Path: utzoo!utgpu!watserv1!watmath!att!linac!pacific.mps.ohio-state.edu!zaphod.mps.ohio-state.edu!caen!stealth
From: stealth@caen.engin.umich.edu (Mike Pelletier)
Newsgroups: news.software.b
Subject: Re: More about C News and barfing.
Message-ID: <1991Feb8.210554.22633@engin.umich.edu>
Date: 8 Feb 91 21:05:54 GMT
References: <1991Feb7.023059.3082@Sandelman.OCUnix.On.Ca>
Sender: news@engin.umich.edu (CAEN Netnews)
Organization: University of Michigan Engineering, Ann Arbor
Lines: 44

In article <1991Feb7.023059.3082@Sandelman.OCUnix.On.Ca>
	mcr@Sandelman.OCUnix.On.Ca (Michael Richardson) writes:
>
>  Recently I read something here to the effect that a (new) C news
>site was producing a large number of barfs on a adjacent B news site?
>  Henry responded to the effect that this can't be as C news as bashed
>quite extensively against B news.
>
>  Well, one of my downfeeds seems to have the same problem -- to
>quote:
>
>To: latour!mcr
>Subject: Bad news batches
>Date: Wed, 6 Feb 91 17:37:29 EST
>Message-Id:  <9102061737.aa05706@revcan.UUCP>
>
>We seem to be getting a significant number of bad news batches which cannot be
>uncompressed.  This is causing us a major problem as the SCO news software
>hangs on a bad batch but continues to write the last log entry over and over
>until the file system fills completely at which time the system is effectively
>dead.  Any ideas, suggestions?
>
>Dave
>
>----
>
>  Does this ring ANY bells at all? I'm afraid that the previous
>discussion has expired off my system already. (I'm quite behind in my
>news.all reading)

With the mention of uncompression failure, my first thought is to check
to make sure the downstream site is using the same number of bits as the
upstream site in the compression algorithm.

># 12-bit compression is the lowest common denominator among news sites,
># and is often almost as good as the much-more-costly 16-bit compression.
>
>echo "#! cunbatch"
_________________
>compress -b 12
^^^^^^^^^^^^^^^^^

IE, perhaps your downfeed is trying to uncompress the batches using 16 bits
instead of 12 bits.  What is "a significant number"?