all articles expire and no space left!

[Russ Allbery]

Edwin <edwin.yau at rdel.co.uk> writes:

> 1) Yesterday something strange happened.  During the day all articles in
> all newsgroups self-expired. (i.e. everything vanished to be replaced
> with a similar message of:

> Error!

> newsgroup server responded: Bad article number

It sounds to me like either history or overview is having problems, and
your additional information below seems to confirm that.

> 2)  My daily news server report says the following:

> Expire messages:
> expire begin Fri Mar 31 03:01:30 BST 2000: (-v1 -c -e
> -z/var/news/log/expire.rm
> -u/var/news/log/expire.index.rm -Z/var/news/log/expire.lowmark)
>     Can't write new history, No space left on device
>     Can't close /var/news/db/history.n, No space left on device
>     Can't close -Z file, No space left on device
>     Expire errors: history files not updated.

You ran out of disk space in the partition where your history file is
stored.  Based on the above symptoms, it sounds like it may have also been
corrupted at the same time.  You may need to run makehistory to rebuild
it.

> I have 6 Gb partition set up where all of news goes into (/var/news)
> including INN and the buffers etc.  I've set up 5 Gb of buffers (3 * 1.5
> Gb and 1 * 0.5 Gb give 5 Gb in total).  /var/news/spool takes up 676 MB.
> I am left with approx 100k of free space.

If your history file in /var/news/db is larger than 100KB, that's not
enough free space.  You need free space equal (roughly) to the size of the
history file and its indexes to successfully do the nightly expire.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>