[Bitcoin-development] DNS seeds unstable

[Bitcoin-development] DNS seeds unstable

[Andreas Schildbach]

I'm bringing this issue up again. The current Bitcoin DNS seed
infrastructure is unstable. I assume this is because of we're using a
custom DNS implementation which is not 100% compatible. There have been
bugs in the past, like a case sensitive match for the domain name.

Current state (seeds taken from bitcoinj):

mainnet:

seed.bitcoin.sipa.be			OK
dnsseed.bluematt.me			OK
dnsseed.bitcoin.dashjr.org		SERVFAIL, tried multiple ISPs
seed.bitcoinstats.com			OK

testnet:

testnet-seed.bitcoin.petertodd.org	SERVFAIL, just from Telefonica
testnet-seed.bluematt.me		OK (but only returns one node)

Note: Telefonica is one of Europe's largest ISPs.

I would try to improve DNS myself, but I'm not capable of writing C. My
"fix" would be to reimplement everything in Java -- I doubt you guys
would be happy with that.

Re: [Bitcoin-development] DNS seeds unstable

[Drak]

[-- Attachment #1: Type: text/plain, Size: 2004 bytes --]

I am sure the failure here is probably more mundane like a a service not
restarted, or not on auto restart when server is rebooted and such like.
The dns seeder works pretty efficiently in my experience. Maybe we need
more seeders and to include the ability for zone transfers so existing
seeders can spread their records to multiple servers running standard DNS
servers like bind etc.


On 15 May 2014 12:50, Andreas Schildbach <andreas@schildbach.de> wrote:

> I'm bringing this issue up again. The current Bitcoin DNS seed
> infrastructure is unstable. I assume this is because of we're using a
> custom DNS implementation which is not 100% compatible. There have been
> bugs in the past, like a case sensitive match for the domain name.
>
> Current state (seeds taken from bitcoinj):
>
> mainnet:
>
> seed.bitcoin.sipa.be                    OK
> dnsseed.bluematt.me                     OK
> dnsseed.bitcoin.dashjr.org              SERVFAIL, tried multiple ISPs
> seed.bitcoinstats.com                   OK
>
> testnet:
>
> testnet-seed.bitcoin.petertodd.org      SERVFAIL, just from Telefonica
> testnet-seed.bluematt.me                OK (but only returns one node)
>
> Note: Telefonica is one of Europe's largest ISPs.
>
> I would try to improve DNS myself, but I'm not capable of writing C. My
> "fix" would be to reimplement everything in Java -- I doubt you guys
> would be happy with that.
>
>
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform
> available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>

[-- Attachment #2: Type: text/html, Size: 3095 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Jeff Garzik]

On Thu, May 15, 2014 at 7:50 AM, Andreas Schildbach
<andreas@schildbach.de> wrote:
> I'm bringing this issue up again. The current Bitcoin DNS seed
> infrastructure is unstable. I assume this is because of we're using a
> custom DNS implementation which is not 100% compatible. There have been
> bugs in the past, like a case sensitive match for the domain name.

While I agree that the problem seeds should be fixed, it also sounds
like a problem with bitcoinj / other layers, if a malfunctioning seed
is a problem for your app.

Seeds are fundamentally a backup for peer exchange.  You are going to
run into problems if you rely 100% on seeds, always.

Further, there are multiple seeds so that we are not impacted if a
couple seeds malfunction or die.  All bitcoin apps must take this into
account.

-- 
Jeff Garzik
Bitcoin core developer and open source evangelist
BitPay, Inc.      https://bitpay.com/

Re: [Bitcoin-development] DNS seeds unstable

[Gregory Maxwell]

On Thu, May 15, 2014 at 4:50 AM, Andreas Schildbach
<andreas@schildbach.de> wrote:
> I'm bringing this issue up again. The current Bitcoin DNS seed
> infrastructure is unstable. I assume this is because of we're using a
> custom DNS implementation which is not 100% compatible. There have been
> bugs in the past, like a case sensitive match for the domain name.

If software is using the DNS seeds in a way where one or two being
unavailable is problematic, then the software may be using them
poorly.

Generally DNS seeds should only be used as fast connectivity hints,
primarily for initial connectivity. Relying on them exclusively
increases isolation vulnerabilities (e.g. because the dns seed
operators or any ISP or network attacker on the path between you and
the seeds can replace the results with ones that isolate you on a
bogus network).

Re: [Bitcoin-development] DNS seeds unstable

[Andreas Schildbach]

On 05/15/2014 07:48 PM, Gregory Maxwell wrote:
> On Thu, May 15, 2014 at 4:50 AM, Andreas Schildbach
> <andreas@schildbach.de> wrote:
>> I'm bringing this issue up again. The current Bitcoin DNS seed
>> infrastructure is unstable. I assume this is because of we're using a
>> custom DNS implementation which is not 100% compatible. There have been
>> bugs in the past, like a case sensitive match for the domain name.
>
> If software is using the DNS seeds in a way where one or two being
> unavailable is problematic, then the software may be using them
> poorly.
>
> Generally DNS seeds should only be used as fast connectivity hints,
> primarily for initial connectivity. Relying on them exclusively
> increases isolation vulnerabilities (e.g. because the dns seed
> operators or any ISP or network attacker on the path between you and
> the seeds can replace the results with ones that isolate you on a
> bogus network).

I just used "nslookup", after seeing the issues in bitcoinj.

I agree that clients should be robust regarding DNS lookups (and
bitcoinj isn't), but still I think the first step needs to be
maintaining a quality infrastructure.

Re: [Bitcoin-development] DNS seeds unstable

[Mike Hearn]

[-- Attachment #1: Type: text/plain, Size: 2180 bytes --]

My android wallet is working OK. Yes it isn't great when seeds have
temporary availability problems but things are still working.

There's a couple of pull reqs outstanding to include hard coded seed peers
and getaddr sourced IPs. Once those are finished and merged in there'll be
more backup paths.
On 16 May 2014 11:16, "Andreas Schildbach" <andreas@schildbach.de> wrote:

> On 05/15/2014 07:48 PM, Gregory Maxwell wrote:
> > On Thu, May 15, 2014 at 4:50 AM, Andreas Schildbach
> > <andreas@schildbach.de> wrote:
> >> I'm bringing this issue up again. The current Bitcoin DNS seed
> >> infrastructure is unstable. I assume this is because of we're using a
> >> custom DNS implementation which is not 100% compatible. There have been
> >> bugs in the past, like a case sensitive match for the domain name.
> >
> > If software is using the DNS seeds in a way where one or two being
> > unavailable is problematic, then the software may be using them
> > poorly.
> >
> > Generally DNS seeds should only be used as fast connectivity hints,
> > primarily for initial connectivity. Relying on them exclusively
> > increases isolation vulnerabilities (e.g. because the dns seed
> > operators or any ISP or network attacker on the path between you and
> > the seeds can replace the results with ones that isolate you on a
> > bogus network).
>
> I just used "nslookup", after seeing the issues in bitcoinj.
>
> I agree that clients should be robust regarding DNS lookups (and
> bitcoinj isn't), but still I think the first step needs to be
> maintaining a quality infrastructure.
>
>
>
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform
> available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>

[-- Attachment #2: Type: text/html, Size: 2930 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Luke Dashjr]

On Thursday, May 15, 2014 11:50:29 AM Andreas Schildbach wrote:
> dnsseed.bitcoin.dashjr.org		SERVFAIL, tried multiple ISPs

FWIW, this may be a routing issue: I notice various ISPs have been unable to 
route to my server over IPv4 today. IPv6 seems to be fine.

Not sure how important DNS seed reliability is anyway; it's just 
bootstrapping, and there are multiple servers listed.

Luke

Re: [Bitcoin-development] DNS seeds unstable

[Andreas Schildbach]

Apparently British Telecom also cannot speak to Peter Todd's server.

That another very large ISP in Europe.


On 05/15/2014 01:50 PM, Andreas Schildbach wrote:
> I'm bringing this issue up again. The current Bitcoin DNS seed
> infrastructure is unstable. I assume this is because of we're using a
> custom DNS implementation which is not 100% compatible. There have been
> bugs in the past, like a case sensitive match for the domain name.
> 
> Current state (seeds taken from bitcoinj):
> 
> mainnet:
> 
> seed.bitcoin.sipa.be			OK
> dnsseed.bluematt.me			OK
> dnsseed.bitcoin.dashjr.org		SERVFAIL, tried multiple ISPs
> seed.bitcoinstats.com			OK
> 
> testnet:
> 
> testnet-seed.bitcoin.petertodd.org	SERVFAIL, just from Telefonica
> testnet-seed.bluematt.me		OK (but only returns one node)
> 
> Note: Telefonica is one of Europe's largest ISPs.
> 
> I would try to improve DNS myself, but I'm not capable of writing C. My
> "fix" would be to reimplement everything in Java -- I doubt you guys
> would be happy with that.
> 
> 
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> 

Re: [Bitcoin-development] DNS seeds unstable

[Matt Whitlock]

Is Peter Todd's server actually up? The Google public DNS resolver at 8.8.8.8 can't resolve testnet-seed.bitcoin.petertodd.org either (SERVFAIL).


On Friday, 16 May 2014, at 6:34 pm, Andreas Schildbach wrote:
> Apparently British Telecom also cannot speak to Peter Todd's server.
> 
> That another very large ISP in Europe.
> 
> 
> On 05/15/2014 01:50 PM, Andreas Schildbach wrote:
> > I'm bringing this issue up again. The current Bitcoin DNS seed
> > infrastructure is unstable. I assume this is because of we're using a
> > custom DNS implementation which is not 100% compatible. There have been
> > bugs in the past, like a case sensitive match for the domain name.
> > 
> > Current state (seeds taken from bitcoinj):
> > 
> > mainnet:
> > 
> > seed.bitcoin.sipa.be			OK
> > dnsseed.bluematt.me			OK
> > dnsseed.bitcoin.dashjr.org		SERVFAIL, tried multiple ISPs
> > seed.bitcoinstats.com			OK
> > 
> > testnet:
> > 
> > testnet-seed.bitcoin.petertodd.org	SERVFAIL, just from Telefonica
> > testnet-seed.bluematt.me		OK (but only returns one node)
> > 
> > Note: Telefonica is one of Europe's largest ISPs.
> > 
> > I would try to improve DNS myself, but I'm not capable of writing C. My
> > "fix" would be to reimplement everything in Java -- I doubt you guys
> > would be happy with that.
> > 
> > 
> > ------------------------------------------------------------------------------
> > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> > Instantly run your Selenium tests across 300+ browser/OS combos.
> > Get unparalleled scalability from the best Selenium testing platform available
> > Simple to use. Nothing to install. Get started now for free."
> > http://p.sf.net/sfu/SauceLabs
> > 
> 
> 
> 
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development

Re: [Bitcoin-development] DNS seeds unstable

[Laszlo Hanyecz]

It looks like it might be firewalled, probably just need to fix the ACL in EC2.

-Laszlo

On May 16, 2014, at 4:34 PM, Andreas Schildbach <andreas@schildbach.de> wrote:

> Apparently British Telecom also cannot speak to Peter Todd's server.
> 
> That another very large ISP in Europe.
> 
> 
> On 05/15/2014 01:50 PM, Andreas Schildbach wrote:
>> I'm bringing this issue up again. The current Bitcoin DNS seed
>> infrastructure is unstable. I assume this is because of we're using a
>> custom DNS implementation which is not 100% compatible. There have been
>> bugs in the past, like a case sensitive match for the domain name.
>> 
>> Current state (seeds taken from bitcoinj):
>> 
>> mainnet:
>> 
>> seed.bitcoin.sipa.be			OK
>> dnsseed.bluematt.me			OK
>> dnsseed.bitcoin.dashjr.org		SERVFAIL, tried multiple ISPs
>> seed.bitcoinstats.com			OK
>> 
>> testnet:
>> 
>> testnet-seed.bitcoin.petertodd.org	SERVFAIL, just from Telefonica
>> testnet-seed.bluematt.me		OK (but only returns one node)
>> 
>> Note: Telefonica is one of Europe's largest ISPs.
>> 
>> I would try to improve DNS myself, but I'm not capable of writing C. My
>> "fix" would be to reimplement everything in Java -- I doubt you guys
>> would be happy with that.
>> 
>> 
>> ------------------------------------------------------------------------------
>> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
>> Instantly run your Selenium tests across 300+ browser/OS combos.
>> Get unparalleled scalability from the best Selenium testing platform available
>> Simple to use. Nothing to install. Get started now for free."
>> http://p.sf.net/sfu/SauceLabs
>> 
> 
> 
> 
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development

Re: [Bitcoin-development] DNS seeds unstable

[Alex Kotenko]

[-- Attachment #1: Type: text/plain, Size: 3913 bytes --]

Hi guys


Just wanted to let you know that Andreas' testnet Bitcoin Wallet doesn't
work because of fail in the peer discovery, and this caused us problems as
we cannot properly demonstrate my XBTerminal POS on the Bitcoin Conference.

Right now I'm booting up an own full node that I will set as trusted peer
in the Wallet settings, hopefully this will work. However this DNS
discovery problem is really a problem, even for testnet. Btw, I had
problems firing up the full bitcoind node also, of the same reason -
discovery failed. I had to ask Andreas to paste me his node list to
manually seed the nodelist.

So I've set up and will run a well connected testnet node, as we need it
for the XBTerminal.
Please let me know if I can somehow help to fix the DNS discovery issue
also.


Best regards,
Alex Kotenko


2014-05-16 17:46 GMT+01:00 Laszlo Hanyecz <laszlo@heliacal.net>:

> It looks like it might be firewalled, probably just need to fix the ACL in
> EC2.
>
> -Laszlo
>
> On May 16, 2014, at 4:34 PM, Andreas Schildbach <andreas@schildbach.de>
> wrote:
>
> > Apparently British Telecom also cannot speak to Peter Todd's server.
> >
> > That another very large ISP in Europe.
> >
> >
> > On 05/15/2014 01:50 PM, Andreas Schildbach wrote:
> >> I'm bringing this issue up again. The current Bitcoin DNS seed
> >> infrastructure is unstable. I assume this is because of we're using a
> >> custom DNS implementation which is not 100% compatible. There have been
> >> bugs in the past, like a case sensitive match for the domain name.
> >>
> >> Current state (seeds taken from bitcoinj):
> >>
> >> mainnet:
> >>
> >> seed.bitcoin.sipa.be                 OK
> >> dnsseed.bluematt.me                  OK
> >> dnsseed.bitcoin.dashjr.org           SERVFAIL, tried multiple ISPs
> >> seed.bitcoinstats.com                        OK
> >>
> >> testnet:
> >>
> >> testnet-seed.bitcoin.petertodd.org   SERVFAIL, just from Telefonica
> >> testnet-seed.bluematt.me             OK (but only returns one node)
> >>
> >> Note: Telefonica is one of Europe's largest ISPs.
> >>
> >> I would try to improve DNS myself, but I'm not capable of writing C. My
> >> "fix" would be to reimplement everything in Java -- I doubt you guys
> >> would be happy with that.
> >>
> >>
> >>
> ------------------------------------------------------------------------------
> >> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> >> Instantly run your Selenium tests across 300+ browser/OS combos.
> >> Get unparalleled scalability from the best Selenium testing platform
> available
> >> Simple to use. Nothing to install. Get started now for free."
> >> http://p.sf.net/sfu/SauceLabs
> >>
> >
> >
> >
> >
> ------------------------------------------------------------------------------
> > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> > Instantly run your Selenium tests across 300+ browser/OS combos.
> > Get unparalleled scalability from the best Selenium testing platform
> available
> > Simple to use. Nothing to install. Get started now for free."
> > http://p.sf.net/sfu/SauceLabs
> > _______________________________________________
> > Bitcoin-development mailing list
> > Bitcoin-development@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>
>
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform
> available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>

[-- Attachment #2: Type: text/html, Size: 6538 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Jeff Garzik]

There are only two testnet seeds listed in bitcoind, and one of them
returns SERVFAIL (testnet-seed.bitcoin.petertodd.org) and the other
just returns one A record (testnet-seed.bluematt.me).  No idea what
seeds bitcoinj uses.

If you are going to depend on testnet, especially for an important
demo... contribute to the infrastructure!  This stuff doesn't just fix
itself for free.




On Fri, May 16, 2014 at 7:07 PM, Alex Kotenko <alexykot@gmail.com> wrote:
> Hi guys
>
>
> Just wanted to let you know that Andreas' testnet Bitcoin Wallet doesn't
> work because of fail in the peer discovery, and this caused us problems as
> we cannot properly demonstrate my XBTerminal POS on the Bitcoin Conference.
>
> Right now I'm booting up an own full node that I will set as trusted peer in
> the Wallet settings, hopefully this will work. However this DNS discovery
> problem is really a problem, even for testnet. Btw, I had problems firing up
> the full bitcoind node also, of the same reason - discovery failed. I had to
> ask Andreas to paste me his node list to manually seed the nodelist.
>
> So I've set up and will run a well connected testnet node, as we need it for
> the XBTerminal.
> Please let me know if I can somehow help to fix the DNS discovery issue
> also.
>
>
> Best regards,
> Alex Kotenko
>
>
> 2014-05-16 17:46 GMT+01:00 Laszlo Hanyecz <laszlo@heliacal.net>:
>
>> It looks like it might be firewalled, probably just need to fix the ACL in
>> EC2.
>>
>> -Laszlo
>>
>> On May 16, 2014, at 4:34 PM, Andreas Schildbach <andreas@schildbach.de>
>> wrote:
>>
>> > Apparently British Telecom also cannot speak to Peter Todd's server.
>> >
>> > That another very large ISP in Europe.
>> >
>> >
>> > On 05/15/2014 01:50 PM, Andreas Schildbach wrote:
>> >> I'm bringing this issue up again. The current Bitcoin DNS seed
>> >> infrastructure is unstable. I assume this is because of we're using a
>> >> custom DNS implementation which is not 100% compatible. There have been
>> >> bugs in the past, like a case sensitive match for the domain name.
>> >>
>> >> Current state (seeds taken from bitcoinj):
>> >>
>> >> mainnet:
>> >>
>> >> seed.bitcoin.sipa.be                 OK
>> >> dnsseed.bluematt.me                  OK
>> >> dnsseed.bitcoin.dashjr.org           SERVFAIL, tried multiple ISPs
>> >> seed.bitcoinstats.com                        OK
>> >>
>> >> testnet:
>> >>
>> >> testnet-seed.bitcoin.petertodd.org   SERVFAIL, just from Telefonica
>> >> testnet-seed.bluematt.me             OK (but only returns one node)
>> >>
>> >> Note: Telefonica is one of Europe's largest ISPs.
>> >>
>> >> I would try to improve DNS myself, but I'm not capable of writing C. My
>> >> "fix" would be to reimplement everything in Java -- I doubt you guys
>> >> would be happy with that.
>> >>
>> >>
>> >>
>> >> ------------------------------------------------------------------------------
>> >> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
>> >> Instantly run your Selenium tests across 300+ browser/OS combos.
>> >> Get unparalleled scalability from the best Selenium testing platform
>> >> available
>> >> Simple to use. Nothing to install. Get started now for free."
>> >> http://p.sf.net/sfu/SauceLabs
>> >>
>> >
>> >
>> >
>> >
>> > ------------------------------------------------------------------------------
>> > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
>> > Instantly run your Selenium tests across 300+ browser/OS combos.
>> > Get unparalleled scalability from the best Selenium testing platform
>> > available
>> > Simple to use. Nothing to install. Get started now for free."
>> > http://p.sf.net/sfu/SauceLabs
>> > _______________________________________________
>> > Bitcoin-development mailing list
>> > Bitcoin-development@lists.sourceforge.net
>> > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>>
>>
>>
>> ------------------------------------------------------------------------------
>> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
>> Instantly run your Selenium tests across 300+ browser/OS combos.
>> Get unparalleled scalability from the best Selenium testing platform
>> available
>> Simple to use. Nothing to install. Get started now for free."
>> http://p.sf.net/sfu/SauceLabs
>> _______________________________________________
>> Bitcoin-development mailing list
>> Bitcoin-development@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>
>
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform
> available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>



-- 
Jeff Garzik
Bitcoin core developer and open source evangelist
BitPay, Inc.      https://bitpay.com/

Re: [Bitcoin-development] DNS seeds unstable

[Alex Kotenko]

[-- Attachment #1: Type: text/plain, Size: 5663 bytes --]

Ok, what do I need to do? How do I host a testnet seed myself?

Best regards,
Alex Kotenko


2014-05-16 23:02 GMT+01:00 Jeff Garzik <jgarzik@bitpay.com>:

> There are only two testnet seeds listed in bitcoind, and one of them
> returns SERVFAIL (testnet-seed.bitcoin.petertodd.org) and the other
> just returns one A record (testnet-seed.bluematt.me).  No idea what
> seeds bitcoinj uses.
>
> If you are going to depend on testnet, especially for an important
> demo... contribute to the infrastructure!  This stuff doesn't just fix
> itself for free.
>
>
>
>
> On Fri, May 16, 2014 at 7:07 PM, Alex Kotenko <alexykot@gmail.com> wrote:
> > Hi guys
> >
> >
> > Just wanted to let you know that Andreas' testnet Bitcoin Wallet doesn't
> > work because of fail in the peer discovery, and this caused us problems
> as
> > we cannot properly demonstrate my XBTerminal POS on the Bitcoin
> Conference.
> >
> > Right now I'm booting up an own full node that I will set as trusted
> peer in
> > the Wallet settings, hopefully this will work. However this DNS discovery
> > problem is really a problem, even for testnet. Btw, I had problems
> firing up
> > the full bitcoind node also, of the same reason - discovery failed. I
> had to
> > ask Andreas to paste me his node list to manually seed the nodelist.
> >
> > So I've set up and will run a well connected testnet node, as we need it
> for
> > the XBTerminal.
> > Please let me know if I can somehow help to fix the DNS discovery issue
> > also.
> >
> >
> > Best regards,
> > Alex Kotenko
> >
> >
> > 2014-05-16 17:46 GMT+01:00 Laszlo Hanyecz <laszlo@heliacal.net>:
> >
> >> It looks like it might be firewalled, probably just need to fix the ACL
> in
> >> EC2.
> >>
> >> -Laszlo
> >>
> >> On May 16, 2014, at 4:34 PM, Andreas Schildbach <andreas@schildbach.de>
> >> wrote:
> >>
> >> > Apparently British Telecom also cannot speak to Peter Todd's server.
> >> >
> >> > That another very large ISP in Europe.
> >> >
> >> >
> >> > On 05/15/2014 01:50 PM, Andreas Schildbach wrote:
> >> >> I'm bringing this issue up again. The current Bitcoin DNS seed
> >> >> infrastructure is unstable. I assume this is because of we're using a
> >> >> custom DNS implementation which is not 100% compatible. There have
> been
> >> >> bugs in the past, like a case sensitive match for the domain name.
> >> >>
> >> >> Current state (seeds taken from bitcoinj):
> >> >>
> >> >> mainnet:
> >> >>
> >> >> seed.bitcoin.sipa.be                 OK
> >> >> dnsseed.bluematt.me                  OK
> >> >> dnsseed.bitcoin.dashjr.org           SERVFAIL, tried multiple ISPs
> >> >> seed.bitcoinstats.com                        OK
> >> >>
> >> >> testnet:
> >> >>
> >> >> testnet-seed.bitcoin.petertodd.org   SERVFAIL, just from Telefonica
> >> >> testnet-seed.bluematt.me             OK (but only returns one node)
> >> >>
> >> >> Note: Telefonica is one of Europe's largest ISPs.
> >> >>
> >> >> I would try to improve DNS myself, but I'm not capable of writing C.
> My
> >> >> "fix" would be to reimplement everything in Java -- I doubt you guys
> >> >> would be happy with that.
> >> >>
> >> >>
> >> >>
> >> >>
> ------------------------------------------------------------------------------
> >> >> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For
> FREE
> >> >> Instantly run your Selenium tests across 300+ browser/OS combos.
> >> >> Get unparalleled scalability from the best Selenium testing platform
> >> >> available
> >> >> Simple to use. Nothing to install. Get started now for free."
> >> >> http://p.sf.net/sfu/SauceLabs
> >> >>
> >> >
> >> >
> >> >
> >> >
> >> >
> ------------------------------------------------------------------------------
> >> > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> >> > Instantly run your Selenium tests across 300+ browser/OS combos.
> >> > Get unparalleled scalability from the best Selenium testing platform
> >> > available
> >> > Simple to use. Nothing to install. Get started now for free."
> >> > http://p.sf.net/sfu/SauceLabs
> >> > _______________________________________________
> >> > Bitcoin-development mailing list
> >> > Bitcoin-development@lists.sourceforge.net
> >> > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> >>
> >>
> >>
> >>
> ------------------------------------------------------------------------------
> >> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> >> Instantly run your Selenium tests across 300+ browser/OS combos.
> >> Get unparalleled scalability from the best Selenium testing platform
> >> available
> >> Simple to use. Nothing to install. Get started now for free."
> >> http://p.sf.net/sfu/SauceLabs
> >> _______________________________________________
> >> Bitcoin-development mailing list
> >> Bitcoin-development@lists.sourceforge.net
> >> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> >
> >
> >
> >
> ------------------------------------------------------------------------------
> > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> > Instantly run your Selenium tests across 300+ browser/OS combos.
> > Get unparalleled scalability from the best Selenium testing platform
> > available
> > Simple to use. Nothing to install. Get started now for free."
> > http://p.sf.net/sfu/SauceLabs
> > _______________________________________________
> > Bitcoin-development mailing list
> > Bitcoin-development@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> >
>
>
>
> --
> Jeff Garzik
> Bitcoin core developer and open source evangelist
> BitPay, Inc.      https://bitpay.com/
>

[-- Attachment #2: Type: text/html, Size: 9249 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Andreas Schildbach]

I think the best way to contribute to the infrastructure is actually
what we're doing: Test the current infrastructure and point out where it
is not working. Trying to find solutions for problems.

There is nothing gained by throwing additional hardware at a problem if
the problem itself isn't understood well.


On 05/17/2014 02:58 AM, Alex Kotenko wrote:
> Ok, what do I need to do? How do I host a testnet seed myself? 
> 
> Best regards, 
> Alex Kotenko
> 
> 
> 2014-05-16 23:02 GMT+01:00 Jeff Garzik <jgarzik@bitpay.com
> <mailto:jgarzik@bitpay.com>>:
> 
>     There are only two testnet seeds listed in bitcoind, and one of them
>     returns SERVFAIL (testnet-seed.bitcoin.petertodd.org
>     <http://testnet-seed.bitcoin.petertodd.org>) and the other
>     just returns one A record (testnet-seed.bluematt.me
>     <http://testnet-seed.bluematt.me>).  No idea what
>     seeds bitcoinj uses.
> 
>     If you are going to depend on testnet, especially for an important
>     demo... contribute to the infrastructure!  This stuff doesn't just fix
>     itself for free.

Re: [Bitcoin-development] DNS seeds unstable

[Alex Kotenko]

[-- Attachment #1: Type: text/plain, Size: 2551 bytes --]

Agree.

So, my understanding is that atm we have no working DNS seeds at the
testnet3, right? There are two DNS seeds known, of which one is unreachable
atm, and another one is giving just one IP address, which is also a dead
node.

If I'll start a DNS seed of my own and make sure it works well, will this
help?
I've found this DNS seeder daemon <https://github.com/sipa/bitcoin-seeder>,
and it seems to be exactly what I need to run a DNS seeder myself.
So if my understanding is correct, I'll setup a DNS seeds for mainnet and
for testnet at bitcoin-seed.alexykot.me and testnet-seed.alexykot.me, and
also a well connected nodes for mainnet and testnet on the same server.
Is this a good plan? Will this all help?


Best regards,
Alex Kotenko


2014-05-17 12:39 GMT+01:00 Andreas Schildbach <andreas@schildbach.de>:

> I think the best way to contribute to the infrastructure is actually
> what we're doing: Test the current infrastructure and point out where it
> is not working. Trying to find solutions for problems.
>
> There is nothing gained by throwing additional hardware at a problem if
> the problem itself isn't understood well.
>
>
> On 05/17/2014 02:58 AM, Alex Kotenko wrote:
> > Ok, what do I need to do? How do I host a testnet seed myself?
> >
> > Best regards,
> > Alex Kotenko
> >
> >
> > 2014-05-16 23:02 GMT+01:00 Jeff Garzik <jgarzik@bitpay.com
> > <mailto:jgarzik@bitpay.com>>:
> >
> >     There are only two testnet seeds listed in bitcoind, and one of them
> >     returns SERVFAIL (testnet-seed.bitcoin.petertodd.org
> >     <http://testnet-seed.bitcoin.petertodd.org>) and the other
> >     just returns one A record (testnet-seed.bluematt.me
> >     <http://testnet-seed.bluematt.me>).  No idea what
> >     seeds bitcoinj uses.
> >
> >     If you are going to depend on testnet, especially for an important
> >     demo... contribute to the infrastructure!  This stuff doesn't just
> fix
> >     itself for free.
>
>
>
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform
> available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>

[-- Attachment #2: Type: text/html, Size: 5272 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Andreas Schildbach]

On 05/17/2014 02:02 PM, Alex Kotenko wrote:

> So, my understanding is that atm we have no working DNS seeds at the
> testnet3, right? There are two DNS seeds known, of which one is
> unreachable atm, and another one is giving just one IP address, which is
> also a dead node. 

Yes, that's my understanding too.

> If I'll start a DNS seed of my own and make sure it works well, will
> this help? 

Yes, definately.

> I've found this DNS seeder daemon
> <https://github.com/sipa/bitcoin-seeder>, and it seems to be exactly
> what I need to run a DNS seeder myself. 

Afaik this is what most of the other seeds are using, yes.

> So if my understanding is correct, I'll setup a DNS seeds for mainnet
> and for testnet at bitcoin-seed.alexykot.me
> <http://bitcoin-seed.alexykot.me> and testnet-seed.alexykot.me
> <http://testnet-seed.alexykot.me>, and also a well connected nodes for
> mainnet and testnet on the same server. 
> Is this a good plan? Will this all help?

Sound great! Let me know if you've got something to test.

Re: [Bitcoin-development] DNS seeds unstable

[Alex Kotenko]

[-- Attachment #1: Type: text/plain, Size: 2374 bytes --]

Hmm, I've mostly setup what's promised, testing DNS seeds now. There is one
problem I see that I can't really solve myself.
This dnsseed daemon cannot serve more than one name at once, which means
that I cannot serve testnet and mainnet seeds off one daemon instance which
means I need to buy two IP addresses for it. That's unfortunate as it needs
much more spendings from me to operate, second IP address will cost nearly
as much as the server itself.

​Can anybody help with this? I cannot into C++ to fix that myself.   ​


Best regards,
Alex Kotenko


2014-05-17 13:39 GMT+01:00 Andreas Schildbach <andreas@schildbach.de>:

> On 05/17/2014 02:02 PM, Alex Kotenko wrote:
>
> > So, my understanding is that atm we have no working DNS seeds at the
> > testnet3, right? There are two DNS seeds known, of which one is
> > unreachable atm, and another one is giving just one IP address, which is
> > also a dead node.
>
> Yes, that's my understanding too.
>
> > If I'll start a DNS seed of my own and make sure it works well, will
> > this help?
>
> Yes, definately.
>
> > I've found this DNS seeder daemon
> > <https://github.com/sipa/bitcoin-seeder>, and it seems to be exactly
> > what I need to run a DNS seeder myself.
>
> Afaik this is what most of the other seeds are using, yes.
>
> > So if my understanding is correct, I'll setup a DNS seeds for mainnet
> > and for testnet at bitcoin-seed.alexykot.me
> > <http://bitcoin-seed.alexykot.me> and testnet-seed.alexykot.me
> > <http://testnet-seed.alexykot.me>, and also a well connected nodes for
> > mainnet and testnet on the same server.
> > Is this a good plan? Will this all help?
>
> Sound great! Let me know if you've got something to test.
>
>
>
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform
> available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>

[-- Attachment #2: Type: text/html, Size: 4214 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Michael Wozniak]

I’m not familiar with how the daemon works, however could you set up two daemons listening local on different ports and with a separate daemon or normal dns server that proxies incoming queries to either domain? I don’t know if standard DNS servers would support that, or if you would need a custom proxy application.

-
Michael Wozniak


On May 19, 2014, at 4:14 PM, Alex Kotenko <alexykot@gmail.com> wrote:

> Hmm, I've mostly setup what's promised, testing DNS seeds now. There is one problem I see that I can't really solve myself. 
> This dnsseed daemon cannot serve more than one name at once, which means that I cannot serve testnet and mainnet seeds off one daemon instance which means I need to buy two IP addresses for it. That's unfortunate as it needs much more spendings from me to operate, second IP address will cost nearly as much as the server itself. 
> 
> ​Can anybody help with this? I cannot into C++ to fix that myself.   ​
> 
> 
> Best regards, 
> Alex Kotenko
> 
> 
> 2014-05-17 13:39 GMT+01:00 Andreas Schildbach <andreas@schildbach.de>:
> On 05/17/2014 02:02 PM, Alex Kotenko wrote:
> 
> > So, my understanding is that atm we have no working DNS seeds at the
> > testnet3, right? There are two DNS seeds known, of which one is
> > unreachable atm, and another one is giving just one IP address, which is
> > also a dead node.
> 
> Yes, that's my understanding too.
> 
> > If I'll start a DNS seed of my own and make sure it works well, will
> > this help?
> 
> Yes, definately.
> 
> > I've found this DNS seeder daemon
> > <https://github.com/sipa/bitcoin-seeder>, and it seems to be exactly
> > what I need to run a DNS seeder myself.
> 
> Afaik this is what most of the other seeds are using, yes.
> 
> > So if my understanding is correct, I'll setup a DNS seeds for mainnet
> > and for testnet at bitcoin-seed.alexykot.me
> > <http://bitcoin-seed.alexykot.me> and testnet-seed.alexykot.me
> > <http://testnet-seed.alexykot.me>, and also a well connected nodes for
> > mainnet and testnet on the same server.
> > Is this a good plan? Will this all help?
> 
> Sound great! Let me know if you've got something to test.
> 
> 
> 
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> 
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs_______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development

Re: [Bitcoin-development] DNS seeds unstable

[Alex Kotenko]

[-- Attachment #1: Type: text/plain, Size: 3832 bytes --]

Well, it's possible theoretically, but will need another piece of custom
software that will understand DNS protocol and proxy it correctly based on
actual incoming DNS queries.
On 19 May 2014 21:22, "Michael Wozniak" <mw@osfda.org> wrote:

> I’m not familiar with how the daemon works, however could you set up two
> daemons listening local on different ports and with a separate daemon or
> normal dns server that proxies incoming queries to either domain? I don’t
> know if standard DNS servers would support that, or if you would need a
> custom proxy application.
>
> -
> Michael Wozniak
>
>
> On May 19, 2014, at 4:14 PM, Alex Kotenko <alexykot@gmail.com> wrote:
>
> > Hmm, I've mostly setup what's promised, testing DNS seeds now. There is
> one problem I see that I can't really solve myself.
> > This dnsseed daemon cannot serve more than one name at once, which means
> that I cannot serve testnet and mainnet seeds off one daemon instance which
> means I need to buy two IP addresses for it. That's unfortunate as it needs
> much more spendings from me to operate, second IP address will cost nearly
> as much as the server itself.
> >
> > ​Can anybody help with this? I cannot into C++ to fix that myself.   ​
> >
> >
> > Best regards,
> > Alex Kotenko
> >
> >
> > 2014-05-17 13:39 GMT+01:00 Andreas Schildbach <andreas@schildbach.de>:
> > On 05/17/2014 02:02 PM, Alex Kotenko wrote:
> >
> > > So, my understanding is that atm we have no working DNS seeds at the
> > > testnet3, right? There are two DNS seeds known, of which one is
> > > unreachable atm, and another one is giving just one IP address, which
> is
> > > also a dead node.
> >
> > Yes, that's my understanding too.
> >
> > > If I'll start a DNS seed of my own and make sure it works well, will
> > > this help?
> >
> > Yes, definately.
> >
> > > I've found this DNS seeder daemon
> > > <https://github.com/sipa/bitcoin-seeder>, and it seems to be exactly
> > > what I need to run a DNS seeder myself.
> >
> > Afaik this is what most of the other seeds are using, yes.
> >
> > > So if my understanding is correct, I'll setup a DNS seeds for mainnet
> > > and for testnet at bitcoin-seed.alexykot.me
> > > <http://bitcoin-seed.alexykot.me> and testnet-seed.alexykot.me
> > > <http://testnet-seed.alexykot.me>, and also a well connected nodes for
> > > mainnet and testnet on the same server.
> > > Is this a good plan? Will this all help?
> >
> > Sound great! Let me know if you've got something to test.
> >
> >
> >
> >
> ------------------------------------------------------------------------------
> > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> > Instantly run your Selenium tests across 300+ browser/OS combos.
> > Get unparalleled scalability from the best Selenium testing platform
> available
> > Simple to use. Nothing to install. Get started now for free."
> > http://p.sf.net/sfu/SauceLabs
> > _______________________________________________
> > Bitcoin-development mailing list
> > Bitcoin-development@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> >
> >
> ------------------------------------------------------------------------------
> > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> > Instantly run your Selenium tests across 300+ browser/OS combos.
> > Get unparalleled scalability from the best Selenium testing platform
> available
> > Simple to use. Nothing to install. Get started now for free."
> >
> http://p.sf.net/sfu/SauceLabs_______________________________________________
> > Bitcoin-development mailing list
> > Bitcoin-development@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>
>

[-- Attachment #2: Type: text/html, Size: 5375 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Robert McKay]

It should be possible to configure bind as a DNS forwarder.. this can 
be done in a zone context.. then you can forward the different zones to 
different dnsseed daemons running on different non-public IPs or two 
different ports on the same IP (or on one single non-public IP since 
there's really no reason to expose the dnsseed directly daemon at all).

Rob

On Mon, 19 May 2014 21:14:32 +0100, Alex Kotenko wrote:
> Hmm, Ive mostly setup whats promised, testing DNS seeds now. There is
> one problem I see that I cant really solve myself. 
> This dnsseed daemon cannot serve more than one name at once, which
> means that I cannot serve testnet and mainnet seeds off one daemon
> instance which means I need to buy two IP addresses for it. Thats
> unfortunate as it needs much more spendings from me to operate, 
> second
> IP address will cost nearly as much as the server itself. 
>
> ​Can anybody help with this? I cannot into C++ to fix that myself.
>   ​
>
> Best regards, 
>
> Alex Kotenko
>
> 2014-05-17 13:39 GMT+01:00 Andreas Schildbach :
>
>> On 05/17/2014 02:02 PM, Alex Kotenko wrote:
>>
>> > So, my understanding is that atm we have no working DNS seeds at
>> the
>> > testnet3, right? There are two DNS seeds known, of which one is
>> > unreachable atm, and another one is giving just one IP address,
>> which is
>> > also a dead node.
>>
>> Yes, thats my understanding too.
>>
>> > If Ill start a DNS seed of my own and make sure it works well,
>> will
>> > this help?
>>
>> Yes, definately.
>>
>> > Ive found this DNS seeder daemon
>> > , and it seems to be exactly
>>
>>> what I need to run a DNS seeder myself.
>>
>> Afaik this is what most of the other seeds are using, yes.
>>
>> > So if my understanding is correct, Ill setup a DNS seeds for
>> mainnet
>> > and for testnet at bitcoin-seed.alexykot.me [2]
>> > and testnet-seed.alexykot.me [4]
>> > , and also a well connected nodes for
>>
>>> mainnet and testnet on the same server.
>> > Is this a good plan? Will this all help?
>>
>> Sound great! Let me know if youve got something to test.
>>
>>
> 
> ------------------------------------------------------------------------------
>> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For
>> FREE
>> Instantly run your Selenium tests across 300+ browser/OS combos.
>> Get unparalleled scalability from the best Selenium testing
>> platform available
>> Simple to use. Nothing to install. Get started now for free."
>> http://p.sf.net/sfu/SauceLabs [6]
>> _______________________________________________
>> Bitcoin-development mailing list
>> Bitcoin-development@lists.sourceforge.net [7]
>> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>> [8]
>
>
>
> Links:
> ------
> [1] https://github.com/sipa/bitcoin-seeder
> [2] http://bitcoin-seed.alexykot.me
> [3] http://bitcoin-seed.alexykot.me
> [4] http://testnet-seed.alexykot.me
> [5] http://testnet-seed.alexykot.me
> [6] http://p.sf.net/sfu/SauceLabs
> [7] mailto:Bitcoin-development@lists.sourceforge.net
> [8] https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> [9] mailto:andreas@schildbach.de

Re: [Bitcoin-development] DNS seeds unstable

[Jeff Garzik]

On Mon, May 19, 2014 at 4:36 PM, Robert McKay <robert@mckay.com> wrote:
> It should be possible to configure bind as a DNS forwarder.. this can
> be done in a zone context.. then you can forward the different zones to
> different dnsseed daemons running on different non-public IPs or two
> different ports on the same IP (or on one single non-public IP since
> there's really no reason to expose the dnsseed directly daemon at all).

Quite the opposite.  dnsseed data rotates through a lot of addresses
if available.  Using the bind/zone-xfer system would result in fewer
total addresses going through to the clients, thanks to the addition
of caching levels that the bind/zone-xfer system brings.

That said, if the choice is between no-service and bind, bind it is ;p

-- 
Jeff Garzik
Bitcoin core developer and open source evangelist
BitPay, Inc.      https://bitpay.com/

Re: [Bitcoin-development] DNS seeds unstable

[Robert McKay]

On Mon, 19 May 2014 19:49:52 -0400, Jeff Garzik wrote:
> On Mon, May 19, 2014 at 4:36 PM, Robert McKay <robert@mckay.com> 
> wrote:
>> It should be possible to configure bind as a DNS forwarder.. this 
>> can
>> be done in a zone context.. then you can forward the different zones 
>> to
>> different dnsseed daemons running on different non-public IPs or two
>> different ports on the same IP (or on one single non-public IP since
>> there's really no reason to expose the dnsseed directly daemon at 
>> all).
>
> Quite the opposite.  dnsseed data rotates through a lot of addresses
> if available.  Using the bind/zone-xfer system would result in fewer
> total addresses going through to the clients, thanks to the addition
> of caching levels that the bind/zone-xfer system brings.
>
> That said, if the choice is between no-service and bind, bind it is 
> ;p

Setting it up as a zone forwarder causes each request to go through to 
the dnsseed backend for each request.

Rob

Re: [Bitcoin-development] DNS seeds unstable

[Robert McKay]

On Tue, 20 May 2014 01:44:29 +0100, Robert McKay wrote:
> On Mon, 19 May 2014 19:49:52 -0400, Jeff Garzik wrote:
>> On Mon, May 19, 2014 at 4:36 PM, Robert McKay <robert@mckay.com>
>> wrote:
>>> It should be possible to configure bind as a DNS forwarder.. this
>>> can
>>> be done in a zone context.. then you can forward the different 
>>> zones
>>> to
>>> different dnsseed daemons running on different non-public IPs or 
>>> two
>>> different ports on the same IP (or on one single non-public IP 
>>> since
>>> there's really no reason to expose the dnsseed directly daemon at
>>> all).
>>
>> Quite the opposite.  dnsseed data rotates through a lot of addresses
>> if available.  Using the bind/zone-xfer system would result in fewer
>> total addresses going through to the clients, thanks to the addition
>> of caching levels that the bind/zone-xfer system brings.
>>
>> That said, if the choice is between no-service and bind, bind it is
>> ;p
>
> Setting it up as a zone forwarder causes each request to go through 
> to
> the dnsseed backend for each request.

This stackoverflow describes a similar situation;

http://stackoverflow.com/questions/15338232/how-to-forward-a-subzone

you can additionally specify the port to forward too;

http://www.zytrax.com/books/dns/ch7/queries.html#forwarders

it should be possible to forward to different ports on 127.0.0.1 for 
each dnsseed instance.

Rob

Re: [Bitcoin-development] DNS seeds unstable

[Alex Kotenko]

[-- Attachment #1: Type: text/plain, Size: 2779 bytes --]

okay, I've set it up with bind forwarding requests to two dnsseeds running
on separate ports. Though I see a problem with testnet DNS seed itself. It
runs, but somehow it only returns one IP address. Exactly same DNS seeder
looking for mainnet nodes is working fine.

You can reach seeds through
mainnet seed:
dig @node.alexykot.me bitcoin-seed.alexykot.me A
or directly
dig -p 8353 @node.alexykot.me bitcoin-seed.alexykot.me A

testnet seed
dig @node.alexykot.me testnet-seed.alexykot.me A
or directly
dig -p 18353 @node.alexykot.me testnet-seed.alexykot.me A

So what can be the problem with testnet DNS seeder?


Best regards,
Alex Kotenko


2014-05-20 1:50 GMT+01:00 Robert McKay <robert@mckay.com>:

> On Tue, 20 May 2014 01:44:29 +0100, Robert McKay wrote:
> > On Mon, 19 May 2014 19:49:52 -0400, Jeff Garzik wrote:
> >> On Mon, May 19, 2014 at 4:36 PM, Robert McKay <robert@mckay.com>
> >> wrote:
> >>> It should be possible to configure bind as a DNS forwarder.. this
> >>> can
> >>> be done in a zone context.. then you can forward the different
> >>> zones
> >>> to
> >>> different dnsseed daemons running on different non-public IPs or
> >>> two
> >>> different ports on the same IP (or on one single non-public IP
> >>> since
> >>> there's really no reason to expose the dnsseed directly daemon at
> >>> all).
> >>
> >> Quite the opposite.  dnsseed data rotates through a lot of addresses
> >> if available.  Using the bind/zone-xfer system would result in fewer
> >> total addresses going through to the clients, thanks to the addition
> >> of caching levels that the bind/zone-xfer system brings.
> >>
> >> That said, if the choice is between no-service and bind, bind it is
> >> ;p
> >
> > Setting it up as a zone forwarder causes each request to go through
> > to
> > the dnsseed backend for each request.
>
> This stackoverflow describes a similar situation;
>
> http://stackoverflow.com/questions/15338232/how-to-forward-a-subzone
>
> you can additionally specify the port to forward too;
>
> http://www.zytrax.com/books/dns/ch7/queries.html#forwarders
>
> it should be possible to forward to different ports on 127.0.0.1 for
> each dnsseed instance.
>
> Rob
>
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform
> available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>

[-- Attachment #2: Type: text/html, Size: 6281 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Andreas Schildbach]

Great, thanks for this contribution!

Do you plan to have your seeds reachable on port 53 eventually?
Currently bitcoinj cannot deal with nonstandard ports I think.


On 05/21/2014 11:23 AM, Alex Kotenko wrote:
> okay, I've set it up with bind forwarding requests to two dnsseeds
> running on separate ports. Though I see a problem with testnet DNS seed
> itself. It runs, but somehow it only returns one IP address. Exactly
> same DNS seeder looking for mainnet nodes is working fine. 
> 
> You can reach seeds through
> mainnet seed:
> dig @node.alexykot.me <http://node.alexykot.me> bitcoin-seed.alexykot.me
> <http://bitcoin-seed.alexykot.me> A
> or directly
> dig -p 8353 @node.alexykot.me <http://node.alexykot.me>
> bitcoin-seed.alexykot.me <http://bitcoin-seed.alexykot.me> A
> 
> testnet seed
> dig @node.alexykot.me <http://node.alexykot.me> testnet-seed.alexykot.me
> <http://testnet-seed.alexykot.me> A
> or directly
> dig -p 18353 @node.alexykot.me <http://node.alexykot.me>
> testnet-seed.alexykot.me <http://testnet-seed.alexykot.me> A
> 
> So what can be the problem with testnet DNS seeder?
> 
> 
> Best regards, 
> Alex Kotenko
> 
> 
> 2014-05-20 1:50 GMT+01:00 Robert McKay <robert@mckay.com
> <mailto:robert@mckay.com>>:
> 
>     On Tue, 20 May 2014 01:44:29 +0100, Robert McKay wrote:
>     > On Mon, 19 May 2014 19:49:52 -0400, Jeff Garzik wrote:
>     >> On Mon, May 19, 2014 at 4:36 PM, Robert McKay <robert@mckay.com
>     <mailto:robert@mckay.com>>
>     >> wrote:
>     >>> It should be possible to configure bind as a DNS forwarder.. this
>     >>> can
>     >>> be done in a zone context.. then you can forward the different
>     >>> zones
>     >>> to
>     >>> different dnsseed daemons running on different non-public IPs or
>     >>> two
>     >>> different ports on the same IP (or on one single non-public IP
>     >>> since
>     >>> there's really no reason to expose the dnsseed directly daemon at
>     >>> all).
>     >>
>     >> Quite the opposite.  dnsseed data rotates through a lot of addresses
>     >> if available.  Using the bind/zone-xfer system would result in fewer
>     >> total addresses going through to the clients, thanks to the addition
>     >> of caching levels that the bind/zone-xfer system brings.
>     >>
>     >> That said, if the choice is between no-service and bind, bind it is
>     >> ;p
>     >
>     > Setting it up as a zone forwarder causes each request to go through
>     > to
>     > the dnsseed backend for each request.
> 
>     This stackoverflow describes a similar situation;
> 
>     http://stackoverflow.com/questions/15338232/how-to-forward-a-subzone
> 
>     you can additionally specify the port to forward too;
> 
>     http://www.zytrax.com/books/dns/ch7/queries.html#forwarders
> 
>     it should be possible to forward to different ports on 127.0.0.1 for
>     each dnsseed instance.
> 
>     Rob
> 
>     ------------------------------------------------------------------------------
>     "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
>     Instantly run your Selenium tests across 300+ browser/OS combos.
>     Get unparalleled scalability from the best Selenium testing platform
>     available
>     Simple to use. Nothing to install. Get started now for free."
>     http://p.sf.net/sfu/SauceLabs
>     _______________________________________________
>     Bitcoin-development mailing list
>     Bitcoin-development@lists.sourceforge.net
>     <mailto:Bitcoin-development@lists.sourceforge.net>
>     https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> 
> 
> 
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> 

Re: [Bitcoin-development] DNS seeds unstable

[Alex Kotenko]

[-- Attachment #1: Type: text/plain, Size: 5260 bytes --]

Misunderstanding. Both seeds are available on port 53 via BIND forwarding.
Just also each DNS seed is available separately on it's own port.


Best regards,
Alex Kotenko


2014-05-21 12:03 GMT+01:00 Andreas Schildbach <andreas@schildbach.de>:

> Great, thanks for this contribution!
>
> Do you plan to have your seeds reachable on port 53 eventually?
> Currently bitcoinj cannot deal with nonstandard ports I think.
>
>
> On 05/21/2014 11:23 AM, Alex Kotenko wrote:
> > okay, I've set it up with bind forwarding requests to two dnsseeds
> > running on separate ports. Though I see a problem with testnet DNS seed
> > itself. It runs, but somehow it only returns one IP address. Exactly
> > same DNS seeder looking for mainnet nodes is working fine.
> >
> > You can reach seeds through
> > mainnet seed:
> > dig @node.alexykot.me <http://node.alexykot.me> bitcoin-seed.alexykot.me
> > <http://bitcoin-seed.alexykot.me> A
> > or directly
> > dig -p 8353 @node.alexykot.me <http://node.alexykot.me>
> > bitcoin-seed.alexykot.me <http://bitcoin-seed.alexykot.me> A
> >
> > testnet seed
> > dig @node.alexykot.me <http://node.alexykot.me> testnet-seed.alexykot.me
> > <http://testnet-seed.alexykot.me> A
> > or directly
> > dig -p 18353 @node.alexykot.me <http://node.alexykot.me>
> > testnet-seed.alexykot.me <http://testnet-seed.alexykot.me> A
> >
> > So what can be the problem with testnet DNS seeder?
> >
> >
> > Best regards,
> > Alex Kotenko
> >
> >
> > 2014-05-20 1:50 GMT+01:00 Robert McKay <robert@mckay.com
> > <mailto:robert@mckay.com>>:
> >
> >     On Tue, 20 May 2014 01:44:29 +0100, Robert McKay wrote:
> >     > On Mon, 19 May 2014 19:49:52 -0400, Jeff Garzik wrote:
> >     >> On Mon, May 19, 2014 at 4:36 PM, Robert McKay <robert@mckay.com
> >     <mailto:robert@mckay.com>>
> >     >> wrote:
> >     >>> It should be possible to configure bind as a DNS forwarder.. this
> >     >>> can
> >     >>> be done in a zone context.. then you can forward the different
> >     >>> zones
> >     >>> to
> >     >>> different dnsseed daemons running on different non-public IPs or
> >     >>> two
> >     >>> different ports on the same IP (or on one single non-public IP
> >     >>> since
> >     >>> there's really no reason to expose the dnsseed directly daemon at
> >     >>> all).
> >     >>
> >     >> Quite the opposite.  dnsseed data rotates through a lot of
> addresses
> >     >> if available.  Using the bind/zone-xfer system would result in
> fewer
> >     >> total addresses going through to the clients, thanks to the
> addition
> >     >> of caching levels that the bind/zone-xfer system brings.
> >     >>
> >     >> That said, if the choice is between no-service and bind, bind it
> is
> >     >> ;p
> >     >
> >     > Setting it up as a zone forwarder causes each request to go through
> >     > to
> >     > the dnsseed backend for each request.
> >
> >     This stackoverflow describes a similar situation;
> >
> >     http://stackoverflow.com/questions/15338232/how-to-forward-a-subzone
> >
> >     you can additionally specify the port to forward too;
> >
> >     http://www.zytrax.com/books/dns/ch7/queries.html#forwarders
> >
> >     it should be possible to forward to different ports on 127.0.0.1 for
> >     each dnsseed instance.
> >
> >     Rob
> >
> >
> ------------------------------------------------------------------------------
> >     "Accelerate Dev Cycles with Automated Cross-Browser Testing - For
> FREE
> >     Instantly run your Selenium tests across 300+ browser/OS combos.
> >     Get unparalleled scalability from the best Selenium testing platform
> >     available
> >     Simple to use. Nothing to install. Get started now for free."
> >     http://p.sf.net/sfu/SauceLabs
> >     _______________________________________________
> >     Bitcoin-development mailing list
> >     Bitcoin-development@lists.sourceforge.net
> >     <mailto:Bitcoin-development@lists.sourceforge.net>
> >     https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> >
> >
> >
> >
> >
> ------------------------------------------------------------------------------
> > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> > Instantly run your Selenium tests across 300+ browser/OS combos.
> > Get unparalleled scalability from the best Selenium testing platform
> available
> > Simple to use. Nothing to install. Get started now for free."
> > http://p.sf.net/sfu/SauceLabs
> >
> >
> >
> > _______________________________________________
> > Bitcoin-development mailing list
> > Bitcoin-development@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> >
>
>
>
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform
> available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>

[-- Attachment #2: Type: text/html, Size: 9257 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Andreas Schildbach]

Hmmm, not for me:

$ nslookup bitcoin-seed.alexykot.me
Server:		127.0.1.1
Address:	127.0.1.1#53

** server can't find bitcoin-seed.alexykot.me: SERVFAIL

$ nslookup testnet-seed.alexykot.me
Server:		127.0.1.1
Address:	127.0.1.1#53

** server can't find testnet-seed.alexykot.me: SERVFAIL


Can you look up in the logfile what requests I just made?



On 05/21/2014 01:10 PM, Alex Kotenko wrote:
> Misunderstanding. Both seeds are available on port 53 via BIND forwarding.
> Just also each DNS seed is available separately on it's own port.
> 
> 
> Best regards, 
> Alex Kotenko
> 
> 
> 2014-05-21 12:03 GMT+01:00 Andreas Schildbach <andreas@schildbach.de
> <mailto:andreas@schildbach.de>>:
> 
>     Great, thanks for this contribution!
> 
>     Do you plan to have your seeds reachable on port 53 eventually?
>     Currently bitcoinj cannot deal with nonstandard ports I think.
> 
> 
>     On 05/21/2014 11:23 AM, Alex Kotenko wrote:
>     > okay, I've set it up with bind forwarding requests to two dnsseeds
>     > running on separate ports. Though I see a problem with testnet DNS
>     seed
>     > itself. It runs, but somehow it only returns one IP address. Exactly
>     > same DNS seeder looking for mainnet nodes is working fine.
>     >
>     > You can reach seeds through
>     > mainnet seed:
>     > dig @node.alexykot.me <http://node.alexykot.me>
>     <http://node.alexykot.me> bitcoin-seed.alexykot.me
>     <http://bitcoin-seed.alexykot.me>
>     > <http://bitcoin-seed.alexykot.me> A
>     > or directly
>     > dig -p 8353 @node.alexykot.me <http://node.alexykot.me>
>     <http://node.alexykot.me>
>     > bitcoin-seed.alexykot.me <http://bitcoin-seed.alexykot.me>
>     <http://bitcoin-seed.alexykot.me> A
>     >
>     > testnet seed
>     > dig @node.alexykot.me <http://node.alexykot.me>
>     <http://node.alexykot.me> testnet-seed.alexykot.me
>     <http://testnet-seed.alexykot.me>
>     > <http://testnet-seed.alexykot.me> A
>     > or directly
>     > dig -p 18353 @node.alexykot.me <http://node.alexykot.me>
>     <http://node.alexykot.me>
>     > testnet-seed.alexykot.me <http://testnet-seed.alexykot.me>
>     <http://testnet-seed.alexykot.me> A
>     >
>     > So what can be the problem with testnet DNS seeder?
>     >
>     >
>     > Best regards,
>     > Alex Kotenko
>     >
>     >
>     > 2014-05-20 1:50 GMT+01:00 Robert McKay <robert@mckay.com
>     <mailto:robert@mckay.com>
>     > <mailto:robert@mckay.com <mailto:robert@mckay.com>>>:
>     >
>     >     On Tue, 20 May 2014 01:44:29 +0100, Robert McKay wrote:
>     >     > On Mon, 19 May 2014 19:49:52 -0400, Jeff Garzik wrote:
>     >     >> On Mon, May 19, 2014 at 4:36 PM, Robert McKay
>     <robert@mckay.com <mailto:robert@mckay.com>
>     >     <mailto:robert@mckay.com <mailto:robert@mckay.com>>>
>     >     >> wrote:
>     >     >>> It should be possible to configure bind as a DNS
>     forwarder.. this
>     >     >>> can
>     >     >>> be done in a zone context.. then you can forward the different
>     >     >>> zones
>     >     >>> to
>     >     >>> different dnsseed daemons running on different non-public
>     IPs or
>     >     >>> two
>     >     >>> different ports on the same IP (or on one single non-public IP
>     >     >>> since
>     >     >>> there's really no reason to expose the dnsseed directly
>     daemon at
>     >     >>> all).
>     >     >>
>     >     >> Quite the opposite.  dnsseed data rotates through a lot of
>     addresses
>     >     >> if available.  Using the bind/zone-xfer system would result
>     in fewer
>     >     >> total addresses going through to the clients, thanks to the
>     addition
>     >     >> of caching levels that the bind/zone-xfer system brings.
>     >     >>
>     >     >> That said, if the choice is between no-service and bind,
>     bind it is
>     >     >> ;p
>     >     >
>     >     > Setting it up as a zone forwarder causes each request to go
>     through
>     >     > to
>     >     > the dnsseed backend for each request.
>     >
>     >     This stackoverflow describes a similar situation;
>     >
>     >    
>     http://stackoverflow.com/questions/15338232/how-to-forward-a-subzone
>     >
>     >     you can additionally specify the port to forward too;
>     >
>     >     http://www.zytrax.com/books/dns/ch7/queries.html#forwarders
>     >
>     >     it should be possible to forward to different ports on
>     127.0.0.1 for
>     >     each dnsseed instance.
>     >
>     >     Rob
>     >
>     >    
>     ------------------------------------------------------------------------------
>     >     "Accelerate Dev Cycles with Automated Cross-Browser Testing -
>     For FREE
>     >     Instantly run your Selenium tests across 300+ browser/OS combos.
>     >     Get unparalleled scalability from the best Selenium testing
>     platform
>     >     available
>     >     Simple to use. Nothing to install. Get started now for free."
>     >     http://p.sf.net/sfu/SauceLabs
>     >     _______________________________________________
>     >     Bitcoin-development mailing list
>     >     Bitcoin-development@lists.sourceforge.net
>     <mailto:Bitcoin-development@lists.sourceforge.net>
>     >     <mailto:Bitcoin-development@lists.sourceforge.net
>     <mailto:Bitcoin-development@lists.sourceforge.net>>
>     >     https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>     >
>     >
>     >
>     >
>     >
>     ------------------------------------------------------------------------------
>     > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
>     > Instantly run your Selenium tests across 300+ browser/OS combos.
>     > Get unparalleled scalability from the best Selenium testing
>     platform available
>     > Simple to use. Nothing to install. Get started now for free."
>     > http://p.sf.net/sfu/SauceLabs
>     >
>     >
>     >
>     > _______________________________________________
>     > Bitcoin-development mailing list
>     > Bitcoin-development@lists.sourceforge.net
>     <mailto:Bitcoin-development@lists.sourceforge.net>
>     > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>     >
> 
> 
> 
>     ------------------------------------------------------------------------------
>     "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
>     Instantly run your Selenium tests across 300+ browser/OS combos.
>     Get unparalleled scalability from the best Selenium testing platform
>     available
>     Simple to use. Nothing to install. Get started now for free."
>     http://p.sf.net/sfu/SauceLabs
>     _______________________________________________
>     Bitcoin-development mailing list
>     Bitcoin-development@lists.sourceforge.net
>     <mailto:Bitcoin-development@lists.sourceforge.net>
>     https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> 
> 
> 
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> 

Re: [Bitcoin-development] DNS seeds unstable

[Alex Kotenko]

[-- Attachment #1: Type: text/plain, Size: 623 bytes --]

Hi all


It took some time, but now my testnet seed is fully operational. In fact
I've dropped an earlier idea of DNS forwarding and now serving only testnet
seed, as it is more important atm than a mainnet one.
Testnet seed is available at testnet-seed.alexykot.me.
Please check and let me know if all looks fine with this seed from your
side.


Best regards,
Alex Kotenko


2014-05-22 9:58 GMT+01:00 Andreas Schildbach <andreas@schildbach.de>:

> Hi Alex,
>
> I'm not sure if you saw this message.
> Your seeds are not reachable from my ISP unfortunately.
>
> Cheers,
>
> Andreas
>
>
> ​
>

[-- Attachment #2: Type: text/html, Size: 2022 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Andreas Schildbach]

Thanks Alex, you're my hero of the day (-:

Your seed works well for me. Here is a PR for bitcoinj:

https://github.com/bitcoinj/bitcoinj/pull/101


On 06/11/2014 03:57 PM, Alex Kotenko wrote:
> Hi all
> 
> 
> It took some time, but now my testnet seed is fully operational. In fact
> I've dropped an earlier idea of DNS forwarding and now serving only
> testnet seed, as it is more important atm than a mainnet one. 
> Testnet seed is available at testnet-seed.alexykot.me
> <http://testnet-seed.alexykot.me>.
> Please check and let me know if all looks fine with this seed from your
> side.
> 
> 
> Best regards, 
> Alex Kotenko
> 
> 
> 2014-05-22 9:58 GMT+01:00 Andreas Schildbach <andreas@schildbach.de
> <mailto:andreas@schildbach.de>>:
> 
>     Hi Alex,
> 
>     I'm not sure if you saw this message.
>     Your seeds are not reachable from my ISP unfortunately.
> 
>     Cheers,
> 
>     Andreas
> 
> 
>     ​
> 
> 
> 
> ------------------------------------------------------------------------------
> HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
> Find What Matters Most in Your Big Data with HPCC Systems
> Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
> Leverages Graph Analysis for Fast Processing & Easy Data Exploration
> http://p.sf.net/sfu/hpccsystems
> 
> 
> 
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
> 

Re: [Bitcoin-development] DNS seeds unstable

[Michael Wozniak]

You would set it up as a forwarder, not as a zone transfer to bind.  That should proxy the request every time and only cache based on any TTL that’s set in the response.

Here’s an example of how it could work:
https://planet.jboss.org/post/setting_up_a_forwarding_dns_server_or_dns_proxy_with_isc_bind


On May 19, 2014, at 7:49 PM, Jeff Garzik <jgarzik@bitpay.com> wrote:

> On Mon, May 19, 2014 at 4:36 PM, Robert McKay <robert@mckay.com> wrote:
>> It should be possible to configure bind as a DNS forwarder.. this can
>> be done in a zone context.. then you can forward the different zones to
>> different dnsseed daemons running on different non-public IPs or two
>> different ports on the same IP (or on one single non-public IP since
>> there's really no reason to expose the dnsseed directly daemon at all).
> 
> Quite the opposite.  dnsseed data rotates through a lot of addresses
> if available.  Using the bind/zone-xfer system would result in fewer
> total addresses going through to the clients, thanks to the addition
> of caching levels that the bind/zone-xfer system brings.
> 
> That said, if the choice is between no-service and bind, bind it is ;p
> 
> -- 
> Jeff Garzik
> Bitcoin core developer and open source evangelist
> BitPay, Inc.      https://bitpay.com/
> 
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development

Re: [Bitcoin-development] DNS seeds unstable

[Rob Golding]

> > dnsseed.bitcoin.dashjr.org		SERVFAIL, tried multiple ISPs

dnsseed.bitcoin.dashjr.org. 60  IN      NS      jun.dashjr.org.
but jun.dashjr.org isn't responding to dns queries (as at 18.10 GMT
2014-05-16)

that would be a fundamental problem with the dns infrastructure for that
domain (and the sub-hosts/records) , with the authoritive server not
replying to the dns query

> > testnet-seed.bitcoin.petertodd.org	SERVFAIL, just from Telefonica

Similarly no response from the alias'd aws dns service on 23.21.243.183
(testnet-seed-ns1.bitcoin.petertodd.org) from various test locations in
Europe

If there's a requirement for a domain & highly redundant dns to hard-code
into something, and one of the dev's drops me an email, I can get that
organised FoC, but these issues look like 'common'
firewall/transit/connectivity issues at first glance.

Rob

Re: [Bitcoin-development] DNS seeds unstable

[Nick Simpson]

[-- Attachment #1: Type: text/plain, Size: 1696 bytes --]

Luke's server has been having issues with ipv4 routing lately anyway, even ignoring any DNS issues.

Nick

On May 16, 2014 12:17:17 PM CDT, Rob Golding <rob.golding@astutium.com> wrote:
>> > dnsseed.bitcoin.dashjr.org		SERVFAIL, tried multiple ISPs
>
>dnsseed.bitcoin.dashjr.org. 60  IN      NS      jun.dashjr.org.
>but jun.dashjr.org isn't responding to dns queries (as at 18.10 GMT
>2014-05-16)
>
>that would be a fundamental problem with the dns infrastructure for
>that
>domain (and the sub-hosts/records) , with the authoritive server not
>replying to the dns query
>
>> > testnet-seed.bitcoin.petertodd.org	SERVFAIL, just from Telefonica
>
>Similarly no response from the alias'd aws dns service on 23.21.243.183
>(testnet-seed-ns1.bitcoin.petertodd.org) from various test locations in
>Europe
>
>If there's a requirement for a domain & highly redundant dns to
>hard-code
>into something, and one of the dev's drops me an email, I can get that
>organised FoC, but these issues look like 'common'
>firewall/transit/connectivity issues at first glance.
>
>Rob
>
>
>------------------------------------------------------------------------------
>"Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
>Instantly run your Selenium tests across 300+ browser/OS combos.
>Get unparalleled scalability from the best Selenium testing platform
>available
>Simple to use. Nothing to install. Get started now for free."
>http://p.sf.net/sfu/SauceLabs
>_______________________________________________
>Bitcoin-development mailing list
>Bitcoin-development@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/bitcoin-development

[-- Attachment #2: Type: text/html, Size: 2905 bytes --]

Re: [Bitcoin-development] DNS seeds unstable

[Luke Dashjr]

On Friday, May 16, 2014 5:17:17 PM Rob Golding wrote:
> > > dnsseed.bitcoin.dashjr.org		SERVFAIL, tried multiple ISPs
> 
> dnsseed.bitcoin.dashjr.org. 60  IN      NS      jun.dashjr.org.
> but jun.dashjr.org isn't responding to dns queries (as at 18.10 GMT
> 2014-05-16)
> 
> that would be a fundamental problem with the dns infrastructure for that
> domain (and the sub-hosts/records) , with the authoritive server not
> replying to the dns query

jun.dashjr.org has resolved 130 requests in the past ~12 hours, and I can see 
it resolving more in tcpdump. I do observe two potential problems, however:

- For the past 24-48 hours, there seem to be some significant routing problems 
on the IPv4 internet, resulting in people using various diverse ISPs 
(including myself at home) being unable to route to any of my servers. This 
issue doesn't affect IPv6, however.

- The DNS seeder seems to only be responding to requests received over IPv4. 
This means IPv6 DNS servers get no response. :(

Luke

Re: [Bitcoin-development] DNS seeds unstable

[Matt Corallo]

This is very strange...when did you run this test and can anyone else
reproduce this?

Matt

On 05/15/14 11:50, Andreas Schildbach wrote:
> testnet-seed.bluematt.me		OK (but only returns one node)

Re: [Bitcoin-development] DNS seeds unstable

[Andreas Schildbach]

For example just now:

$ nslookup testnet-seed.bluematt.me
Server:		127.0.1.1
Address:	127.0.1.1#53

Non-authoritative answer:
testnet-seed.bluematt.me	canonical name = bitcoin-seednode.bluematt.me.
bitcoin-seednode.bluematt.me	canonical name = desktopv2.bluematt.me.
Name:	desktopv2.bluematt.me
Address: 152.23.202.18

And that address doesn't connect on port 18333.


On 05/16/2014 08:53 PM, Matt Corallo wrote:
> This is very strange...when did you run this test and can anyone else
> reproduce this?
> 
> Matt
> 
> On 05/15/14 11:50, Andreas Schildbach wrote:
>> testnet-seed.bluematt.me		OK (but only returns one node)
> 
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> 

Re: [Bitcoin-development] DNS seeds unstable

[Matt Corallo]

Oops, missed the lost

On May 16, 2014 3:04:40 PM EDT, Matt Corallo <bitcoin-list@bluematt.me> wrote:
>Oh, I missed that this was the testnet seed. Yea, that one never got
>set
>up properly and was just pointing to a static seed node (that is now
>down...). The mainnet seed actually works.
>
>On 05/16/14 19:01, Laszlo Hanyecz wrote:
>> Matt,
>> 
>> I get the same:
>> 
>> $ host testnet-seed.bluematt.me
>> testnet-seed.bluematt.me is an alias for
>bitcoin-seednode.bluematt.me.
>> bitcoin-seednode.bluematt.me is an alias for desktopv2.bluematt.me.
>> desktopv2.bluematt.me has address 152.23.202.18
>> 
>> 
>> $ dig +trace desktopv2.bluematt.me. any
>> 
>> ; <<>> DiG 9.8.5-P1 <<>> +trace desktopv2.bluematt.me. any
>> ;; global options: +cmd
>> .			451792	IN	NS	l.root-servers.net.
>> .			451792	IN	NS	d.root-servers.net.
>> .			451792	IN	NS	e.root-servers.net.
>> .			451792	IN	NS	c.root-servers.net.
>> .			451792	IN	NS	j.root-servers.net.
>> .			451792	IN	NS	b.root-servers.net.
>> .			451792	IN	NS	h.root-servers.net.
>> .			451792	IN	NS	f.root-servers.net.
>> .			451792	IN	NS	g.root-servers.net.
>> .			451792	IN	NS	a.root-servers.net.
>> .			451792	IN	NS	i.root-servers.net.
>> .			451792	IN	NS	m.root-servers.net.
>> .			451792	IN	NS	k.root-servers.net.
>> ;; Received 512 bytes from 2a00:5540:5014::1#53(2a00:5540:5014::1) in
>2 ms
>> 
>> me.			172800	IN	NS	b2.me.afilias-nst.org.
>> me.			172800	IN	NS	ns2.nic.me.
>> me.			172800	IN	NS	a0.cctld.afilias-nst.info.
>> me.			172800	IN	NS	b0.cctld.afilias-nst.org.
>> me.			172800	IN	NS	c0.cctld.afilias-nst.info.
>> me.			172800	IN	NS	d0.cctld.afilias-nst.org.
>> me.			172800	IN	NS	a2.me.afilias-nst.info.
>> me.			172800	IN	NS	ns.nic.me.
>> ;; Received 509 bytes from 2001:7fe::53#53(i.root-servers.net) in
>1807 ms
>> 
>> bluematt.me.		86400	IN	NS	ns.bluematt.me.
>> bluematt.me.		86400	IN	NS	ns3.he.net.
>> bluematt.me.		86400	IN	NS	ns4.he.net.
>> bluematt.me.		86400	IN	NS	ns1.rollernet.us.
>> bluematt.me.		86400	IN	NS	ns2.rollernet.us.
>> ;; Received 162 bytes from
>2001:500:26::1#53(b0.cctld.afilias-nst.org) in 118 ms
>> 
>> desktopv2.bluematt.me.	3600	IN	A	152.23.202.18
>> bluematt.me.		86400	IN	NS	ns4.he.net.
>> bluematt.me.		86400	IN	NS	ns3.he.net.
>> bluematt.me.		86400	IN	NS	ns2.rollernet.us.
>> bluematt.me.		86400	IN	NS	ns1.rollernet.us.
>> bluematt.me.		86400	IN	NS	ns.bluematt.me.
>> ;; Received 178 bytes from 2607:fe70:0:4::b#53(ns2.rollernet.us) in
>126 ms
>> 
>> 
>> 
>> 
>> Thanks,
>> Laszlo
>> 
>> 
>> On May 16, 2014, at 6:53 PM, Matt Corallo <bitcoin-list@bluematt.me>
>wrote:
>> 
>>> This is very strange...when did you run this test and can anyone
>else
>>> reproduce this?
>>>
>>> Matt
>>>
>>> On 05/15/14 11:50, Andreas Schildbach wrote:
>>>> testnet-seed.bluematt.me		OK (but only returns one node)
>>>
>>>
>------------------------------------------------------------------------------
>>> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For
>FREE
>>> Instantly run your Selenium tests across 300+ browser/OS combos.
>>> Get unparalleled scalability from the best Selenium testing platform
>available
>>> Simple to use. Nothing to install. Get started now for free."
>>> http://p.sf.net/sfu/SauceLabs
>>> _______________________________________________
>>> Bitcoin-development mailing list
>>> Bitcoin-development@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>>