I recall seeing a post in this list (this year) about digests having
this type of problem. The solution was to edit the digest script and
add the word sort somewhere, but you would have to check the archives to
see where. I just don't remember.
Dan Liston
Joe Hartley wrote:
>
> A while back, I wrote:
> > > > The details: MD 1.94.5 on a Linux box of unknown flavor (possibly
> > > > slackware 7.0.0, can't tell for sure), perl 5.6.0
> > > >
> > > > The problem: the system is sending out one digest per post to a list
> > > > I run called tft. It used to work fine, but the host has undergone
> > > > some changes lately and now I've got this problem.
> > > > ( snip )
> > > > PS: while I was writing this, I may have found the problem. There were a
> > > > couple of old files in the work directory:
> > > >
> > > > (36) root@endless:/usr/local/majordomo/digest-work/tft-digest# ls -l
> > > > total 8
> > > > -rw-rw---- 1 majordom daemon 1514 May 25 08:50 001
> > > > -rw-rw---- 1 majordom daemon 2297 May 26 22:25 005
>
> I've reinstalled MJ 1.94.5 from scratch, but I still have the problem,
> but there's more to it now.
>
> The list goes along happily building up the messages in the digest work area,
> waiting for that one message that will kick off the digest-making process:
>
> /usr/local/majordomo# ls -l queue-tft/*
> -rw-rw---- 1 majordom daemon 1988 Jun 25 11:01 queue-tft/001
> -rw-rw---- 1 majordom daemon 1972 Jun 25 11:01 queue-tft/002
> -rw-rw---- 1 majordom daemon 2392 Jun 25 11:01 queue-tft/003
> -rw-rw---- 1 majordom daemon 1341 Jun 25 11:01 queue-tft/004
> -rw-rw---- 1 majordom daemon 2076 Jun 25 11:56 queue-tft/005
> -rw-rw---- 1 majordom daemon 2291 Jun 25 14:11 queue-tft/006
> -rw-rw---- 1 majordom daemon 1980 Jun 25 14:54 queue-tft/007
> -rw-rw---- 1 majordom daemon 1580 Jun 25 17:26 queue-tft/008
> -rw-rw---- 1 majordom daemon 1961 Jun 25 17:27 queue-tft/009
> -rw-rw---- 1 majordom daemon 2686 Jun 25 17:50 queue-tft/010
> -rw-rw---- 1 majordom daemon 3157 Jun 25 18:22 queue-tft/011
> -rw-rw---- 1 majordom daemon 1969 Jun 25 21:16 queue-tft/012
> -rw-rw---- 1 majordom daemon 1558 Jun 25 21:24 queue-tft/013
> -rw-rw---- 1 majordom daemon 2958 Jun 25 22:01 queue-tft/014
> -rw-rw---- 1 majordom daemon 1965 Jun 25 22:37 queue-tft/015
> -rw-rw---- 1 majordom daemon 2587 Jun 25 23:52 queue-tft/016
> -rw-rw---- 1 majordom daemon 1882 Jun 26 00:12 queue-tft/017
> -rw-rw---- 1 majordom daemon 1478 Jun 26 00:43 queue-tft/018
> -rw-rw---- 1 majordom daemon 2307 Jun 26 02:32 queue-tft/019
> -rw-rw---- 1 majordom daemon 1237 Jun 26 02:32 queue-tft/020
> -rw-rw---- 1 majordom daemon 2137 Jun 26 05:44 queue-tft/021
> -rw-rw---- 1 majordom daemon 2473 Jun 26 09:45 queue-tft/022
>
> When the digest gets made, though, not all the posts make it into the
> digest! Here's what got left over:
>
> # ls -l queue-tft/*
> -rw-rw---- 1 majordom daemon 1988 Jun 25 11:01 queue-tft/001
> -rw-rw---- 1 majordom daemon 1580 Jun 25 17:26 queue-tft/008
> -rw-rw---- 1 majordom daemon 2307 Jun 26 02:32 queue-tft/019
>
> If I leave these messages in the digest work area numbered like this,
> at some point the digest will break and start sending out one digest per
> post. I have not left it long enough to find out whether this happens
> as soon as new posts start coming in or whether it waits until it runs into
> the situation where a post would have the same number as an existing
> file in the work area.
>
> I can fix things by moving the leftovers to numerically sequential files
> (001, 002, etc.) and things will be fine until the next day when the digest
> goes out and it happens again.
>
> The files that get left behind are not the same numerically each time, and
> I can't find anything wrong in the headers of the messages like an incorrect
> date caused by a wrong date set on someone's PC.
>
> Furthermore, the messages in the digest do not appear in the order in
> which they were posted!!
>
> In the digest.config file, I have the following settings:
> digest_maxdays = 1
> digest_maxlines =
> maxlength = 60000
>
> This is driving me absolutely buggy. I've been administering MJ lists
> for a long time on a number of different servers of various flavors. I've
> never seen anything like this before. I've got a digested list running on
> a Sun that never ever has a problem; I have no clue why this particular
> list is giving me such a headache.
>
> Any clues, guesses and prognostications gratefully accepted!
>
> ======================================================================
> Joe Hartley - UNIX/network Consultant - jh@brainiac.com
> 12 Emma G Lane, Narragansett, RI 02882 - vox 401.782.9042
> Without deviation from the norm, "progress" is not possible. - FZappa
--
-- Dan Liston --------------------- dliston@netscape.com --
-- Netscape Consultant ------------------ (605) 232-3170 --
-- iPlanet E-Commerce Solutions, A Sun Netscape Alliance --
Follow-Ups:
References:
|
|
