Whiskers wrote:
> On 2012-05-27, John F. Morse <jo...@example.invalid> wrote:
>
>> Whiskers wrote:
>>
>>>> Do you now see why I said what I said?
>>>>
>>> Not entirely, as you clearly know what a correctly created Path header
>>> looks like.
>>>
>> Then why do you question me? I don't care if you do, but there must be some
>> reason, like misunderstanding?
>>
>
> Until you provided evidence that you do indeed know what a Path header
> should look like, I had no reason to suppose that you'd ever seen
> anything other than the wierd abomination you quoted.
>
OK, I'll accept your appearance that everything you read must be proven
as a fact.
Having read many run-of-the-mill statements by unknown (unqualified)
people, who obviously only have a minimal understanding of the subjects,
I'm not so different myself.
I am a believer in what I read on a sign posted on a backbar many years ago:
"In God we trust -- All others, cash!"
>>> If you'd posted using a Highwinds server, your comments about what they
>>> do to headers would have carried more weight.
>>>
>> Now I'm beginning to question your understanding.... ;-)
>>
>
> Accusing Highwinds of doing something whilst not yourself posting via
> Highwinds, leaves room for doubt as to the accuracy of your accusation.
>
Again, I see your point here concerning suspicion. I'll attempt to make
myself more clear and hope you will believe what I post is the truth. I
have no reason to lie.
>> If I would have posted using a Highwinds server, then the Path would
>> only show the minimal "not-for-mail" (and possibly the .POSTED entry
>> depending on the software).
>>
>
> To you apparently, if you read from Highwinds. Those reading from other
> servers would have seen a Path header starting with Highwinds and
> ending with the server being read from - like this:
>
I am aware of what the Path looks like when reading from other servers,
but the whole point of my original comment was when reading from Highwinds.
The Path is stripped in that case. Their software does it. Why I do not
know, and I doubt if they care to tell me.
If an article is posted on Highwinds, their server(s) are in the Path
when the article is read on any other NSP. It must be there or all of
these other servers would attempt to send the article back to Highwinds.
That is the main function of the Path -- to stop propagating articles
back to servers that have previously seen the articles.
The Path can also be used for tracing, but any article can propagate
over many different paths. Additionally, tracing using the Path will
only show successfully propagated articles. Those which were lost have
no Path to examine! So it is rather limited to showing a NSP
administrator which peer sent the article.
Then you must be careful which newsreader you use. For instance, Pan
will show the same article and header no matter which news server you
use to examine that article. Once downloaded, only the first article
will be presented.
>> What I provided was a Path from a message that was posted on some
>> other server and fed to Highwinds. It shows there is nothing remaining
>> in the Path to help with a trace.
>>
>
> You did indeed, when pressed for evidence.
>
Wasn't that in my original message? Or at least my reply to your
original disbelief?
>> Since you still question me,
>>
>
> I'm not yet convinced that there isn't something odd happening after the
> messages leave Highwinds and reach you. The spurious MIME header
> that also appeared in your example, is something I'd normally expect
> only to be inserted by a usenet client (and it wasn't put there by the
> sender's client, namely mine). Where did that header come from, and
> how and why? (That can go unanswered unless the answer is easy - and
> of course I can work out 'Highwinds must have done it', that's /too/
> easy!).
>
I don't find the "MIME-Version 1.0" in your articles. Only in an
article I posted using Thunderbird (quoted further down). Perhaps to
support the Format-Flowed?
Don't know, and don't really care to spend time on this fork from the
original Path topic, which itself is a fork from the original subject. ;-)
My posting and reading access was directly to and from Highwinds for
this discussion, unless I use a different NSP for an example. I do have
reader access to 32 NNTP (NNRP) Usenet servers around the world. They
are numerous, but needed for propagation testing. I use my own server
for most all of my Usenet (and private groups) activities.
Highwinds purchased
usenetserver.com, and is now providing the service.
My cable ISP has no physical news servers. They simply forward any
request for "
news.everestkc.net" (and others, see below) to the
Highwinds IP(s), where the servers are located.
Additionally the Everest cable ISP was purchased by SureWest (of
California). In every case I've tried, either FQDN can be used.
A little history....
Back in 2002 when Everest was started, I was a Time Warner (RoadRunner)
subscriber (1999-2002), as well as AT&T Worldnet (1996-2009) for backup
connectivity, I formerly used Earthlink (1998-1999), SBIS (
swbell.net)
(1997-1998), and PSI-Interramp (1995-1996). My son worked in the AOL NOC
in 1996, so I had access there FWIW.
After fighting Time Warner constantly for the previous six months of
poor or no service. I switched to Everest on November 7, 2002. Time
Warner had reported a "bad splice" -- their lingo for the center
conductor in a coax connector at a device. When cold (nights without
sunlight) the coax was contracting and the slack loop was insufficient,
so the center conductor was pulling out of the connector.
I might add that I was a CATV designer for Broadband Data Communications
(
bdc.net) in 1998-2000, mainly for Jones Intercable (now owned by
Comcast), who served Independence and Blue Springs, Missouri, and
Olathe, Kansas. I also designed for Time Warner serving the Greater
Kansas City Area, but primarily the Johnson County, Kansas portion,
which formerly was Telecable, and TCI before that, dating back into the
1960s,
I also designed the complete city of Belleville, Kansas, and most of the
DelMarVa area on the East Coast. These all were upgrades for
bidirectional CATV, higher bandwidth, and for Internet over cable.
Another project I designed was for the Los Angeles Public Schools, which
was a CCTV system. LA schools tend to have construction like a "strip
mall" where students enter individual classrooms from an exterior door,
and use multiple structures.
When the CATV industry upgrade bloom slowed down, and since I was paid
on "piece work" which was now running thin to nothing, I resigned and
moved to Black and Veatch, a major engineering business in Overland
Park, Kansas.
http://bv.com
In the Telecom Division at BV, I designed several Metricom sites
(Internet via radio, similar to cellular telephone), before Metricom
experienced financial problems.
http://en.wikipedia.org/wiki/Metricom
With my previous CATV experience, I then spent a few months leading a
group investigating an AT&T CATV plant upgrade in the Chicago, Illinois,
area. We had to review every design map, tally every device, cable,
strand, etc., and submit the results. This was for a legal action
against one or more contractors who claimed more than AT&T believed.
I designed most of the fiber ring around San Jose, California.
I Investigated local government (city/county) ordinances, laws and
building permit mandates in several states, for either increasing the
height of cell towers, or building new towers, for adding additional
Sprint cellular antennas.
My final project at BV Telecom Division was using a pole load program to
compute the requirements of utility poles which were going to be used to
support the upcoming new Everest plant. So I actually assisted in the
design of the Everest ISP I now use.
Then several dozen engineers and designers, who were stationed in south
Asia, were sent back to BV headquarters because GE lost their contract
to build nuclear power plants. Had something to do with a new
king/dictator/whatever not wanting nukes. Since these guys had more
seniority than me, I was laid off.
On my way out the door, I stopped by the Advanced Technologies Devision
at BV, to see if they had any work. I had designed a few projects for
them in the past when in Telecom, so I knew a few of the designers,
engineers and supervisors.
My services were needed, so I moved to ATD, working primarily in the
electrical engineering discipline. I designed huge server farms for
Sprint, Genuity and 360 Networks, where hundreds of cabinets were used,
and a dozen 100 kW diesel generators for emergency power (computers and
the necessary air conditioning load). Security, alarms, telephone,
racks, fiber, etc.
Then the "Dot Com" bubble burst. Laid off again, I checked the BV Power
Division on the fifth floor, and they hired me. Here I designed power
generating plant retrofits. That assignment lasted about six months, and
the work ran out.
The last stop was the BV Power Transmissio9n Division. They didn't have
enough work to hire me, and since I didn't care to work in their Water
Division, I left. Spent the next two years in a small (two-man) CNC
machine shop, running VMCs at night, building parts for airport electronics.
Getting back to Everest....
Everest never built their own in-house news servers. They farmed out the
task to
newsfeeds.com and had at least four different FQDNs available,
which were forwarded as shown here:
news.everestkc.net ==>
corp-goliath.newsfeeds.com 128.242.171.114
30, 492 groups, fast
news-binaries.everestkc.net ==>
corp-binaries.newsfeeds.com 128.242.173.230
7177 groups
news-corp.everestkc.net ==>
corp-news.newsfeeds.com 209.189.89.243
113,476 groups, very slow
news-goliath.everestkc.net ==> no record remaining
8703 groups, slow
I don't know if
newsfeeds.com was sold to
usenetserver.com, or if it was
another outsourcing. According to the cross reference, requests for
newsfeeds.com were "Redirected to others" in their table at:
http://www.newsgroupservers.net/newsgroup_server_resellers
Some time after, Highwinds outright bought
usenetserver.com, so they are
now the owner and NSP (they reportedly do not sell to individuals
though). See:
http://www.newsgroupservers.net/usenetserver_newshosting_highwinds_review
On February 21, 2007, Everest moved one of my IPs to their
newly-assigned netblock for growth. I discovered I could no longer
access
usenetserver.com using this new 74.x.x.x netblock. I checked with
usenetserver.com directly because nobody at Everest had enough
experience to even understand what I was reporting.
The smart-ass employee at
usenetserver.com was telling me I need
authentication (username and password). I told him that I don't need any
authentication, didn't for the old IP, and don't on the 64.x.x.x IP
which didn't change. My authentication is by cable modem MAC, and
assignment to a particular cable node. Nobody can access
usenetserver.com through a cable ISP unless that ISP authorizes the
access. Plus the ISP is paying for the news service that the NSP is
supposed to provide.
Well, this fiasco went round and round, and after consulting with
Everest management (some who used to work with me at BDC), I was told
usenetserver.com actually blocked Everest because someone had hacked
into Everest and stolen a bunch of passwords, then posted a lot of spam.
I don't swallow because there are no passwords -- no authentication.
Usenetserver.com is/was run by a bunch of nincompoops, or liars. Now
Highwinds is the owner. Go figger.
Afterwards I attempted to talk Everest into setting up their own Usenet
server(s). The person qualified to do this work checked with the new
owners of Everest, SureWest, and informed me that SureWest didn't want
to invest in news servers. Oh well. An "A" for trying.
What about the here and now?
traceroute to
news.everestkc.net (74.209.131.13), 30 hops max, 40 byte
packets
15
unknown.hwng.net (74.209.131.13) 54.383 ms 54.718 ms 54.959 ms
traceroute to
news-binaries.everestkc.net (74.209.131.13), 30 hops max,
40 byte packets
14 *
unknown.hwng.net (74.209.131.13) 47.054 ms 47.245 ms
traceroute to
news-corp.everestkc.net (74.209.131.13), 30 hops max, 40
byte packets
15 *
unknown.hwng.net (74.209.131.13) 62.497 ms 62.743 ms
traceroute to
news-goliath.everestkc.net (74.209.131.13), 30 hops max,
40 byte packets
12
unknown.hwng.net (74.209.131.13) 68.909 ms * *
As you can see they all resolve to 74.209.131.13. The "hwng" stands for
High Winds News Groups, or something similar.
john@noc:~$ whois 74.209.131.13
UNS Holdings, Inc. HIGHWINDS3-UNS2 (NET-74-209-130-0-1) 74.209.130.0 -
74.209.131.255
Highwinds Network Group, Inc. HIGHWINDS3 (NET-74-209-128-0-1)
74.209.128.0 - 74.209.143.255
I suspect "UNS" is the
usenetserver.com name that Highwinds uses for
tracking where articles arrived.
Correct, it is added by Highwinds to signify the message was posted by
someone connecting (through
usenetserver.com) from an Everest account.
Evidentially Highwinds has totally dropped the
usenetserver.com from the
path, using only their own servers and FQDNs.
But when the traceroute and whois are compared, the truth is exposed.
See my previous comment, and also realize that an incoming.conf file
configuration can allow using an alias. The M-ID was supplied by a
Highwinds server.
There are no outsourced NSP in that header. The article was POSTED on
the "newsfe13.iad.POSTED" server at Highwinds (see the "iad" and compare
with the next server in the Path).
Highwinds then sent it to
news2.google.com which sent it to
postnews.google.com. From there it went to
fu-berlin.de and then to
uni-berlin.de which should be the server where you read it (or possibly
you use and removed the hop to
individual.net).
How about these from the above Path header:
npeer03.iad.highwinds-media.com
news.highwinds-media.com
feed-me.highwinds-media.com
Even the
extra.newsguy.com is a transit server for Newsguy, another
Highwinds partner.
Perhaps the
newsp.newsguy.com is a "News Portal" meaning a reader
server, or a portal from the drn server.
Newsguy is famous for fiddling with the header in order to hide their
massive spam flooding.
Check these links (maybe search for "newsguy" to easily locate the
references):
http://www.mixmin.net/cleanfeed/regex.html#bad_nph_hosts (the
bad_nph_hosts parameter description)
Such twiddling as "news7" without a FQDN, and even "
newsdawg.com" are used.
http://www.mixmin.net/cleanfeed/files.html#local_flag_spamsource (an
example of the local_flag_spamsource parameter description)
The evidence points to questionable operations of Highwinds and their
other brands and partners.
> Reading between the lines of your posts in this thread, I infer that
> "Xanadu-BBS-net" is in fact a conventional news-server, despite the
> name, and that it is publicly accessible and well peered, and that you
> are the administrator of it not merely a user. If that is so, then I
> can understand why you might be annoyed that didn't know.
>
Yes,
xanadu-bbs.net is a conventional news server farm operating
24/7/365, using multiple INN 2.5.2 servers. I am the sole administrator
and also the owner, and completely finance the operation.
It is a "private project" accessible by invitation only, and peers with
multiple Usenet servers around the world.
Today it ranks at #32 on the Top1000 Usenet server list at:
http://top1000.anthologeek.net/topsum.current.txt Last week it was
ranked at #26. Today there are six news servers which I peer with in the
top 30, and three in the top 10.
Used not only for standard Usenet text groups, and a half-dozen binary
groups, it has 32,679 newsgroups in the active file, of which 481 groups
are private and not propagated to Usenet.
Beside normal newsgroup access, the mission includes development and
testing, and supporting Usenet with a multiple transit server configuration.
The NNTP operation originally started in 2000, using the free Highwinds
Twister server, which later became bCandid's property. Then I switched
over to using RumorMill software on a 120 MHz Power Macintosh 7600 in 2001.
In 2005, Debian 3.1 "Sarge" and INN 2.4.3 were used for the operation.
Today the OS is all Debian 6.0.4 "Squeeze" and INN 2.5.2 for the
production servers.
In 1977 the original "Xanadu BBS" was a private dial-up BBS running
homemade software on an Apple ][ computer. It was improved over the
years, and wound up running the Hermes BBS software on a Mac Plus in
1990, then on a Mac IIsi in 1995, until all the users departed for the
Internet around 2001. At the height we had several hundred local users,
and three phone lines.
I just kept using the name.