Discussion:
Are you messing again with the Binary News Groups
(too old to reply)
One2Go
2003-10-30 08:37:57 UTC
Permalink
Since last Friday a very stable and prooven set-up on this end is now
experiencing CRC errors when downloading completed posts. This is not the
case with EasyNews. Why can't you leave them alone once they are working?

News Group alt.binaries.multimedia

tarzan-1x04-rules.of.engagement-pdtv.svcd.rar__yEnc_CRCerror_part_3.txt

Path:
auth.uk.news.easynet.net!numbering.news.easynet.net!spool0.uk.news.easynet.
net!easynet-quince!easynet.net!news-
hub.cableinet.net!blueyonder!peer01.cox.net!peer02.cox.net!cox.net!pd7cy1no
!pd7cy2so!shaw.ca!news.alt.net!news.astraweb.com!news-big.astraweb.com!not-
for-mail
From: ***@Kandor (Bizarro)
Sender: ***@Kandor
Newsgroups: alt.binaries.multimedia
Subject: Tarzan 1x04 - "Rules Of Engagment" [STNJ] - 05 of 68 - tarzan-
1x04-rules.of.engagement-pdtv.svcd.rar (03/24)
X-Newsposter: YENC-POWER-POST-A&A-v10C (Modified POWER-POST
www.CosmicWolf.com)
Date: 29 Oct 2003 23:47:41 GMT
Lines: 5081
Message-ID: <3fa0516a$0$75972$***@news.astraweb.com>
Organization: Unlimited download news at news.astraweb.com
NNTP-Posting-Host: f74485a0.news.astraweb.com
X-Trace:
DXC=lTOZ8NVN<=HOObP71A_HVDL?***@J41>hRIa\<>***@T`iJ3213@^Ibef]OAL0LN8Sc_
K4YejK??Y21VLZ[9H
Xref: auth.uk.news.easynet.net alt.binaries.multimedia:26336483

This is 1 of 4 just in that RAR file that had CRC errors.

Please try to fix this.

Hans
Chris Needham
2003-10-30 19:00:54 UTC
Permalink
Post by One2Go
Since last Friday a very stable and prooven set-up on this end is now
experiencing CRC errors when downloading completed posts. This is not the
case with EasyNews. Why can't you leave them alone once they are working?
I see this too, I use 3 servers and parts downloaded from Easynet are
corrupt, also I've noticed parts don't exist even though the headers
show them as being there.
One2Go
2003-10-30 19:38:57 UTC
Permalink
Post by Chris Needham
Post by One2Go
Since last Friday a very stable and prooven set-up on this end is now
experiencing CRC errors when downloading completed posts. This is not
the case with EasyNews. Why can't you leave them alone once they are
working?
I see this too, I use 3 servers and parts downloaded from Easynet are
corrupt, also I've noticed parts don't exist even though the headers
show them as being there.
Chris,

Please join the discussion on easynet.news.support Rob Partington is
helping out there. Any additional information to help these guys to fix
their news server is appreciated. Please join in.

I am so frustrated with their binary news server it works for a couple of
weeks and than some jerk messes with it and the next thing incompletion or
corrupted files or even worst not even showing up. I thing I am out of here
and get myself a different ISP. I been with EasyNet for over 3 years and I
believe I join the voices of many that at this time I can no longer
recommend EasyNet as an ISP.

Thanks
Hans
Rob Partington
2003-10-30 20:42:58 UTC
Permalink
On 30 Oct 2003 19:38:57 GMT,
Post by One2Go
I am so frustrated with their binary news server it works for a couple of
weeks and than some jerk messes with it
That jerk would be me. I'm about to mess with it some more, to see if
I can improve things temporarily. Let me know if things are better
tomorrow.
Post by One2Go
and the next thing incompletion or
corrupted files or even worst not even showing up. I thing I am out of here
and get myself a different ISP.
That's a shame, but if our 7 day retention, fast speed, and pretty damned
good completion* isn't good enough, I'm sure there's another UK ISP who's
news server is better.

*reads demon.service, nildram.support and claranet.support*

But I think you might be out of luck there, since it appears that easynet
are the only people actively caring about providing a good news service
and yes, it does go wrong, but now we have 7 day retention instead of the
3 days we had a month ago[1], we've got a whole load of new stuff coming
online in the next couple of months, and, frankly, we're pretty good at
news now and I'm pretty sure we'll be the best non-news-provider news
provider in the UK when all this work is done.

* most of the time we're comparable to the pay-for US providers.

[1] which no-one appears to care about and is exceedingly frustrating
since people spent months complaining about retention and now we have
good retention, no-one has even noticed!
Rob Partington
2003-10-31 08:13:30 UTC
Permalink
On 31 Oct 2003 06:48:28 GMT,
Sorry Rob I had this vision of some network administrator that decided to
make changes and it had an effect on the NGs. If I am talking to the person
that can do something about it, accept my apologies.
Don't worry about it, I'm as annoyed as you all are about this.

(If I have done something that caused this, "jerk" would be about right.)
Now things had gotten worst over night.
Yeah, I'm trying to track down what's going on and that means a small
amount of disruption. However, there is some good news -- it appears
that at least some of the articles (I only tested a few last night)
are perfectly ok on the spool, they're just becoming corrupted on the
way to the readers.

So, in theory, once I fix that problem, assuming it takes less than
a couple days, you should be able to get the articles you want.
But recently the service has really deteriated. I don't mind 3
to 5 days retention if it means quality. Iknow you are trying hard to
accomplish both but it seems to be not working.
I'm at a total loss as to what's happening. Nothing has changed with the
reader boxes or secondary spool for nearly 18 months. We've added some
new storage, but that started six months ago and again a couple of weeks
back. It's not a problem of software or hardware because the cluster
is made up of a diverse set of machines (historical accident, but it helps
for resilience.)
Here are two headers from last night the first is part 3 of the second is
part 5. Part 4 does not exist on the server. For a posting on the 29th you
go figure.
The missing part 4 is down to my testing at the moment. It does exist,
because I downloaded it myself last night as part of the testing. You'll
be able to fetch it once the source of the corruption is isolated.

This does have my top priority at the moment and hopefully I'll be able
to figure out what's going on before the end of today. I'll keep you
all posted as to progress, fear not.
Rob Partington
2003-10-31 13:33:20 UTC
Permalink
On 31 Oct 2003 12:55:07 GMT,
What a service. Thanks Rob and I certainly didn't want to piss off the one
guy that can do something about it. If I can't get it via EasyNet I will
get it via EasyNews.
Ok, so far, on a random sampling from alt.binaries.multimedia and about
20 broken parts from things posted this morning, I think I've isolated
the problem. Some more testing will help determine if this is actually
where things are breaking. If so, there's not much I can do to fix it
retrospectively except to remove any articles that might be affected,
some (~50%) of which will be on another spool, but the rest won't be.
Mr News E. Checker
2003-10-30 20:25:33 UTC
Permalink
[Sorry, can't access the other group anonymously]

Another example is <NfSnb.49050$***@fe08.atl2.webusenet.com>. I
did a compare of this part between Easynet and Easynews and there is
just one byte difference, a ">" character that's being changed to a
":" on Easynet. Is this only affecting Yenc files, do you think?

On the bright side if the posts have PAR2 files you will only need a
few of them to correct your corrupt downloads as they are only a few
bytes out.
Rob Partington
2003-10-30 20:43:50 UTC
Permalink
On Thu, 30 Oct 2003 20:25:33 GMT,
Post by Mr News E. Checker
did a compare of this part between Easynet and Easynews and there is
just one byte difference, a ">" character that's being changed to a
":" on Easynet. Is this only affecting Yenc files, do you think?
Ta, I'll investigate that. It does seem odd that in 18 months of
continuous operation, this has never happened before and now it's
suddenly happening all over the place. But we shall see...
Mr News E Checker
2003-10-30 22:06:20 UTC
Permalink
Post by Rob Partington
On Thu, 30 Oct 2003 20:25:33 GMT,
Post by Mr News E. Checker
did a compare of this part between Easynet and Easynews and there is
just one byte difference, a ">" character that's being changed to a
":" on Easynet. Is this only affecting Yenc files, do you think?
Ta, I'll investigate that. It does seem odd that in 18 months of
continuous operation, this has never happened before and now it's
suddenly happening all over the place. But we shall see...
Thanks, let me know if you need any more. For what it's worth here is
a binary compare of the 15MB file decoded from the 40 parts of the
same file as the above Message-ID. The Easynews (good) file is the
left-hand column. To get the part number divide the offset by about
375,000. Each incorrect byte is plus 4 so looks like it's the same bit
going astray every time.

Comparing files Copy(1) of 5x5.r00 and 5X5.R00
00171610: 01 05
00171ED6: A1 A5
0040D80F: 70 74
00412610: 49 4D
0041628B: 40 44
004184B2: 89 8D
0042131A: 01 05
0055AB09: 29 2D
00561D23: 00 04
00566B32: 28 2C
005672FE: 01 05
005673F5: 89 8D
0056AF77: 81 85
0056DD4D: 49 4D
0056EA01: 09 0D
007CF2A8: F1 F5
007D6EAC: 28 2C
007E38CA: 18 1C
007E4A65: D9 DD
0081295B: 99 9D
0081AB26: D0 D4
00825AEE: 31 35
00A203EA: B9 BD
00A39706: 81 85
Rob Partington
2003-10-31 09:20:19 UTC
Permalink
On Thu, 30 Oct 2003 22:06:20 GMT,
Post by Mr News E Checker
Thanks, let me know if you need any more.
If you could do another couple of those, that'd be fantastic. Finding
that it's always the same bit being set when it shouldn't will probably
help a lot in tracking the problem down.
Mr News E Checker
2003-10-31 11:31:00 UTC
Permalink
Post by Rob Partington
On Thu, 30 Oct 2003 22:06:20 GMT,
Post by Mr News E Checker
Thanks, let me know if you need any more.
If you could do another couple of those, that'd be fantastic. Finding
that it's always the same bit being set when it shouldn't will probably
help a lot in tracking the problem down.
OK, two more files downloaded today from
alt.binaries.multimedia.buffy-v-slayer. The first one starts off with
the same pattern of an occasional extra bit 2 in the first few megs
but then goes out of sync as the Easynet file is two bytes longer. The
latter problem could be the more usual file propagation corruption.

Comparing files easynews angel.505.hdtv-lol.part24.rar and
EASYNET ANGEL.505.HDTV-LOL.PART24.RAR
001C341D: F1 F5
00631CFC: 48 4C
00638FAC: E0 E4
00640CCB: 60 64
00642606: B8 BC
0083E75A: 29 2D
0083FFA6: 39 3D
0084E875: 58 5C
00949A66: 89 8D
0094D57C: AF E4
0094D57D: 29 AF
0094D57E: 49 29
0094D57F: 26 49
0094D580: 0C 26
0094D581: 2F 0C
(etc. for rest of file)

The next file only has one bit difference in the whole 15 megs:

Comparing files easynews angel.505.hdtv-lol.part22.rar and
EASYNET ANGEL.505.HDTV-LOL.PART22.RAR

00B6230A: 99 9D

I think we have a pattern here. :-)
Loading...