*BSD News Article 92759


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.ecn.uoknor.edu!solace!news-feed.inet.tele.dk!news.nacamar.de!howland.erols.net!news.mathworks.com!uunet!in1.uu.net!204.147.80.3!news.uswest.net!not-for-mail
From: Brett Rabe <brett@uswest.net>
Newsgroups: news.software.nntp,comp.unix.bsd.freebsd.misc
Subject: File exists writing symlinking article file
Date: Thu, 03 Apr 1997 15:12:39 -0600
Organization: US West !nteract Services
Lines: 56
Message-ID: <33441D47.15FB7483@uswest.net>
NNTP-Posting-Host: phouka.oss.uswest.net
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: Mozilla 3.01 (X11; I; FreeBSD 2.2.1-RELEASE i386)
Xref: euryale.cc.adfa.oz.au news.software.nntp:31805 comp.unix.bsd.freebsd.misc:38422

Alright. I can't seem to get things running again. :-)

Server throttled File exists writing symlinking article file --
throttling

That's the error. Here's the setup, and then I'll describe what I've
done. I have, btw, read the advice in the FAQ (and followed it, I
think).

I've got one main server accepting an incoming feed. It's running
just fine and dandy. I have a secondary server on which innd is
running in slave mode to the primary machine. The idea is to split
the reader load between the machines and utilize the second machine
for outgoing feeds. It had been working slick until the symlinking
error started showing up.

The machines:
P200's running FreeBSD 2.1.7.1. 256M ram. Kernel config options
that might be pertinent: AHC_TAGENABLE, AHC_ALLOW_MEMIO,
AHC_SCBPAGING_ENABLE. Spool partitions are split across two
ccd partitions (alt.binaries.* in one and everything else in
another, for a total spool of about 28gig).

Inn:
Version 1.5.1. Using MMAP and MMAP_SYNC.

What I've done (I've gotten slightly more drastic with each step --
the problem continued to occur after each step):

1) Followed the advice in the FAQ, to wit: 
Renumbered the active file with innd running via "ctlinnd renumber''".

2) Continued to follow the advice of the FAQ:
  - Renumbered the active file as above,
  - Remade the history file (makehistory)
  - Renumbered the active file as above.

3) Shut down the server and remade the active and history files.
Used makeactive and makehistory. Prior to running makehistory, I
blew away all of the links (it's just news, right?) as the man
page for makehistory suggests.

My next option is pretty radical (newfs the partitions and rebuild
the spool from the server). Realistically, that's not that big of a
deal -- it'd only take a few days to catch up, and the primary server
is, for now, absorbing the load quite well. But I'd like to get a
better grasp on the problem and its solution for the future.... so
any suggestions/advice/etc would be welcome.

Thanks.

Brett

-- 
Brett Rabe
Systems Administrator - !nteract Services