has the news gone bad?  I'm running remote trn on a Sun4 (OS 4.1.1)
using a Decstation 5000 as the news server.



	Thanx in advance,

--
INTERNET:  abrams@bnl.gov               |/ |    ^
BITNET:    abrams@bnlux0.BITNET         |\ |__ /-\

bill@bilver.uucp (Bill Vermillion) (05/06/91)

In article <1991May3.195034.9911@bnlux1.bnl.gov> abrams@dan.ccd.bnl.gov (The Ancient Programmer) writes:
 
>	Can anyone tell me what is happening here? Is Trn screwing up
>or has the news gone bad?  I'm running remote trn on a Sun4 (OS 4.1.1)
>using a Decstation 5000 as the news server.

Well I would suspect that some news article/directory is bad.

I had a similar occurance.  Mthreads would get to one point, and
stop and leave the rest unprocessed.

I ran mthreads with about 3 or 4 v's on the option line.  Each
extra v forces an extra level, and more output in the log.

When I found where the output of the log stopped, I checked that
group and found no problems.  So I looked at active 2 to find out
which newsgroup was to be processed next after the last one that was
processed.  When I went to that directory I found problems.

One I found was that instead of a file .threads I had a directory
called .threads.    I found one or two other corrupt portions
there.  Rm'd the offending files and directories and all was fine.

This may not be your problem, but it sounds similar.


-- 
Bill Vermillion - UUCP: uunet!tarpit!bilver!bill
                      : bill@bilver.UUCP