Path: utzoo!attcan!uunet!husc6!bloom-beacon!tut.cis.ohio-state.edu!mailrus!uwmcsd1!nic.MR.NET!eta!pwcs!elric!corum!root
From: root@corum.UUCP (System Administration)
Newsgroups: news.software.b
Subject: concurrent execution of rnews and inews
Keywords: rnews hang xenix_2.1.3 News_B2.11.14
Message-ID: <284@corum.UUCP>
Date: 11 Sep 88 00:29:59 GMT
Organization: Unisys Inc., Eagan, MN, 55121, USA
Lines: 45

Sigh... I'm running News B2.11.14 on a relatively small Xenix system at os
release 2.1.3.

I have News set up for spooling incoming news into .rnews for later unspooling
at non-peak hours in order to massively reduce the load on the system during
the day.  The batches are NOT compressed.

I call rnews -U from news' crontab at irregular intervals -- mostly during the
late night and early morning hours.  I also have an entry which generates
UUCP/USENET traffic reports every morning at 0730 (for the use of the
mn.traffic newsgroup).

The problem is that i can log on later on during the day and doing a ps -funews
see that, for example, an rnew 88* process has 5 seconds of cpu time on it and
has been sitting there since 0530!  Inews has also been sitting around with a
little bit of time on it since 0730.  If i just leave things the way they are,
they will simply sit there, with more inews processes appearing every day at
0730 until news runs out of processes (max count) or the system is rebooted,
or i intervene by killing off either the rnews procs or the inews procs,
whichever occurs first.

The only thing i can really think of is that somehow the two processes (inews
and rnews) are locking each other out and are in some sort of race condition.
In fact, i was sort of under the impression that a .rnews.lock file was
supposed to have been generated/checked by these two programs, but during these
trying times (i.e., when the two procs are "racing"?) i haven't noticed any
.rnews.lock file sitting around.

I would really hate to have to either run inews manually or add code to the
shell that invokes it to check to see if rnews is running first; besides that
would only solve the problem of inews running if rnews is already running.  It
wouldn't alter rnews firing up after inews.  

Perhaps if i touch a .rnews.lock file, even though inews is supposed to do
this already?

So, whats a poor sysadmin to do?

Any suggestions?  (patches??)  (Telling me to get either c-news or news 3.0
won't help since they are apparently unreleased as of yet  - unless i missed
the announcement)

derek
-- 
Derek Terveer	root@corum.UUCP		..!clyde!lily!corum!root