• another peeve with newsmaster who dont do common sense

    From noel@deletethis@invalid.lan to news.admin.peering on Sat Aug 1 09:56:44 2026
    From Newsgroup: news.admin.peering

    Just discovered (after a 21 day grace period) an overnight alert from our monitoring we weren't getting anything (since July 10) from blueworld
    hosting, it seems after some checks his IP changed, and i'll be fucked if
    I can find any reference or warning that was to happen let alone an after
    the fact, all completed blah blah blah, and of consequence to that
    fail2ban blackholed them for repeatedly trying to send whilst getting
    denied :)

    Now BWH is not the first peer who's done that, and sadly wont be the
    last, but all it takes is a minute of time to send post to this group if
    you dont bother keeping an internal email list of peers to notify...


    I'll rmeove the ban and correct IP's in access file, if feed returns it returns, if not, it doesnt.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nomen Nescio@nobody@dizum.com to news.admin.peering on Sat Aug 1 03:13:58 2026
    From Newsgroup: news.admin.peering

    Do you not use DNS to refresh your name -> IP mapping regularly?

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Marco Moock@mm@dorfdsl.de to news.admin.peering on Sat Aug 1 06:18:37 2026
    From Newsgroup: news.admin.peering

    Am 01.08.26 um 05:13 schrieb Nomen Nescio:
    Do you not use DNS to refresh your name -> IP mapping regularly?

    IIRC INN resolved the names one time and then never does it again until restarted. Although that is from hearsay.
    --
    Gru|f
    Marco

    Junk-Mail bitte an trashcan@stinkedores.dorfdsl.de
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Ray Banana@rayban@raybanana.net to news.admin.peering on Sat Aug 1 08:52:09 2026
    From Newsgroup: news.admin.peering

    Thus spake Marco Moock <mm@dorfdsl.de>

    Am 01.08.26 um 05:13 schrieb Nomen Nescio:
    Do you not use DNS to refresh your name -> IP mapping regularly?
    IIRC INN resolved the names one time and then never does it again
    until restarted. Although that is from hearsay.

    That would indicate that a daily cron job is missing:

    0 0 * * * test -x /usr/local/news/bin/ctlinnd &&
    /usr/local/news/bin/ctlinnd -t 300 -s reload incoming.conf "flush cache"

    Has been working for me for decades.
    --
    -f-a|U-e-u-+ rCo -a-a-|-+-+|U
    https://www.eternal-september.org
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Kevin Bowling@kevin.bowling@kev009.com to news.admin.peering on Sat Aug 1 01:10:08 2026
    From Newsgroup: news.admin.peering

    On 7/31/26 23:52, Ray Banana wrote:
    Thus spake Marco Moock <mm@dorfdsl.de>

    Am 01.08.26 um 05:13 schrieb Nomen Nescio:
    Do you not use DNS to refresh your name -> IP mapping regularly?
    IIRC INN resolved the names one time and then never does it again
    until restarted. Although that is from hearsay.

    That would indicate that a daily cron job is missing:

    0 0 * * * test -x /usr/local/news/bin/ctlinnd &&
    /usr/local/news/bin/ctlinnd -t 300 -s reload incoming.conf "flush cache"

    Has been working for me for decades.

    Yeah you got it, that's the intention with inn right now and there's an
    issue on GitHub describing how to fix: https://github.com/InterNetNews/inn/issues/89.

    There are plenty of reasons to not expect a peer to be fixed to one IP
    unless you have a mutual understanding of that case.

    Regards,
    Kevin
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From noel@deletethis@invalid.lan to news.admin.peering on Sat Aug 1 20:07:33 2026
    From Newsgroup: news.admin.peering

    On Sat, 01 Aug 2026 03:13:58 +0000, Nomen Nescio wrote:

    Do you not use DNS to refresh your name -> IP mapping regularly?

    No, real news servers (as in ISP's) have since the 90s granted ihave
    access to IP's, it was generally frowned upon to use hostnames for
    obvious reasons, IP's have been used in my own, and prior to that
    Tesltra's servers (ad I know optusnet, iprimus, also did, I also know
    many of the US ISP's used IP as well, Jesse is very aware I use IP as I
    told him at outset, and the last time he changed his IP.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From noel@deletethis@invalid.lan to news.admin.peering on Sat Aug 1 20:08:12 2026
    From Newsgroup: news.admin.peering

    On Sat, 01 Aug 2026 08:52:09 +0200, Ray Banana wrote:

    Thus spake Marco Moock <mm@dorfdsl.de>

    Am 01.08.26 um 05:13 schrieb Nomen Nescio:
    Do you not use DNS to refresh your name -> IP mapping regularly?
    IIRC INN resolved the names one time and then never does it again until
    restarted. Although that is from hearsay.

    That would indicate that a daily cron job is missing:

    0 0 * * * test -x /usr/local/news/bin/ctlinnd &&
    /usr/local/news/bin/ctlinnd -t 300 -s reload incoming.conf "flush
    cache"

    Has been working for me for decades.

    That wont do shit for me, since we dont use INN :)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From noel@deletethis@invalid.lan to news.admin.peering on Sat Aug 1 20:55:44 2026
    From Newsgroup: news.admin.peering

    On Sat, 01 Aug 2026 01:10:08 -0700, Kevin Bowling wrote:

    On 7/31/26 23:52, Ray Banana wrote:
    Thus spake Marco Moock <mm@dorfdsl.de>

    Am 01.08.26 um 05:13 schrieb Nomen Nescio:
    Do you not use DNS to refresh your name -> IP mapping regularly?
    IIRC INN resolved the names one time and then never does it again
    until restarted. Although that is from hearsay.

    That would indicate that a daily cron job is missing:

    0 0 * * * test -x /usr/local/news/bin/ctlinnd &&
    /usr/local/news/bin/ctlinnd -t 300 -s reload incoming.conf "flush
    cache"

    Has been working for me for decades.

    Yeah you got it, that's the intention with inn right now and there's an
    issue on GitHub describing how to fix: https://github.com/InterNetNews/inn/issues/89.

    There are plenty of reasons to not expect a peer to be fixed to one IP
    unless you have a mutual understanding of that case.

    Regards,
    Kevin

    even running multiple front ends using a load balancer which is what we
    did at $previous_life -not here in my own organisation as these days
    usage is obviously very different from 30 yrs ago, or even 10 years ago, everything can run on one server when doing text only groups, even with decades of retention, we have a HP dl380 gen4 in early days, only had 16G
    ram, it could happly handle 600 concurrent usenet sessions, and the load
    was still under 1, I guess thats why its hard to get me to dump dnews.

    Anyways I have enough peers, and as most my peers peer with each other
    (why I've never approached Ray), I fail to see why I too need 30 peers,
    I'm happy with the 5, I have (excluding another 4 what I call downstream
    peers which are universities and dont tend to post much outside their
    local groups).

    Some of my peers have changed IP's over time for various reasons, in
    recent times, one we've probably all seen one with their thoughtful notifications, that are appreciated by some, and they should be seen as
    an example on how to communicate changes, well done there, over the
    ditch ;)

    So Kevin I see no valid reason for a peer not to be fixed to one IP (if multiple you let your peers know -how its always been done), in fact isnt
    it a case that most of us here have policies requiring a static IP to
    even be considered for peering :)

    All in all its just common courtesy to give your peers a heads up, its
    fine if Jesse no longer wants peering, jesus, it took me 3 weeks to
    discover he was MIA courtesy of the pending articles list file hitting
    10Mb which triggered the email alert, anyways he can either accept the
    advice or ignore it, the sun will still rise in the morning.


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Kevin Bowling@kevin.bowling@kev009.com to news.admin.peering on Sat Aug 1 04:50:28 2026
    From Newsgroup: news.admin.peering

    On 8/1/26 03:55, noel wrote:
    On Sat, 01 Aug 2026 01:10:08 -0700, Kevin Bowling wrote:

    On 7/31/26 23:52, Ray Banana wrote:
    Thus spake Marco Moock <mm@dorfdsl.de>

    Am 01.08.26 um 05:13 schrieb Nomen Nescio:
    Do you not use DNS to refresh your name -> IP mapping regularly?
    IIRC INN resolved the names one time and then never does it again
    until restarted. Although that is from hearsay.

    That would indicate that a daily cron job is missing:

    0 0 * * * test -x /usr/local/news/bin/ctlinnd &&
    /usr/local/news/bin/ctlinnd -t 300 -s reload incoming.conf "flush
    cache"

    Has been working for me for decades.

    Yeah you got it, that's the intention with inn right now and there's an
    issue on GitHub describing how to fix:
    https://github.com/InterNetNews/inn/issues/89.

    There are plenty of reasons to not expect a peer to be fixed to one IP
    unless you have a mutual understanding of that case.

    Regards,
    Kevin

    even running multiple front ends using a load balancer which is what we
    did at $previous_life -not here in my own organisation as these days
    usage is obviously very different from 30 yrs ago, or even 10 years ago, everything can run on one server when doing text only groups, even with decades of retention, we have a HP dl380 gen4 in early days, only had 16G ram, it could happly handle 600 concurrent usenet sessions, and the load
    was still under 1, I guess thats why its hard to get me to dump dnews.

    Anyways I have enough peers, and as most my peers peer with each other
    (why I've never approached Ray), I fail to see why I too need 30 peers,
    I'm happy with the 5, I have (excluding another 4 what I call downstream peers which are universities and dont tend to post much outside their
    local groups).

    Some of my peers have changed IP's over time for various reasons, in
    recent times, one we've probably all seen one with their thoughtful notifications, that are appreciated by some, and they should be seen as
    an example on how to communicate changes, well done there, over the
    ditch ;)

    So Kevin I see no valid reason for a peer not to be fixed to one IP (if multiple you let your peers know -how its always been done), in fact isnt
    it a case that most of us here have policies requiring a static IP to
    even be considered for peering :)

    Many larger providers use multiple A records for good reasons, it's
    valid to design around articles flooding into networks and you really shouldn't put a load balancer in front of feeders.. just use the
    protocols effectively. Inferring that there must be a relationship to
    IPs or that an operator is forever bound to one set of systems/PoPs
    doesn't reflect how these systems are run. You need to have proper rDNS
    set up and the 3 major news server software handle this per design.

    All in all its just common courtesy to give your peers a heads up, its
    fine if Jesse no longer wants peering, jesus, it took me 3 weeks to
    discover he was MIA courtesy of the pending articles list file hitting
    10Mb which triggered the email alert, anyways he can either accept the
    advice or ignore it, the sun will still rise in the morning.
    Maybe look at the bigger picture of what he is doing and that most of
    use are cramming administration into active life. Your server your
    rules, so to each their own.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From noel@deletethis@invalid.lan to news.admin.peering on Sat Aug 1 23:54:14 2026
    From Newsgroup: news.admin.peering

    On Sat, 01 Aug 2026 04:50:28 -0700, Kevin Bowling wrote:


    Many larger providers use multiple A records for good reasons, it's

    yeah, they're too cheap to use what the big boys use hardware load
    balancers

    valid to design around articles flooding into networks and you really shouldn't put a load balancer in front of feeders.. just use the

    /me shakes head

    protocols effectively. Inferring that there must be a relationship to
    IPs or that an operator is forever bound to one set of systems/PoPs
    doesn't reflect how these systems are run. You need to have proper rDNS

    thats contrary to my experience in the real world of running ISP's and carriers for over 30 years, crap thats over 35 years now actiually

    set up and the 3 major news server software handle this per design.


    whoooooooooooppppieeeeeeeeeeeeeeee who cares, are you trying to tell me i should adjust the way i do things, the same way national carriers have
    been doing this for over 30 years just because you say so, just to
    resolve one persons M.O.O? you're a bigger fucking idiot than I was
    starting to think if so.



    All in all its just common courtesy to give your peers a heads up, its
    fine if Jesse no longer wants peering, jesus, it took me 3 weeks to
    discover he was MIA courtesy of the pending articles list file hitting
    10Mb which triggered the email alert, anyways he can either accept the
    advice or ignore it, the sun will still rise in the morning.

    Maybe look at the bigger picture of what he is doing and that most of
    use are cramming administration into active life. Your server your
    rules, so to each their own.

    ohh wow really, and i thought running my organisation 15 hours a day 6
    days a week was because i was bored and wanted to avoid the family..
    thats where i went wrong.. jesus fucking christ. Jesse miught lack common courtesy, but you lack common sense, and thats worse.
    dont waste your time trolling me again.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Ray Banana@rayban@raybanana.net to news.admin.peering on Sat Aug 1 16:35:17 2026
    From Newsgroup: news.admin.peering

    Thus spake noel <deletethis@invalid.lan>

    On Sat, 01 Aug 2026 08:52:09 +0200, Ray Banana wrote:
    Thus spake Marco Moock <mm@dorfdsl.de>

    Am 01.08.26 um 05:13 schrieb Nomen Nescio:
    Do you not use DNS to refresh your name -> IP mapping regularly?
    IIRC INN resolved the names one time and then never does it again until
    restarted. Although that is from hearsay.

    That would indicate that a daily cron job is missing:

    0 0 * * * test -x /usr/local/news/bin/ctlinnd &&
    /usr/local/news/bin/ctlinnd -t 300 -s reload incoming.conf "flush
    cache"

    Has been working for me for decades.

    That wont do shit for me, since we dont use INN :)

    I know you are using D-News. Marco, however, referred explicitly to INN
    ;-)
    --
    -f-a|U-e-u-+ rCo -a-a-|-+-+|U
    https://www.eternal-september.org
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Jesse Rehmer@jesse.rehmer@blueworldhosting.com to news.admin.peering on Sat Aug 1 16:25:25 2026
    From Newsgroup: news.admin.peering

    On Aug 1, 2026 at 8:54:14rC>AM CDT, "noel" <deletethis@invalid.lan> wrote:

    On Sat, 01 Aug 2026 04:50:28 -0700, Kevin Bowling wrote:


    Many larger providers use multiple A records for good reasons, it's

    yeah, they're too cheap to use what the big boys use hardware load
    balancers

    valid to design around articles flooding into networks and you really
    shouldn't put a load balancer in front of feeders.. just use the

    /me shakes head

    protocols effectively. Inferring that there must be a relationship to
    IPs or that an operator is forever bound to one set of systems/PoPs
    doesn't reflect how these systems are run. You need to have proper rDNS

    thats contrary to my experience in the real world of running ISP's and carriers for over 30 years, crap thats over 35 years now actiually

    set up and the 3 major news server software handle this per design.


    whoooooooooooppppieeeeeeeeeeeeeeee who cares, are you trying to tell me i should adjust the way i do things, the same way national carriers have
    been doing this for over 30 years just because you say so, just to
    resolve one persons M.O.O? you're a bigger fucking idiot than I was
    starting to think if so.



    All in all its just common courtesy to give your peers a heads up, its
    fine if Jesse no longer wants peering, jesus, it took me 3 weeks to
    discover he was MIA courtesy of the pending articles list file hitting
    10Mb which triggered the email alert, anyways he can either accept the
    advice or ignore it, the sun will still rise in the morning.

    Maybe look at the bigger picture of what he is doing and that most of
    use are cramming administration into active life. Your server your
    rules, so to each their own.

    ohh wow really, and i thought running my organisation 15 hours a day 6
    days a week was because i was bored and wanted to avoid the family..
    thats where i went wrong.. jesus fucking christ. Jesse miught lack common courtesy, but you lack common sense, and thats worse.
    dont waste your time trolling me again.

    Jesus fucking christ. Just de-peer with me if you feel so strongly that I lack common courtesy.

    I'm not a commercial entity, I can't monitor everything 24x7, and a 10-second e-mail would have resolved this quickly.

    I had to quickly move providers. Sometimes shit happens and you aren't left with any choice. I didn't notice issues with incoming/outgoing feeds. One peer not accepting articles isn't something I feel the urge to look into urgently. Maybe after several months, but there's extremely little tax on my system to have a peer that doesn't respond.

    To each their own.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Ivo Gandolfo@usenet@bofh.team to news.admin.peering on Sat Aug 1 21:02:57 2026
    From Newsgroup: news.admin.peering

    Il 01/08/2026 15:54, noel ha scritto:
    [cut]

    ohh wow really, and i thought running my organisation 15 hours a day 6
    days a week was because i was bored and wanted to avoid the family..
    thats where i went wrong.. jesus fucking christ. Jesse miught lack common courtesy, but you lack common sense, and thats worse.
    dont waste your time trolling me again.

    No one's trolling you. And to add to what Jesse already said, to make a
    long story short, it's your fault. If you'd bothered to read the RFCs
    and all the manuals, you'd find it written practically EVERYWHERE that
    they no longer use IPs but hosts. Because IPs change more frequently
    than the hosts themselves.
    I myself went through something similar recently. I went from having a
    server to having an infrastructure. When I made the switch from a single server -> multiple feeder servers and played with the DNS, within an
    hour all my peers automatically synchronized, without much fuss, and I
    bet many (if not all) of them didn't even notice the change.
    I'm also opening PoPs pretty much everywhere. Every time I add a server
    to the pool, I should call every newsadmin I peer with and notify them
    (or if the system detects a server is down and removes it from the pool)
    "Hey! I removed/added server x.y.z.a.! Update the tables!" It doesn't
    seem like a great idea, and we haven't returned to the ARPAnet era when
    system administrators called each other, especially since there aren't
    just 10 servers in the world anymore.

    The person who literally invented RFCs (and IPv4) said, rCLBe conservative
    in what you do, be liberal in what you accept from others,rCY and itrCOs called the Robustness Principle. I highly recommend you give it a read
    (and not just that).

    BTW, if you run a company in 2026 and insist and persist in using
    software that was admittedly abandoned in the '90s and don't intend to
    update, it's YOUR problem, not everyone else's.

    At least five newsmasters responded to you (including me), and all five slapped you. This confirms to me (and should give you pause) that maybe,
    just maybe, it's not all Jesse's fault, but the problem stay in your side.


    Sincerely
    --
    Ivo Gandolfo
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From noel@deletethis@invalid.lan to news.admin.peering on Sun Aug 2 09:04:12 2026
    From Newsgroup: news.admin.peering

    On Sat, 01 Aug 2026 21:02:57 +0200, Ivo Gandolfo wrote:



    No one's trolling you. And to add to what Jesse already said, to make a
    long story short, it's your fault. If you'd bothered to read the RFCs
    and all the manuals, you'd find it written practically EVERYWHERE that
    they no longer use IPs but hosts. Because IPs change more frequently
    than the hosts themselves.
    I myself went through something similar recently. I went from having a
    server to having an infrastructure. When I made the switch from a single server -> multiple feeder servers and played with the DNS, within an


    ohh so its my fault i come from an ISP world - remember us, back in teh
    day we were the only ones providing usenet, now any diskhead can lease a
    VPS and spin one up.

    my bad, then i might as well not only terminate jesse's peer, I might as
    well terminate all of them, I was looking for hardware to expand our
    mirror service, now I found it.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Kevin Bowling@kevin.bowling@kev009.com to news.admin.peering on Sat Aug 1 16:25:49 2026
    From Newsgroup: news.admin.peering

    On 8/1/26 06:54, noel wrote:
    On Sat, 01 Aug 2026 04:50:28 -0700, Kevin Bowling wrote:


    Many larger providers use multiple A records for good reasons, it's

    yeah, they're too cheap to use what the big boys use hardware load
    balancers

    valid to design around articles flooding into networks and you really
    shouldn't put a load balancer in front of feeders.. just use the

    /me shakes head

    protocols effectively. Inferring that there must be a relationship to
    IPs or that an operator is forever bound to one set of systems/PoPs
    doesn't reflect how these systems are run. You need to have proper rDNS

    thats contrary to my experience in the real world of running ISP's and carriers for over 30 years, crap thats over 35 years now actiually

    set up and the 3 major news server software handle this per design.


    whoooooooooooppppieeeeeeeeeeeeeeee who cares, are you trying to tell me i should adjust the way i do things, the same way national carriers have
    been doing this for over 30 years just because you say so, just to
    resolve one persons M.O.O? you're a bigger fucking idiot than I was
    starting to think if so.

    I was guiding you nicely to the fact that you have no idea what you are talking about. I've helped build two of the largest CDNs in the world,
    and you do it by using the Internet Protocols as designed never hardware
    load balancers which cannot scale and are additional failure domain.



    All in all its just common courtesy to give your peers a heads up, its
    fine if Jesse no longer wants peering, jesus, it took me 3 weeks to
    discover he was MIA courtesy of the pending articles list file hitting
    10Mb which triggered the email alert, anyways he can either accept the
    advice or ignore it, the sun will still rise in the morning.

    Maybe look at the bigger picture of what he is doing and that most of
    use are cramming administration into active life. Your server your
    rules, so to each their own.

    ohh wow really, and i thought running my organisation 15 hours a day 6
    days a week was because i was bored and wanted to avoid the family..
    thats where i went wrong.. jesus fucking christ. Jesse miught lack common courtesy, but you lack common sense, and thats worse.
    dont waste your time trolling me again.

    The mercurial reaction to your own Dunning-Kruger limits is at least
    comical. If you don't want peers, drop them. You're the one making the spectacle and an ass out of yourself.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Eddie Martin@ejm@nospammy.myplace.de to news.admin.peering on Sun Aug 2 10:50:26 2026
    From Newsgroup: news.admin.peering

    On 2/08/2026 9:25 am, Kevin Bowling wrote:
    On 8/1/26 06:54, noel wrote:
    On Sat, 01 Aug 2026 04:50:28 -0700, Kevin Bowling wrote:


    Many larger providers use multiple A records for good reasons, it's

    yeah, they're too cheap to use what the big boys use hardware load
    balancers

    valid to design around articles flooding into networks and you really
    shouldn't put a load balancer in front of feeders.. just use the

    /me shakes head

    protocols effectively.-a Inferring that there must be a relationship to
    IPs or that an operator is forever bound to one set of systems/PoPs
    doesn't reflect how these systems are run.-a You need to have proper rDNS >>
    thats contrary to my experience in the real world of running ISP's and
    carriers for over 30 years, crap thats over 35 years now actiually

    set up and the 3 major news server software handle this per design.


    whoooooooooooppppieeeeeeeeeeeeeeee who cares, are you trying to tell me i
    should adjust the way i do things, the same way national carriers have
    been doing this for over 30 years just because you say so, just to
    resolve one persons M.O.O? you're a bigger fucking idiot than I was
    starting to think if so.

    I was guiding you nicely to the fact that you have no idea what you are talking about.-a I've helped build two of the largest CDNs in the world,
    and you do it by using the Internet Protocols as designed never hardware load balancers which cannot scale and are additional failure domain.



    Good grief, I can't remain silent in this who's got the bigger dick
    contest any longer, how you do things today is nothing like how it was
    done before, having worked at high levels in ATT for decades (thankfully
    I am recently retired so I no longer need to deal with rubbish like
    this), hardware load balancers have long and are still very common
    today, or were 7 months ago anyway. So what if you have helped build
    (I've never heard your name before a couple months ago when you started contributing to INN2)a CDN, a CDN is a CDN, it solves a mostly
    different problem, or perhaps your CDN was at Alphabet or Facebook, it
    would explain your arrogances.


    All in all its just common courtesy to give your peers a heads up, its >>>> fine if Jesse no longer wants peering, jesus, it took me 3 weeks to
    discover he was MIA courtesy of the pending articles list file hitting >>>> 10Mb which triggered the email alert, anyways he can either accept the >>>> advice or ignore it, the sun will still rise in the morning.

    Maybe look at the bigger picture of what he is doing and that most of
    use are cramming administration into active life.-a Your server your
    rules, so to each their own.

    ohh wow really, and i thought running my organisation 15 hours a day 6
    days a week was because i was bored and wanted to avoid the family..
    thats where i went wrong.. jesus fucking christ. Jesse miught lack common
    courtesy, but you lack common sense, and thats worse.
    dont waste your time trolling me again.

    The mercurial reaction to your own Dunning-Kruger limits is at least comical.-a If you don't want peers, drop them.-a You're the one making the spectacle and an ass out of yourself.

    On the contrary you all are making an ass out of yourselves. Both you
    and Noel need time out, Jesse does need to learn if he wants to provide
    a service, he needs to consider everyone he provides services to, or
    request they use hostnames or thank but decline to peer, yes, stuff
    happens, but it takes 30 seconds to post a note in here even after the
    fact saying "hey guys" he instead probably spent more energy attacking
    Noel then he did in putting in thought to how his consequences affect
    other peers (and now prospective other peers who see how they get
    treated), and yes Noel should have fired an Email to Jesse instead of
    venting here, it had merit, but should have been raised in private.
    Although as he alludes, maybe in jest, that he wont be using Jesse's or anyone's peering shortly either way, a shame to see an apparently long
    time service pulled because of usenet admin squabbles, and I find it
    strange that Kevin thinks he can speak to this where from what I can
    tell, he does not and has never run an NNTP server, let alone service,
    so not sure where he thinks he can speak on this matter, only one who
    has run them for a substantial amount of time actually understands,
    perhaps his misguided beliefs that CDN and usenet are the same.

    You all need time out to think and reflect.

    Enjoy your weekend.



    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Kevin Bowling@kevin.bowling@kev009.com to news.admin.peering on Sat Aug 1 18:19:15 2026
    From Newsgroup: news.admin.peering

    On 8/1/26 17:50, Eddie Martin wrote:
    On 2/08/2026 9:25 am, Kevin Bowling wrote:
    On 8/1/26 06:54, noel wrote:
    On Sat, 01 Aug 2026 04:50:28 -0700, Kevin Bowling wrote:


    Many larger providers use multiple A records for good reasons, it's

    yeah, they're too cheap to use what the big boys use hardware load
    balancers

    valid to design around articles flooding into networks and you really
    shouldn't put a load balancer in front of feeders.. just use the

    /me shakes head

    protocols effectively.-a Inferring that there must be a relationship to >>>> IPs or that an operator is forever bound to one set of systems/PoPs
    doesn't reflect how these systems are run.-a You need to have proper
    rDNS

    thats contrary to my experience in the real world of running ISP's and
    carriers for over 30 years, crap thats over 35 years now actiually

    set up and the 3 major news server software handle this per design.


    whoooooooooooppppieeeeeeeeeeeeeeee who cares, are you trying to tell
    me i
    should adjust the way i do things, the same way national carriers have
    been doing this for over 30 years just because you say so, just to
    resolve one persons M.O.O? you're a bigger fucking idiot than I was
    starting to think if so.

    I was guiding you nicely to the fact that you have no idea what you
    are talking about.-a I've helped build two of the largest CDNs in the
    world, and you do it by using the Internet Protocols as designed never
    hardware load balancers which cannot scale and are additional failure
    domain.



    Good grief, I can't remain silent in this who's got the bigger dick
    contest any longer, how you do things today is nothing like how it was
    done before, having worked at high levels in ATT for decades (thankfully
    I am recently retired so I no longer need to deal with rubbish like
    this), hardware load balancers have long and are still very common
    today, or were 7 months ago anyway. So what if you have helped build
    (I've never heard your name before a couple months ago when you started contributing to INN2)a CDN, a CDN is a CDN, it solves a-a mostly
    different problem, or perhaps your CDN was at Alphabet or Facebook, it
    would explain your arrogances.

    The Internet Protocols are the same, only the L7 differs. Inbound NNTP
    is basically the worst possible place to use a load balancer, you just
    have the feeders peer with each other and outward and everything flood
    fills without any additional complexity. Reader servers might benefit
    from a sticky session LB, and a _proxy_ (not LB) would be less
    ridiculous for inbound if it was providing routing or message oracle for
    some distributed backend.


    All in all its just common courtesy to give your peers a heads up, its >>>>> fine if Jesse no longer wants peering, jesus, it took me 3 weeks to
    discover he was MIA courtesy of the pending articles list file hitting >>>>> 10Mb which triggered the email alert, anyways he can either accept the >>>>> advice or ignore it, the sun will still rise in the morning.

    Maybe look at the bigger picture of what he is doing and that most of
    use are cramming administration into active life.-a Your server your
    rules, so to each their own.

    ohh wow really, and i thought running my organisation 15 hours a day 6
    days a week was because i was bored and wanted to avoid the family..
    thats where i went wrong.. jesus fucking christ. Jesse miught lack
    common
    courtesy, but you lack common sense, and thats worse.
    dont waste your time trolling me again.

    The mercurial reaction to your own Dunning-Kruger limits is at least
    comical.-a If you don't want peers, drop them.-a You're the one making
    the spectacle and an ass out of yourself.

    On the contrary you all are making an ass out of yourselves. Both you
    and Noel need time out, Jesse does need to learn if he wants to provide
    a service, he needs to consider everyone he provides services to, or
    request they use hostnames or thank but decline to peer, yes, stuff
    happens, but it takes 30 seconds to post a note in here even after the
    fact saying "hey guys" he instead probably spent more energy attacking
    Noel then he did in putting in thought to how his consequences affect
    other peers (and now prospective other peers who see how they get
    treated), and yes Noel should have fired an Email to Jesse instead of venting here, it had merit, but should have been raised in private.
    Although as he alludes, maybe in jest, that he wont be using Jesse's or anyone's peering shortly either way, a shame to see an apparently long
    time service pulled because of usenet admin squabbles, and I find it
    strange that Kevin thinks he can speak to this where from what I can
    tell, he does not and has never run an NNTP server, let alone service,
    so not sure where he thinks he can speak on this matter, only one who
    has run them for a substantial amount of time actually understands,
    perhaps his misguided beliefs that CDN and usenet are the same.

    Another hyperbolic joker. I've run a peer for 15 years, not hard to
    observe.

    You all need time out to think and reflect.

    Enjoy your weekend.




    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Eddie Martin@ejm@nospammy.myplace.de to news.admin.peering on Sun Aug 2 13:29:17 2026
    From Newsgroup: news.admin.peering

    On 2/08/2026 11:19 am, Kevin Bowling wrote:
    On 8/1/26 17:50, Eddie Martin wrote:
    On 2/08/2026 9:25 am, Kevin Bowling wrote:
    On 8/1/26 06:54, noel wrote:
    On Sat, 01 Aug 2026 04:50:28 -0700, Kevin Bowling wrote:


    Many larger providers use multiple A records for good reasons, it's

    yeah, they're too cheap to use what the big boys use hardware load
    balancers

    valid to design around articles flooding into networks and you really >>>>> shouldn't put a load balancer in front of feeders.. just use the

    /me shakes head

    protocols effectively.-a Inferring that there must be a relationship to >>>>> IPs or that an operator is forever bound to one set of systems/PoPs
    doesn't reflect how these systems are run.-a You need to have proper >>>>> rDNS

    thats contrary to my experience in the real world of running ISP's and >>>> carriers for over 30 years, crap thats over 35 years now actiually

    set up and the 3 major news server software handle this per design.


    whoooooooooooppppieeeeeeeeeeeeeeee who cares, are you trying to tell
    me i
    should adjust the way i do things, the same way national carriers have >>>> been doing this for over 30 years just because you say so, just to
    resolve one persons M.O.O? you're a bigger fucking idiot than I was
    starting to think if so.

    I was guiding you nicely to the fact that you have no idea what you
    are talking about.-a I've helped build two of the largest CDNs in the
    world, and you do it by using the Internet Protocols as designed
    never hardware load balancers which cannot scale and are additional
    failure domain.



    Good grief, I can't remain silent in this who's got the bigger dick
    contest any longer, how you do things today is nothing like how it was
    done before, having worked at high levels in ATT for decades
    (thankfully I am recently retired so I no longer need to deal with
    rubbish like this), hardware load balancers have long and are still
    very common today, or were 7 months ago anyway. So what if you have
    helped build (I've never heard your name before a couple months ago
    when you started contributing to INN2)a CDN, a CDN is a CDN, it solves
    a-a mostly different problem, or perhaps your CDN was at Alphabet or
    Facebook, it would explain your arrogances.

    The Internet Protocols are the same, only the L7 differs.-a Inbound NNTP
    is basically the worst possible place to use a load balancer, you just
    have the feeders peer with each other and outward and everything flood
    fills without any additional complexity.-a Reader servers might benefit
    from a sticky session LB, and a _proxy_ (not LB) would be less
    ridiculous for inbound if it was providing routing or message oracle for some distributed backend.


    All in all its just common courtesy to give your peers a heads up, >>>>>> its
    fine if Jesse no longer wants peering, jesus, it took me 3 weeks to >>>>>> discover he was MIA courtesy of the pending articles list file
    hitting
    10Mb which triggered the email alert, anyways he can either accept >>>>>> the
    advice or ignore it, the sun will still rise in the morning.

    Maybe look at the bigger picture of what he is doing and that most of >>>>> use are cramming administration into active life.-a Your server your >>>>> rules, so to each their own.

    ohh wow really, and i thought running my organisation 15 hours a day 6 >>>> days a week was because i was bored and wanted to avoid the family..
    thats where i went wrong.. jesus fucking christ. Jesse miught lack
    common
    courtesy, but you lack common sense, and thats worse.
    dont waste your time trolling me again.

    The mercurial reaction to your own Dunning-Kruger limits is at least
    comical.-a If you don't want peers, drop them.-a You're the one making
    the spectacle and an ass out of yourself.

    On the contrary you all are making an ass out of yourselves. Both you
    and Noel need time out, Jesse does need to learn if he wants to
    provide a service, he needs to consider everyone he provides services
    to, or request they use hostnames or thank but decline to peer, yes,
    stuff happens, but it takes 30 seconds to post a note in here even
    after the fact saying "hey guys" he instead probably spent more energy
    attacking Noel then he did in putting in thought to how his
    consequences affect other peers (and now prospective other peers who
    see how they get treated), and yes Noel should have fired an Email to
    Jesse instead of venting here, it had merit, but should have been
    raised in private. Although as he alludes, maybe in jest, that he wont
    be using Jesse's or anyone's peering shortly either way, a shame to
    see an apparently long time service pulled because of usenet admin
    squabbles, and I find it strange that Kevin thinks he can speak to
    this where from what I can tell, he does not and has never run an NNTP
    server, let alone service, so not sure where he thinks he can speak on
    this matter, only one who has run them for a substantial amount of
    time actually understands, perhaps his misguided beliefs that CDN and
    usenet are the same.

    Another hyperbolic joker.-a I've run a peer for 15 years, not hard to observe.



    You are right Kevin, I apologize, I should have taken a moment to check
    you out, I never really cared before.

    Kevin, we have a problem, there appears no evidence you had any part in
    a CDN development (anyone can set up reverse proxies, ATS, caches, etc),
    you do however run a small MSP computer and software design business in
    Chino Valley, despite no real world peering presence with no AS in the
    routing databases, a cursory look shows you do lease a /29 prefix, but
    Kevin, you need a minimum of a /24 prefix for BGP, and anyone running
    CDN's have decent IPAS, anything smaller and networks discard the route
    opting for the next up, unlike with csiph usenet I see you run Kevin,
    where you could do that on dialup with single IP, so perhaps I am
    starting to see Kevin, why you dislike networks using IP's and want them
    to use host names, so much.

    Another thing puzzles me Kevin, since you are the CDN expert, why is
    there no talks at NANOG, why not share that all encompassing knowledge
    to those mere mortals beneath you, I see you only graduated college in
    2011, so I understand why you are confused and can not comprehend how
    large networks are actually designed and run, since you were not even
    born when some people here were running them, however Kevin, I can
    respect the high bar you've set for yourself, I really do, and perhaps
    one day, you'll get there, but I suspect because you detest humble pie,
    it's a bit off yet, but that's not a real problem Kevin, as you're still
    a young fellow.





    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Kevin Bowling@kevin.bowling@kev009.com to news.admin.peering on Sat Aug 1 21:02:23 2026
    From Newsgroup: news.admin.peering

    On 8/1/26 20:29, Eddie Martin wrote:
    On 2/08/2026 11:19 am, Kevin Bowling wrote:
    On 8/1/26 17:50, Eddie Martin wrote:
    On 2/08/2026 9:25 am, Kevin Bowling wrote:
    On 8/1/26 06:54, noel wrote:
    On Sat, 01 Aug 2026 04:50:28 -0700, Kevin Bowling wrote:


    Many larger providers use multiple A records for good reasons, it's >>>>>
    yeah, they're too cheap to use what the big boys use hardware load
    balancers

    valid to design around articles flooding into networks and you really >>>>>> shouldn't put a load balancer in front of feeders.. just use the

    /me shakes head

    protocols effectively.-a Inferring that there must be a
    relationship to
    IPs or that an operator is forever bound to one set of systems/PoPs >>>>>> doesn't reflect how these systems are run.-a You need to have
    proper rDNS

    thats contrary to my experience in the real world of running ISP's and >>>>> carriers for over 30 years, crap thats over 35 years now actiually

    set up and the 3 major news server software handle this per design. >>>>>

    whoooooooooooppppieeeeeeeeeeeeeeee who cares, are you trying to
    tell me i
    should adjust the way i do things, the same way national carriers have >>>>> been doing this for over 30 years just because you say so, just to
    resolve one persons M.O.O? you're a bigger fucking idiot than I was
    starting to think if so.

    I was guiding you nicely to the fact that you have no idea what you
    are talking about.-a I've helped build two of the largest CDNs in the >>>> world, and you do it by using the Internet Protocols as designed
    never hardware load balancers which cannot scale and are additional
    failure domain.



    Good grief, I can't remain silent in this who's got the bigger dick
    contest any longer, how you do things today is nothing like how it
    was done before, having worked at high levels in ATT for decades
    (thankfully I am recently retired so I no longer need to deal with
    rubbish like this), hardware load balancers have long and are still
    very common today, or were 7 months ago anyway. So what if you have
    helped build (I've never heard your name before a couple months ago
    when you started contributing to INN2)a CDN, a CDN is a CDN, it
    solves a-a mostly different problem, or perhaps your CDN was at
    Alphabet or Facebook, it would explain your arrogances.

    The Internet Protocols are the same, only the L7 differs.-a Inbound
    NNTP is basically the worst possible place to use a load balancer, you
    just have the feeders peer with each other and outward and everything
    flood fills without any additional complexity.-a Reader servers might
    benefit from a sticky session LB, and a _proxy_ (not LB) would be less
    ridiculous for inbound if it was providing routing or message oracle
    for some distributed backend.


    All in all its just common courtesy to give your peers a heads
    up, its
    fine if Jesse no longer wants peering, jesus, it took me 3 weeks to >>>>>>> discover he was MIA courtesy of the pending articles list file
    hitting
    10Mb which triggered the email alert, anyways he can either
    accept the
    advice or ignore it, the sun will still rise in the morning.

    Maybe look at the bigger picture of what he is doing and that most of >>>>>> use are cramming administration into active life.-a Your server your >>>>>> rules, so to each their own.

    ohh wow really, and i thought running my organisation 15 hours a day 6 >>>>> days a week was because i was bored and wanted to avoid the family.. >>>>> thats where i went wrong.. jesus fucking christ. Jesse miught lack
    common
    courtesy, but you lack common sense, and thats worse.
    dont waste your time trolling me again.

    The mercurial reaction to your own Dunning-Kruger limits is at least
    comical.-a If you don't want peers, drop them.-a You're the one making >>>> the spectacle and an ass out of yourself.

    On the contrary you all are making an ass out of yourselves. Both you
    and Noel need time out, Jesse does need to learn if he wants to
    provide a service, he needs to consider everyone he provides services
    to, or request they use hostnames or thank but decline to peer, yes,
    stuff happens, but it takes 30 seconds to post a note in here even
    after the fact saying "hey guys" he instead probably spent more
    energy attacking Noel then he did in putting in thought to how his
    consequences affect other peers (and now prospective other peers who
    see how they get treated), and yes Noel should have fired an Email to
    Jesse instead of venting here, it had merit, but should have been
    raised in private. Although as he alludes, maybe in jest, that he
    wont be using Jesse's or anyone's peering shortly either way, a shame
    to see an apparently long time service pulled because of usenet admin
    squabbles, and I find it strange that Kevin thinks he can speak to
    this where from what I can tell, he does not and has never run an
    NNTP server, let alone service, so not sure where he thinks he can
    speak on this matter, only one who has run them for a substantial
    amount of time actually understands, perhaps his misguided beliefs
    that CDN and usenet are the same.

    Another hyperbolic joker.-a I've run a peer for 15 years, not hard to
    observe.



    You are right Kevin, I apologize, I should have taken a moment to check
    you out, I never really cared before.

    Kevin, we have a problem, there appears no evidence you had any part in
    a CDN development (anyone can set up reverse proxies, ATS, caches, etc),
    you do however run a small MSP computer and software design business in Chino Valley, despite no real world peering presence with no AS in the routing databases, a cursory look shows you do lease a /29 prefix, but Kevin, you need a minimum of a /24 prefix for BGP, and anyone running
    CDN's have decent IPAS, anything smaller and networks discard the route opting for the next up, unlike with csiph usenet I see you run Kevin,
    where you could do that on dialup with single IP, so perhaps I am
    starting to see Kevin, why you dislike networks using IP's and want them
    to use host names, so much.

    csiph is not a CDN nor claimed to be, it's a single PoP done on my own
    time and dime for 15 years. And also trivially observable currently
    serving around 10-30 million articles a day, making it one of the
    busiest free text readers. Meanwhile you are using someone else's system.


    Another thing puzzles me Kevin, since you are the CDN expert, why is
    there no talks at NANOG, why not share that all encompassing knowledge
    to those mere mortals beneath you, I see you only graduated college in
    2011, so I understand why you are confused and can not comprehend how
    large networks are actually designed and run, since you were not even
    born when some people here were running them, however Kevin, I can
    respect the high bar you've set for yourself, I really do, and perhaps
    one day, you'll get there, but I suspect because you detest humble pie,
    it's a bit off yet, but that's not a real problem Kevin, as you're still
    a young fellow.

    Multiple public talks, again trivial to find. NANOG is concentrated
    with people that configure and operate network devices, not software developers.

    You're going off again on things you have no idea about, convinced you
    are right with your Dunning-Krueger twin.

    The root point remains, if you have some notion of how a peer should
    look to peer with you, deal with it privately or simply disconnect them.
    Making a public spectacle with poor technical judgement to attempt to
    shame people is sheer idiocy.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Marco Moock@mm@dorfdsl.de to news.admin.peering on Sun Aug 2 06:41:29 2026
    From Newsgroup: news.admin.peering

    Am 02.08.26 um 03:19 schrieb Kevin Bowling:

    The Internet Protocols are the same, only the L7 differs.-a Inbound NNTP
    is basically the worst possible place to use a load balancer, you just
    have the feeders peer with each other and outward and everything flood
    fills without any additional complexity.-a Reader servers might benefit
    from a sticky session LB, and a _proxy_ (not LB) would be less
    ridiculous for inbound if it was providing routing or message oracle for some distributed backend.

    Is a reader server for text-only groups in need of a load balancer or
    can current hardware handle all the sessions?
    --
    Gru|f
    Marco

    Junk-Mail bitte an trashcan@stinkedores.dorfdsl.de
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From J@J@M to news.admin.peering on Sun Aug 2 06:47:49 2026
    From Newsgroup: news.admin.peering

    On Sat, 1 Aug 2026 18:19:15 -0700, Kevin Bowling wrote:
    On 8/1/26 17:50, Eddie Martin wrote:
    On 2/08/2026 9:25 am, Kevin Bowling wrote:
    On 8/1/26 06:54, noel wrote:
    On Sat, 01 Aug 2026 04:50:28 -0700, Kevin Bowling wrote:
    Many larger providers use multiple A records for good reasons, it's

    snips

    The Internet Protocols are the same, only the L7 differs. Inbound NNTP
    is basically the worst possible place to use a load balancer, you just
    have the feeders peer with each other and outward and everything flood
    fills without any additional complexity. Reader servers might benefit
    from a sticky session LB, and a _proxy_ (not LB) would be less
    ridiculous for inbound if it was providing routing or message oracle for some distributed backend.

    All in all its just common courtesy to give your peers a heads up, its >>>>>> fine if Jesse no longer wants peering, jesus, it took me 3 weeks to >>>>>> discover he was MIA courtesy of the pending articles list file hitting >>>>>> 10Mb which triggered the email alert, anyways he can either accept the >>>>>> advice or ignore it, the sun will still rise in the morning.

    Maybe look at the bigger picture of what he is doing and that most of >>>>> use are cramming administration into active life.a Your server your
    rules, so to each their own.

    ohh wow really, and i thought running my organisation 15 hours a day 6 >>>> days a week was because i was bored and wanted to avoid the family..
    thats where i went wrong.. jesus fucking christ. Jesse miught lack
    common
    courtesy, but you lack common sense, and thats worse.
    dont waste your time trolling me again.

    The mercurial reaction to your own Dunning-Kruger limits is at least
    comical.a If you don't want peers, drop them.a You're the one making
    the spectacle and an ass out of yourself.

    On the contrary you all are making an ass out of yourselves. Both you
    and Noel need time out, Jesse does need to learn if he wants to provide
    a service, he needs to consider everyone he provides services to, or
    request they use hostnames or thank but decline to peer, yes, stuff
    happens, but it takes 30 seconds to post a note in here even after the
    fact saying "hey guys" he instead probably spent more energy attacking
    Noel then he did in putting in thought to how his consequences affect
    other peers (and now prospective other peers who see how they get
    treated), and yes Noel should have fired an Email to Jesse instead of
    venting here, it had merit, but should have been raised in private.
    Although as he alludes, maybe in jest, that he wont be using Jesse's or
    anyone's peering shortly either way, a shame to see an apparently long
    time service pulled because of usenet admin squabbles, and I find it
    strange that Kevin thinks he can speak to this where from what I can
    tell, he does not and has never run an NNTP server, let alone service,
    so not sure where he thinks he can speak on this matter, only one who
    has run them for a substantial amount of time actually understands,
    perhaps his misguided beliefs that CDN and usenet are the same.

    Another hyperbolic joker. I've run a peer for 15 years, not hard to observe.

    myself an amateur but devoted user of usenet newsgroups, this search:

    https://csiph.com/search?q=from%3A%22Eddie+Martin%22&group=&after=&before=&sort=relevance&lang=

    would suggest that only two posts within the past decade (10.8 years),
    and using this "stage name"-style alias . . . if it looks like a duck,
    or quacks like ai-enhanced trolling-chatbots, but maybe these popular
    usenet newsservers, in this case, blueworld and csiph, both with very
    long and complete article retention (especially bw, mostly from 1981),
    are helpful and useful to the public, which is anathema to (redacted)

    such trolling seems strangely out of place even in unmoderated groups,
    as if the common practice to ignore trolls has become the global norm,
    so they're taunting system administrators, to help the chatbots learn,
    and they've been at it since circa 1993, their agenda is causing gigo





































    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Kevin Bowling@kevin.bowling@kev009.com to news.admin.peering on Sat Aug 1 22:11:43 2026
    From Newsgroup: news.admin.peering

    On 8/1/26 21:41, Marco Moock wrote:
    Am 02.08.26 um 03:19 schrieb Kevin Bowling:

    The Internet Protocols are the same, only the L7 differs.-a Inbound
    NNTP is basically the worst possible place to use a load balancer, you
    just have the feeders peer with each other and outward and everything
    flood fills without any additional complexity.-a Reader servers might
    benefit from a sticky session LB, and a _proxy_ (not LB) would be less
    ridiculous for inbound if it was providing routing or message oracle
    for some distributed backend.

    Is a reader server for text-only groups in need of a load balancer or
    can current hardware handle all the sessions?


    For the most part a single system can handle it if designed for. It
    might be a little tight for a place like Eternal September that might
    see 10k idle connections.

    The current issue with inn specifically is that it forks a new nnrpd
    process for every reader connection. That means each connection costs
    many megabytes of RAM, when the actual weight of a connection could be kilobytes. The way to solve this is to make the connections cheap, with
    an event loop, and have nnrpd service many connections as they notify
    (via the kernel) of read/write readiness. But there is a lot of work to
    get there, and you still want some fan out to processes (or threads) to continue taking advantage of multiple CPU cores.

    In the mean time something like
    https://github.com/mjc/nntp-proxy/tree/main would be helpful to an inn
    reader deployment where idle connections sit on a cheap event loop, and
    are multiplexed back to the expensive nnrpd processes. Note that I
    haven't looked deeply at that code so I don't know how complete or
    effective it is. To highlight the complexity, csiph grew a fairly sophisticated connection pooler for similar reasons (the complexity is
    to deal with things like a WWW request aborting.. I want the NNTP conn
    to stay alive and be in a known good state so it finishes the NNTP
    transaction anyway which is cheaper than tearing down and setting up a
    new conn).


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Ivo Gandolfo@usenet@bofh.team to news.admin.peering on Sun Aug 2 12:54:01 2026
    From Newsgroup: news.admin.peering

    Il 02/08/2026 01:04, noel ha scritto:
    ohh so its my fault i come from an ISP world - remember us, back in teh
    day we were the only ones providing usenet, now any diskhead can lease a
    VPS and spin one up.

    my bad, then i might as well not only terminate jesse's peer, I might as well terminate all of them, I was looking for hardware to expand our
    mirror service, now I found it.

    noel,

    for the record I also come from the ISP world. I worked in that
    environment for years, and today I work in R&D, so I am reasonably
    familiar both with how things used to be done and with the inconvenient
    modern habit of occasionally updating onerCOs knowledge.

    Yes, there was a time when Usenet was largely operated by ISPs. Hardware
    was expensive, connectivity was limited, and ordinary people could not
    simply rent a server and run services themselves.

    It must have been a comforting arrangement. A small number of
    organisations had the infrastructure, the expertise and, apparently, the exclusive right to decide who was sufficiently qualified to participate.

    Unfortunately, the calendar now says 2026.

    Today, hardware is accessible, software is widely available,
    documentation exists, and even the dreaded rCLdiskheadsrCY can operate
    systems that were once the exclusive territory of ISPs. Whether that is pleasant for those nostalgic for the old monopoly is beside the point. Technology moved on without requesting permission.

    Coming from the ISP world does not provide a permanent exemption from
    reading current documentation, checking RFCs, or updating
    configurations. Past experience is certainly valuable, but it is not a lifetime support contract for obsolete assumptions.

    As for terminating JesserCOs peer rCoor perhaps all of themrCo because someone pointed out that your configuration was wrong, that is of course your
    choice. They are your machines, your bandwidth and your peers.

    But announcing that you may shut everything down because the discussion
    did not go your way does not sound particularly like the measured
    decision of an experienced network operator, it sounds more like the
    child who owns the football announcing that the match is over because
    the other children had the audacity not to play according to his
    personal rules.

    rCLI own the ball, therefore I winrCY may work very well in a playground. It is a rather less impressive principle for operating network infrastructure.

    If you genuinely no longer want to provide the service, then shut it
    down. Nobody can force you to continue. But dressing up a tantrum as an infrastructure decision does not make it technical, and punishing
    unrelated peers will not retroactively make your original claim correct.

    Nobody is disputing that you have experience. The issue is whether that experience is still being used to understand how things work today, or
    merely displayed as a museum exhibit from the era when ISPs believed
    they owned Usenet.

    The past deserves to be remembered. It does not need to be preserved indefinitely in production configurations or used as an excuse whenever reality fails to remain compatible with it.


    Sincerely
    --
    Ivo Gandolfo

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Ivo Gandolfo@usenet@bofh.team to news.admin.peering on Sun Aug 2 13:02:34 2026
    From Newsgroup: news.admin.peering

    Il 02/08/2026 06:41, Marco Moock ha scritto:
    Is a reader server for text-only groups in need of a load balancer or
    can current hardware handle all the sessions?


    I recompiled INN to have more FD available, but as Kevin told you it's
    more of a CPU/RAM/SSD issue these days.

    My experience over the years has led me to believe that if you have a
    fairly large user base, exposing the reader directly is
    counterproductive unless you have very high-performance hardware.
    In fact, in anticipation of expanding my service, I almost rewrote an nntp-proxy service from scratch.
    The INN servers will remain only as a backend and history server, while
    the nntp frontend will take care of everything else leaving me plenty of options for caching and storing everything in the database for very fast searches.

    I've currently seen several read timeouts on my servers, mostly due to
    the extensive searches performed. If everything is indexed a priori, the result is almost instantaneous.


    Sincerely
    --
    Ivo Gandolfo
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nomen Nescio@nobody@dizum.com to news.admin.peering on Sun Aug 2 11:45:13 2026
    From Newsgroup: news.admin.peering

    noel <deletethis@invalid.lan> wrote:

    ohh so its my fault i come from an ISP world - remember us, back in teh
    day we were the only ones providing usenet, now any diskhead can lease a
    VPS and spin one up.

    So your 35-year old paradigms don't necessarily apply anymore.

    my bad, then i might as well not only terminate jesse's peer, I might as well terminate all of them, I was looking for hardware to expand our
    mirror service, now I found it.

    k, bye?
    What a weird flex.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From noel@deletethis@invalid.lan to news.admin.peering on Tue Aug 4 08:23:59 2026
    From Newsgroup: news.admin.peering

    On Sun, 02 Aug 2026 12:54:01 +0200, Ivo Gandolfo wrote:


    noel,

    for the record I also come from the ISP world. I worked in that
    environment for years, and today I work in R&D, so I am reasonably
    familiar both with how things used to be done and with the inconvenient modern habit of occasionally updating onerCOs knowledge.


    I dont see the need to change here where everything else runs smooth like that, I look at usenet once in a while when Im bored, I guess all those
    mail servers should have the backend shutdown and converted to hostname
    CDN eh, nah, the load balancers do their job their too, I'm not going to
    fux, whats not fuxed.

    Yes, there was a time when Usenet was largely operated by ISPs. Hardware
    was expensive,

    considering the HP and Dell servers we use are 15K each, I'd argue still
    are :) But yeah, these days anyone can put a 100 dollar refurbed PC with
    a daemon on it.


    It must have been a comforting arrangement. A small number of

    It was actually incredibly involved to get peering back then, wasn't an overight Email or contact form request thing and your up and running.

    organisations had the infrastructure, the expertise and, apparently, the exclusive right to decide who was sufficiently qualified to participate.


    and recall how litle garbage there was back then compared to today,
    sometimes change is a good thing, but sometimes its not a good thing.


    Unfortunately, the calendar now says 2026.

    Which is a good thing for me, since it means almost retirement.


    Coming from the ISP world does not provide a permanent exemption from
    reading current documentation, checking RFCs, or updating

    I must have missed teh RFC that says thou "MUST" use hostnames in usenet,
    got that RFC No. handy? Might be worth a 5 minute read with my coffee.

    configurations. Past experience is certainly valuable, but it is not a lifetime support contract for obsolete assumptions.

    But it also shows we learn from our mistakes and dont change what works, certainly not just because some kids who wouldnt survive in a NOC today
    thinks you're doing it wrong, because you surely are not so naive to
    think I'm the only one running this way, Eddie from ATT (AT&T ? or is
    there a closely named ATT as well?) appears to show I'm not, and if AT&T,
    they are one hell of a lot bigger and shows I'm not an outlier, and if we
    are, then i'm in some damn fine company lol.

    As for terminating JesserCOs peer rCoor perhaps all of themrCo because someone
    pointed out that your configuration was wrong, that is of course your
    choice. They are your machines, your bandwidth and your peers.

    No, my configuration is not wrong, it's different, especially to what
    kids of today think because when they started running these out of their parents family basement 10-15 years ago they could only get one IP, so
    had to share it.

    I use what works and still works everywhere except for usenet - and only
    one usenet peer at that, do you think I'm going to change for that or
    because some snotty nosed kid who thanks to Eddie we see is not the big
    megga network man he tried to have us all beleive, he's just a small fry,
    and yes even he has a right to his beliefs, doesn't make them right, nor wrong.


    But announcing that you may shut everything down because the discussion
    did not go your way does not sound particularly like the measured
    decision of an experienced network operator, it sounds more like the
    child who owns the football announcing that the match is over because
    the other children had the audacity not to play according to his
    personal rules.


    Did not go my way? what because some kid whos been caught out trying to
    big note himself found to be full of shit, says I'm not doing it teh way
    he thinks it should be done? ROFLMFAO I don't care what he or other
    people think, the ONLY peoples whos opinion of me I care about are my
    family and friends, the opinions of others means nothing, running this
    and being an IRC oper for several decades, ensures you have very, VERY,
    thick skin :)

    And I was serious when I said our mirror service needs additional
    hardware, its 15K I save by repurposing it, I am still favouring that direction, at least I'll have less fuxwits to deal with.


    MY whole point to my OP was communication or lack thereof, jesse's made
    it clear to me what he thinks of communicating and then as good as told
    me to fuck off, at a time I was considering doing exactly that, that is
    not a case of take my bat and ball and go home, its a case of i was
    already half way across teh fields on my way when he yelled out.

    That's all I have to say on teh matter, other flametards no doubt will continue becasue thats what kids caught lying do. Cheers Ivo.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Jesse Rehmer@jesse.rehmer@blueworldhosting.com to news.admin.peering on Tue Aug 4 02:34:17 2026
    From Newsgroup: news.admin.peering

    On Aug 3, 2026 at 5:23:59rC>PM CDT, "noel" <deletethis@invalid.lan> wrote:

    I look at usenet once in a while when Im bored

    Me too, that's why I didn't notice one peer out of ~35 not working.

    MY whole point to my OP was communication or lack thereof, jesse's made
    it clear to me what he thinks of communicating and then as good as told
    me to fuck off, at a time I was considering doing exactly that, that is
    not a case of take my bat and ball and go home, its a case of i was
    already half way across teh fields on my way when he yelled out.

    You mention my lack of common sense and communication skills, but you started this thread when it should have been an e-mail.

    When I cut everything over incoming/outgoing articles were flowing within minutes. I don't have time to babysit every connection the feeder makes. Eventually, I'd notice and would have sent an e-mail.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Kevin Bowling@kevin.bowling@kev009.com to news.admin.peering on Thu Aug 6 00:12:11 2026
    From Newsgroup: news.admin.peering

    On 8/3/26 15:23, noel wrote:
    On Sun, 02 Aug 2026 12:54:01 +0200, Ivo Gandolfo wrote:


    noel,

    for the record I also come from the ISP world. I worked in that
    environment for years, and today I work in R&D, so I am reasonably
    familiar both with how things used to be done and with the inconvenient
    modern habit of occasionally updating onerCOs knowledge.


    I dont see the need to change here where everything else runs smooth like that, I look at usenet once in a while when Im bored, I guess all those
    mail servers should have the backend shutdown and converted to hostname
    CDN eh, nah, the load balancers do their job their too, I'm not going to
    fux, whats not fuxed.

    Yes, there was a time when Usenet was largely operated by ISPs. Hardware
    was expensive,

    considering the HP and Dell servers we use are 15K each, I'd argue still
    are :) But yeah, these days anyone can put a 100 dollar refurbed PC with
    a daemon on it.


    It must have been a comforting arrangement. A small number of

    It was actually incredibly involved to get peering back then, wasn't an overight Email or contact form request thing and your up and running.

    organisations had the infrastructure, the expertise and, apparently, the
    exclusive right to decide who was sufficiently qualified to participate.


    and recall how litle garbage there was back then compared to today,
    sometimes change is a good thing, but sometimes its not a good thing.


    Unfortunately, the calendar now says 2026.

    Which is a good thing for me, since it means almost retirement.


    Coming from the ISP world does not provide a permanent exemption from
    reading current documentation, checking RFCs, or updating

    I must have missed teh RFC that says thou "MUST" use hostnames in usenet,
    got that RFC No. handy? Might be worth a 5 minute read with my coffee.

    configurations. Past experience is certainly valuable, but it is not a
    lifetime support contract for obsolete assumptions.

    But it also shows we learn from our mistakes and dont change what works, certainly not just because some kids who wouldnt survive in a NOC today thinks you're doing it wrong, because you surely are not so naive to
    think I'm the only one running this way, Eddie from ATT (AT&T ? or is
    there a closely named ATT as well?) appears to show I'm not, and if AT&T, they are one hell of a lot bigger and shows I'm not an outlier, and if we are, then i'm in some damn fine company lol.

    As for terminating JesserCOs peer rCoor perhaps all of themrCo because someone
    pointed out that your configuration was wrong, that is of course your
    choice. They are your machines, your bandwidth and your peers.

    No, my configuration is not wrong, it's different, especially to what
    kids of today think because when they started running these out of their parents family basement 10-15 years ago they could only get one IP, so
    had to share it.

    I use what works and still works everywhere except for usenet - and only
    one usenet peer at that, do you think I'm going to change for that or
    because some snotty nosed kid who thanks to Eddie we see is not the big
    megga network man he tried to have us all beleive, he's just a small fry,
    and yes even he has a right to his beliefs, doesn't make them right, nor wrong.


    But announcing that you may shut everything down because the discussion
    did not go your way does not sound particularly like the measured
    decision of an experienced network operator, it sounds more like the
    child who owns the football announcing that the match is over because
    the other children had the audacity not to play according to his
    personal rules.


    Did not go my way? what because some kid whos been caught out trying to
    big note himself found to be full of shit, says I'm not doing it teh way
    he thinks it should be done? ROFLMFAO I don't care what he or other
    people think, the ONLY peoples whos opinion of me I care about are my
    family and friends, the opinions of others means nothing, running this
    and being an IRC oper for several decades, ensures you have very, VERY,
    thick skin :)

    You are a fool. You came to drop caustic nonsense. And not that it particularly matters to the conversation, but just further highlight
    your stupidity, it's not hard to look up my work history or commits and
    see I've done host networking at Limelight Networks, Netskope, and
    Netflix. Your Eddie sock puppet account is as dumb as you are.

    And I was serious when I said our mirror service needs additional
    hardware, its 15K I save by repurposing it, I am still favouring that direction, at least I'll have less fuxwits to deal with.


    MY whole point to my OP was communication or lack thereof, jesse's made
    it clear to me what he thinks of communicating and then as good as told
    me to fuck off, at a time I was considering doing exactly that, that is
    not a case of take my bat and ball and go home, its a case of i was
    already half way across teh fields on my way when he yelled out.

    That's all I have to say on teh matter, other flametards no doubt will continue becasue thats what kids caught lying do. Cheers Ivo.


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nigel Reed@sysop@endofthelinebbs.com to news.admin.peering on Sat Aug 15 01:47:25 2026
    From Newsgroup: news.admin.peering

    On 1 Aug 2026 09:56:44 +1000
    noel <deletethis@invalid.lan> wrote:

    Just discovered (after a 21 day grace period) an overnight alert from
    our monitoring we weren't getting anything (since July 10) from
    blueworld hosting, it seems after some checks his IP changed, and
    i'll be fucked if I can find any reference or warning that was to
    happen let alone an after the fact, all completed blah blah blah, and
    of consequence to that fail2ban blackholed them for repeatedly trying
    to send whilst getting denied :)

    Now BWH is not the first peer who's done that, and sadly wont be the
    last, but all it takes is a minute of time to send post to this group
    if you dont bother keeping an internal email list of peers to
    notify...


    I'll rmeove the ban and correct IP's in access file, if feed returns
    it returns, if not, it doesnt.

    I use ipsets for blocking IPs but all my usenet peers are in their own
    set which bypasses the blocks. It seems like BWH couldn't be bothered
    to me know they changed their IP either.

    toot@newsfeed:/news/spool/innfeed# host usenet.blueworldhosting.com usenet.blueworldhosting.com has address 76.235.89.209 toot@newsfeed:/news/spool/innfeed# ipset list |grep 76.235.89.209

    Lucky for them I don't have 76.235.89.0/24 blocked so we're still
    exchanging articles, but still, as a matter of courtesy, you need to
    let your peers know if you change your config in some way that effects
    them.
    --
    End Of The Line BBS - Plano, TX
    telnet endofthelinebbs.com 23


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Jesse Rehmer@jesse.rehmer@blueworldhosting.com to news.admin.peering on Sat Aug 15 13:17:41 2026
    From Newsgroup: news.admin.peering

    On Aug 15, 2026 at 1:47:25rC>AM CDT, "Nigel Reed" <sysop@endofthelinebbs.com> wrote:

    I use ipsets for blocking IPs but all my usenet peers are in their own
    set which bypasses the blocks. It seems like BWH couldn't be bothered
    to me know they changed their IP either.

    toot@newsfeed:/news/spool/innfeed# host usenet.blueworldhosting.com usenet.blueworldhosting.com has address 76.235.89.209 toot@newsfeed:/news/spool/innfeed# ipset list |grep 76.235.89.209

    Lucky for them I don't have 76.235.89.0/24 blocked so we're still
    exchanging articles, but still, as a matter of courtesy, you need to
    let your peers know if you change your config in some way that effects
    them.

    I've blocked your IPs, you can remove the peering configuration.

    For someone complaining about notification and visibility, your website with peering details hasn't been accessible for some time, and in the past when you've changed IPs multiple times you didn't always include the new IP information.
    --- Synchronet 3.22a-Linux NewsLink 1.2