trouble upgrading to BETA
Charlie Fluffy
FLUFFIE at MEOWVAX.INT.TELE.DK
Tue Jun 27 14:10:22 UTC 2000
On Tue 2492 Sep 1993, Russ Allbery wrote:
> > I saw this in a BETA snapshot of two weeks ago, and I see it again in
> > the snapshot I just grabbed a few hours ago. Yes, >2GB buffers.
>
> Is your Perl compiled with large file support? I'm guessing that's the
> problem, and may also be the problem here:
See, I knew I should have waited until I could copy somewhat
more useful information... STAND BACK, I GOT A MOUSE IN MY POCKET
AND I'M NOT AFRAID TO USE HIM
Here is the output I see from the latest /news/bin/cnfsstat -a :
[15:40:53]news at news02.inet.tele.dk:/arts/alt/binaries/nude{635}$ /news/bin/cnfsstat -a
Class PR0NVIDZ for groups matching "alt.binaries.multimedia.erotic*,alt.binarie
s.erotica.vcd*,alt.binaries.movies.erotic*,alt.binaries.nospam.multimedia.erotic
*"
Buffer ONE, size: 0.00 kBytes, position: 0.00 kBytes 1.00 cycles
Newest: 2000-06-27 15:41:00, 0 days, 0:00:02 ago
Oldest: 2000-05-27 15:48:48, 30 days, 23:52:14 ago
Class PR0NPIX for groups matching "alt.binaries.pictures.erotic*,alt.sex.pictur
es*,alt.binaries.pictures.suze*,alt.binaries.fullpost*,alt.binaries.erotic*,alt.
binaries.nospam*", article size min/max: 0/200000
Buffer FIVE, size: 0.00 kBytes, position: 0.00 kBytes 0.00 cycles
Newest: 2000-06-27 15:40:55, 0 days, 0:00:07 ago
Oldest: 2000-06-18 17:10:00, 8 days, 22:31:02 ago
Buffer NINE, size: 0.00 kBytes, position: 0.00 kBytes 1.00 cycles
Newest: 2000-06-27 15:40:55, 0 days, 0:00:07 ago
Oldest: 2000-05-26 1:55:53, 32 days, 13:45:09 ago
Class SOUNDZ for groups matching "alt.binaries.sounds.mp3*,alt.binaries.mp3*,al
t.binaries.sounds.jpop*,alt.binaries.sounds.music*,alt.binaries.sounds.samples*"
Buffer TWO, size: 0.00 kBytes, position: 0.00 kBytes 3.00 cycles
Newest: 2000-06-27 15:41:00, 0 days, 0:00:03 ago
Oldest: 2000-06-21 3:36:01, 6 days, 12:05:02 ago
Buffer SIX, size: 0.00 kBytes, position: 0.00 kBytes 2.00 cycles
Newest: 2000-06-27 15:41:01, 0 days, 0:00:03 ago
Oldest: 2000-06-21 3:30:46, 6 days, 12:10:18 ago
Class MULTIMEDIA for groups matching "alt.binaries.multimedia*,alt.binaries.mov
ies*,alt.binaries.mpeg*,alt.binaries.vcd*,alt.anime*,alt.binaries.anime*"
Buffer SEVEN, size: 0.00 kBytes, position: 0.00 kBytes 4.00 cycles
Newest: 2000-06-27 15:41:01, 0 days, 0:00:04 ago
Oldest: 2000-06-25 5:23:39, 2 days, 10:17:26 ago
Class WAREZ for groups matching "alt.*warez*,alt.binaries.emulators*,alt.binari
es.cbts*,alt.binaries.mac*,alt.binaries.images*,alt.binaries.cd*,alt.binaries.ga
mes*,alt.binaries.w-software*,alt.binaries.pictures.leek*"
Buffer THREE, size: 0.00 kBytes, position: 0.00 kBytes 11.00 cycles
Newest: 2000-06-27 15:41:05, 0 days, 0:00:00 ago
Oldest: 2000-06-26 5:22:13, 1 days, 10:18:52 ago
[15:44:42]news at news02.inet.tele.dk:/arts/alt/binaries/nude{644}$ head !:0
head /news/bin/cnfsstat
#! /usr/local/bin/perl
require '/news/lib/innshellvars.pl';
# $Id: cnfsstat.in,v 1.5.2.5 2000/06/18 16:12:54 kondou Exp $
The fact that I do get the expected output from the following
would lead me to believe that perl has the largefile support,
even if I didn't cheat and look at the source config options...
[15:46:52]news at news02.inet.tele.dk:/arts/alt/binaries/nude{646}$ ~/cnfsstat -a
Class PR0NVIDZ for groups matching "alt.binaries.multimedia.erotic*,alt.binarie
s.erotica.vcd*,alt.binaries.movies.erotic*,alt.binaries.nospam.multimedia.erotic
*"
Buffer ONE, len: 42480.47 Mbytes, used: 7941.88 Mbytes (18.7%) 1 cycles
Newest: 2000-06-27 15:47:07, 0 days, 0:00:03 ago
Oldest: 2000-05-27 16:47:13, 30 days, 22:59:57 ago
Class PR0NPIX for groups matching "alt.binaries.pictures.erotic*,alt.sex.pictur
es*,alt.binaries.pictures.suze*,alt.binaries.fullpost*,alt.binaries.erotic*,alt.
binaries.nospam*", article size min/max: 0/200000
Buffer FIVE, len: 42480.47 Mbytes, used: 28308.18 Mbytes (66.6%) 0 cycles
Newest: 2000-06-27 15:47:07, 0 days, 0:00:04 ago
Oldest: 2000-06-18 17:10:00, 8 days, 22:37:11 ago
Buffer NINE, len: 42480.47 Mbytes, used: 6047.62 Mbytes (14.2%) 1 cycles
Newest: 2000-06-27 15:47:07, 0 days, 0:00:04 ago
Oldest: 2000-05-26 1:58:14, 32 days, 13:48:57 ago
Class SOUNDZ for groups matching "alt.binaries.sounds.mp3*,alt.binaries.mp3*,al
t.binaries.sounds.jpop*,alt.binaries.sounds.music*,alt.binaries.sounds.samples*"
Buffer TWO, len: 42480.47 Mbytes, used: 10979.90 Mbytes (25.8%) 3 cycles
Newest: 2000-06-27 15:47:16, 0 days, 0:00:05 ago
Oldest: 2000-06-21 5:32:10, 6 days, 10:15:11 ago
Buffer SIX, len: 42480.47 Mbytes, used: 17633.30 Mbytes (41.5%) 2 cycles
Newest: 2000-06-27 15:47:54, 0 days, 0:00:02 ago
Oldest: 2000-06-21 3:44:43, 6 days, 12:03:13 ago
Class MULTIMEDIA for groups matching "alt.binaries.multimedia*,alt.binaries.mov
ies*,alt.binaries.mpeg*,alt.binaries.vcd*,alt.anime*,alt.binaries.anime*"
Buffer SEVEN, len: 42480.47 Mbytes, used: 5324.29 Mbytes (12.5%) 4 cycles
Newest: 2000-06-27 15:47:55, 0 days, 0:00:01 ago
Oldest: 2000-06-25 5:31:49, 2 days, 10:16:07 ago
Class WAREZ for groups matching "alt.*warez*,alt.binaries.emulators*,alt.binari
es.cbts*,alt.binaries.mac*,alt.binaries.images*,alt.binaries.cd*,alt.binaries.ga
mes*,alt.binaries.w-software*,alt.binaries.pictures.leek*"
^C
[15:48:57]news at news02.inet.tele.dk:/arts/alt/binaries/nude{647}$
There I got bored while ~/cnfsstat was chewing around in the
buffer, and interrupted it after half a minute or so.
[15:48:57]news at news02.inet.tele.dk:/arts/alt/binaries/nude{647}$ head !:0
head ~/cnfsstat
#! /usr/bin/perl
require '/news/lib/innshellvars.pl';
# $Id: cnfsstat.in,v 1.4 2000/02/10 17:37:24 kondou Exp $
[15:50:34]news at news02.inet.tele.dk:/arts/alt/binaries/nude{650}$ ls -li /usr/bin/perl /usr/local/bin/perl*
268783 -rwxr-xr-x 3 root other 17360 May 10 15:40 /usr/bin/perl
268783 -rwxr-xr-x 3 root other 17360 May 10 15:40 /usr/local/bin/perl
268783 -rwxr-xr-x 3 root other 17360 May 10 15:40 /usr/local/bin/perl5.6.0
> > I also have a copy of `cnfsstat' from an older -CURRENT, which does give
> > the sizes, but seems to have the problem of getting stuck on a
> > particular buffer while the latest BETA has no problem zipping through
> > them all.
I'm pretty sure this older cnfsstat can get through all the buffers
quickly if the swerver is somehow not running, so my guess is that
the warez buffer, with a 1+ day retention time, which cycles the
fastest, is just being updated by new articles too fast for the older
cnfsstat to catch up or something, while the newest one (/news/bin/cnfsstat
at top) which doesn't report these %usage/current-pointer stats
properly, isn't doing this.
But what do I know. I did a quick diff between the two skr1ptZ
and saw that it was non-trivial, and more than my tired brane
could handle at that time, and I've avoided doing anything with
it since -- hey, the older one gave me the info I was after, and
besides, pokemon
More information about the inn-workers
mailing list