[news.software.b] Cnews history problem

paine@fungus.dec.com (Willy Paine) (02/20/91)

hi..
 
System:  Interactive 2.2.1
ulimit:  10000
Cnews:   version 90Sept
 
problem:
 
relaynews: error writting `/usr/lib/news/history', probably the disk filled
(File too large)
 
relaynews: error closing history file.
 
 
I have 2215443 bytes in history file.  I have 25 megB free in root
file system and 51 megB free in /usr file system.
 
It works fine if I delete history file but it can grow back in three days
to get same error message.
 
Please advice me and thanks on advanced...
 
It would be really nice to have FAQ file for cnews.
 
Thanks again.
 
willy
 
uunet!fungus.enet.dec.com!paine
uunet!fungus.zso.dec.com!paine  or
seaeast.wa.com!willyp
fax  1 206 827-4099
 
 
 

art@pilikia.pegasus.com (Art Neilson) (03/01/91)

I recently gave all the latest Cnews patches to a local site and
they promptly put in all the patches and rebuilt Cnews.  Now their
history files are monstrous.  The guy who applied the patches and
rebuilt Cnews said his history.pag was on the order of 400~ MB,
larger than the drive itself ... His system is running SCO UNIX.
I recall seeing articles on similar problems recently which were
easily fixed once one knew what to do.  Can anyone give me a clue
as to help this guy ?  BTW he is using dbz.
-- 
Arthur W. Neilson III		| INET: art@pilikia.pegasus.com
Bank of Hawaii Tech Support	| UUCP: uunet!ucsd!nosc!pilikia!art

henry@zoo.toronto.edu (Henry Spencer) (03/02/91)

In article <1991Mar01.054939.25486@pilikia.pegasus.com> art@pilikia.pegasus.com (Art Neilson) writes:
>I recall seeing articles on similar problems recently which were
>easily fixed once one knew what to do.  Can anyone give me a clue
>as to help this guy ?  BTW he is using dbz.

Well, the *real* advice is that everyone should at least skim
notebook/problems when they run into trouble right off the bat.
That's where we put all the portability advice we can think of.

Tell him to recompile dbz without -O.
-- 
"But this *is* the simplified version   | Henry Spencer @ U of Toronto Zoology
for the general public."     -S. Harris |  henry@zoo.toronto.edu  utzoo!henry

ronald@robobar.co.uk (Ronald S H Khoo) (03/03/91)

art@pilikia.pegasus.com (Art Neilson) writes:

> history files are monstrous
> His system is running SCO UNIX.
> he is using dbz.

And the answer, as stated in notebook/problems, is to compile dbz without -O.

Henry, this non-reading of notebook/problems is a bit of a nuisance
isn't it ? Any chance of putting a bigger pointer in ?
The other thing that catches this is running the regression tests.
What about making that the default, at least in dbz/ ?
-- 
Ronald Khoo <ronald@robobar.co.uk> +44 81 991 1142 (O) +44 71 229 7741 (H)

nreadwin@micrognosis.co.uk (Neil Readwin) (03/04/91)

In article <1991Mar3.010217.8750@robobar.co.uk> ronald@robobar.co.uk (Ronald S H Khoo) writes:
>Henry, this non-reading of notebook/problems is a bit of a nuisance
>isn't it ? Any chance of putting a bigger pointer in ?

Something a bit bigger than

|This directory is the C News Implementor's Notebook, containing various
|internal documents describing what goes on inside.

My first reaction when I read that was that since I wasn't very interested
in the internals of C News there was no reason to look at any of the stuff
in notebook/
-- 
 Disclaimer: I do not speak for Micrognosis, CDC or CSK.
 Phone: +44 71 528 8282  E-mail: nreadwin@micrognosis.co.uk
 You will know me by my river, you will know me by the Clyde ...

henry@zoo.toronto.edu (Henry Spencer) (03/05/91)

In article <1991Mar3.010217.8750@robobar.co.uk> ronald@robobar.co.uk (Ronald S H Khoo) writes:
>Henry, this non-reading of notebook/problems is a bit of a nuisance
>isn't it ? Any chance of putting a bigger pointer in ?

There's already a pointer specifically to n/p in the top-level README file.
What is really needed is a "Troubleshooting Guide" document, both as
something that would be very useful in general and as the best place to
put a pointer saying "look at n/p if it doesn't work after installation".

>The other thing that catches this is running the regression tests.
>What about making that the default, at least in dbz/ ?

I'm thinking about this.  It may happen as part of makefile revisions.
One problem is that some of the regression tests -- notably the one for
dbz, unfortunately! -- can take quite a while to run.
-- 
"But this *is* the simplified version   | Henry Spencer @ U of Toronto Zoology
for the general public."     -S. Harris |  henry@zoo.toronto.edu  utzoo!henry