Re: [PDB Tech] PeeringDB API Discrepency

[James Bensley]

On 4 October 2016 at 22:18, Matt Griswold  wrote:
> From looking through the code, this was never implemented in v2, we can't
> think of any technical reasons so it was likely just my screw up while
> migrating the database. There are 84 dupe v4 and 51 dupe v6, so we'll fix
> the data and add the constraint.
>
> The deploy automation needs updating from switching hosts before, but we'll
> get this out ASAP.


Thanks for l looking into this guys.

I meat to sent a follow up email to say I have found lots more, trying
to pull infomration from many other IX's it was quickly clear the
issue was wide spread, so thanks for picking up on the fact it's not
just that specific exchange.

Also I might have found another issue but I think  that may also be an
issue with my script so I won't say anything else because I haven't
had a chance to look into it yet.

Cheers,
James.
___
Pdb-tech mailing list
Pdb-tech@lists.peeringdb.com
http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech

Re: [PDB Tech] PeeringDB API Discrepency

[Chris Van Fossen]

*> There are 84 dupe v4 and 51 dupe v6, so we'll fix the data and add the
constraint.*

That constraint can be a problem.

Example:
Network ABC is given an IP by an IX and updates PeeringDB. Later on,
Network ABC leaves the IX, or goes out of business, or stops caring about
peering. Their PeeringDB entry stays up. The IX decides to give the IP to
new network XYZ. Now the new network cannot put it into PeeringDB, as the
old entry is still sitting with Network ABC.

Happened to me in the past. Got around it by adding a space at the end of
the IP (had to remember to strip out the space when processing data from
PeeringDB).

If we add the constraint, does the resolution become reaching out to
PeeringDB support and waiting for admin to sort it?

Best,
Chris

On Tue, Oct 4, 2016 at 2:18 PM, Matt Griswold  wrote:

> From looking through the code, this was never implemented in v2, we can't
> think of any technical reasons so it was likely just my screw up while
> migrating the database. There are 84 dupe v4 and 51 dupe v6, so we'll fix
> the data and add the constraint.
>
> The deploy automation needs updating from switching hosts before, but
> we'll get this out ASAP.
>
> On Oct 4, 2016 3:28 PM, "Arnold Nipper"  wrote:
>
>> On 04.10.2016 18:53, James Bensley wrote:
>> > On 13 September 2016 at 11:00, Stefan Pratter  wrote:
>> >> Hi,
>> >>
>> >> Thank you for bringing this to our attention, i have opened a github
>> ticket
>> >> for it here:
>> >>
>> >> https://github.com/peeringdb/peeringdb/issues/68
>> >>
>> >> Stefan
>> >
>> >
>> > Whilst playing the the PDB API some more I have found another "issue".
>> > It's not an issue with the API as such but perhaps a feature request;
>> > there seems to be no duplication detection.
>> >
>> > I was trying to automate some stuff at this IX:
>> https://www.peeringdb.com/ix/745
>> >
>> > If you look there are two peers listed at that exchange with the same
>> > IPv4 address (AS 8468 and AS 60688, both with IP 195.66.246.11). This
>> > kind of thing should be easily detectable. Any scope for this sort of
>> > function to be added, where the AS number and company are different
>> > there shouldn't be a matching IP, the IP can only be announced by one
>> > AS number at an exchange.
>> >
>>
>> Woow ... how could this happen? Afaik it is not possible to add an IP
>> twice or even more often.
>>
>> Thanks for spotting. We will follow up with LINX to sort out the issue.
>>
>>
>> Arnold
>> --
>> Arnold Nipper
>> Chief Technology Evangelist and Co-Founder
>>
>> DE-CIX Management GmbH | Lindleystrasse 12 | 60314 Frankfurt am Main |
>> Germany | www.de-cix.net | Phone +49 69 1730902 22 |
>> Mobile +49 172 2650958 | Fax +49 69 4056 2716 |
>> arnold.nip...@de-cix.net | Geschaeftsfuehrer Harald A. Summa |
>> Registergericht AG Koeln HRB 51135
>>
>>
>> ___
>> Pdb-tech mailing list
>> Pdb-tech@lists.peeringdb.com
>> http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech
>>
>>
> ___
> Pdb-tech mailing list
> Pdb-tech@lists.peeringdb.com
> http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech
>
___
Pdb-tech mailing list
Pdb-tech@lists.peeringdb.com
http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech

Re: [PDB Tech] PeeringDB API Discrepency

[Matt Griswold]

>From looking through the code, this was never implemented in v2, we can't
think of any technical reasons so it was likely just my screw up while
migrating the database. There are 84 dupe v4 and 51 dupe v6, so we'll fix
the data and add the constraint.

The deploy automation needs updating from switching hosts before, but we'll
get this out ASAP.

On Oct 4, 2016 3:28 PM, "Arnold Nipper"  wrote:

> On 04.10.2016 18:53, James Bensley wrote:
> > On 13 September 2016 at 11:00, Stefan Pratter  wrote:
> >> Hi,
> >>
> >> Thank you for bringing this to our attention, i have opened a github
> ticket
> >> for it here:
> >>
> >> https://github.com/peeringdb/peeringdb/issues/68
> >>
> >> Stefan
> >
> >
> > Whilst playing the the PDB API some more I have found another "issue".
> > It's not an issue with the API as such but perhaps a feature request;
> > there seems to be no duplication detection.
> >
> > I was trying to automate some stuff at this IX:
> https://www.peeringdb.com/ix/745
> >
> > If you look there are two peers listed at that exchange with the same
> > IPv4 address (AS 8468 and AS 60688, both with IP 195.66.246.11). This
> > kind of thing should be easily detectable. Any scope for this sort of
> > function to be added, where the AS number and company are different
> > there shouldn't be a matching IP, the IP can only be announced by one
> > AS number at an exchange.
> >
>
> Woow ... how could this happen? Afaik it is not possible to add an IP
> twice or even more often.
>
> Thanks for spotting. We will follow up with LINX to sort out the issue.
>
>
> Arnold
> --
> Arnold Nipper
> Chief Technology Evangelist and Co-Founder
>
> DE-CIX Management GmbH | Lindleystrasse 12 | 60314 Frankfurt am Main |
> Germany | www.de-cix.net | Phone +49 69 1730902 22 |
> Mobile +49 172 2650958 | Fax +49 69 4056 2716 |
> arnold.nip...@de-cix.net | Geschaeftsfuehrer Harald A. Summa |
> Registergericht AG Koeln HRB 51135
>
>
> ___
> Pdb-tech mailing list
> Pdb-tech@lists.peeringdb.com
> http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech
>
>
___
Pdb-tech mailing list
Pdb-tech@lists.peeringdb.com
http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech

Re: [PDB Tech] PeeringDB API Discrepency

[Chris Caputo]

Slightly related: It would be nice if in addition to the Peer Name column, 
the IPv4/IPv6 column and the Speed column could have natural sorts 
available.  (Natural sort by IPv4 addr, rather than IPv6 addr, unless IPv6 
is put into a separate column.)

Chris

On Tue, 4 Oct 2016, Matt Griswold wrote:
> Agreed, not sure how that got in, made 
> https://github.com/peeringdb/peeringdb/issues/71 and will check it out / 
> add better tests.
> 
> 
> On Oct 4, 2016 3:46 PM, "Arnold Nipper"  wrote:
>   On 04.10.2016 22:29, Chris Caputo wrote:
>   > On Tue, 4 Oct 2016, Arnold Nipper wrote:
>   >> On 04.10.2016 18:53, James Bensley wrote:
>   >>> On 13 September 2016 at 11:00, Stefan Pratter  
> wrote:
>    Hi,
>   
>    Thank you for bringing this to our attention, i have opened a 
> github ticket
>    for it here:
>   
>    https://github.com/peeringdb/peeringdb/issues/68
>   
>    Stefan
>   >>>
>   >>>
>   >>> Whilst playing the the PDB API some more I have found another 
> "issue".
>   >>> It's not an issue with the API as such but perhaps a feature 
> request;
>   >>> there seems to be no duplication detection.
>   >>>
>   >>> I was trying to automate some stuff at this IX: 
> https://www.peeringdb.com/ix/745
>   >>>
>   >>> If you look there are two peers listed at that exchange with the 
> same
>   >>> IPv4 address (AS 8468 and AS 60688, both with IP 195.66.246.11). 
> This
>   >>> kind of thing should be easily detectable. Any scope for this sort 
> of
>   >>> function to be added, where the AS number and company are different
>   >>> there shouldn't be a matching IP, the IP can only be announced by 
> one
>   >>> AS number at an exchange.
>   >>
>   >> Woow ... how could this happen? Afaik it is not possible to add an IP
>   >> twice or even more often.
>   >>
>   >> Thanks for spotting. We will follow up with LINX to sort out the 
> issue.
>   >
>   > It is possible.  We just had someone do this at the SIX in the past 
> week.
>   > They added our route server addresses, which were registered already 
> in
>   > PeeringDB.
>   >
> 
>   Definitely a severe bug imho and should be fixed immediately. I'm really
>   astonished that this pops up only now.
> 
> 
>   Arnold
>   --
>   Arnold Nipper
>   Chief Technology Evangelist and Co-Founder
> 
>   DE-CIX Management GmbH | Lindleystrasse 12 | 60314 Frankfurt am Main |
>   Germany | www.de-cix.net | Phone +49 69 1730902 22 |
>   Mobile +49 172 2650958 | Fax +49 69 4056 2716 |
>   arnold.nip...@de-cix.net | Geschaeftsfuehrer Harald A. Summa |
>   Registergericht AG Koeln HRB 51135
> 
> 
>   ___
>   Pdb-tech mailing list
>   Pdb-tech@lists.peeringdb.com
>   http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech
> 
> 
> ___
Pdb-tech mailing list
Pdb-tech@lists.peeringdb.com
http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech

Re: [PDB Tech] PeeringDB API Discrepency

[Matt Griswold]

Agreed, not sure how that got in, made
https://github.com/peeringdb/peeringdb/issues/71 and will check it out /
add better tests.

On Oct 4, 2016 3:46 PM, "Arnold Nipper"  wrote:

> On 04.10.2016 22:29, Chris Caputo wrote:
> > On Tue, 4 Oct 2016, Arnold Nipper wrote:
> >> On 04.10.2016 18:53, James Bensley wrote:
> >>> On 13 September 2016 at 11:00, Stefan Pratter  wrote:
>  Hi,
> 
>  Thank you for bringing this to our attention, i have opened a github
> ticket
>  for it here:
> 
>  https://github.com/peeringdb/peeringdb/issues/68
> 
>  Stefan
> >>>
> >>>
> >>> Whilst playing the the PDB API some more I have found another "issue".
> >>> It's not an issue with the API as such but perhaps a feature request;
> >>> there seems to be no duplication detection.
> >>>
> >>> I was trying to automate some stuff at this IX:
> https://www.peeringdb.com/ix/745
> >>>
> >>> If you look there are two peers listed at that exchange with the same
> >>> IPv4 address (AS 8468 and AS 60688, both with IP 195.66.246.11). This
> >>> kind of thing should be easily detectable. Any scope for this sort of
> >>> function to be added, where the AS number and company are different
> >>> there shouldn't be a matching IP, the IP can only be announced by one
> >>> AS number at an exchange.
> >>
> >> Woow ... how could this happen? Afaik it is not possible to add an IP
> >> twice or even more often.
> >>
> >> Thanks for spotting. We will follow up with LINX to sort out the issue.
> >
> > It is possible.  We just had someone do this at the SIX in the past week.
> > They added our route server addresses, which were registered already in
> > PeeringDB.
> >
>
> Definitely a severe bug imho and should be fixed immediately. I'm really
> astonished that this pops up only now.
>
>
> Arnold
> --
> Arnold Nipper
> Chief Technology Evangelist and Co-Founder
>
> DE-CIX Management GmbH | Lindleystrasse 12 | 60314 Frankfurt am Main |
> Germany | www.de-cix.net | Phone +49 69 1730902 22 |
> Mobile +49 172 2650958 | Fax +49 69 4056 2716 |
> arnold.nip...@de-cix.net | Geschaeftsfuehrer Harald A. Summa |
> Registergericht AG Koeln HRB 51135
>
>
> ___
> Pdb-tech mailing list
> Pdb-tech@lists.peeringdb.com
> http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech
>
>
___
Pdb-tech mailing list
Pdb-tech@lists.peeringdb.com
http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech

Re: [PDB Tech] PeeringDB API Discrepency

[Chris Caputo]

On Tue, 4 Oct 2016, Arnold Nipper wrote:
> On 04.10.2016 18:53, James Bensley wrote:
> > On 13 September 2016 at 11:00, Stefan Pratter  wrote:
> >> Hi,
> >>
> >> Thank you for bringing this to our attention, i have opened a github ticket
> >> for it here:
> >>
> >> https://github.com/peeringdb/peeringdb/issues/68
> >>
> >> Stefan
> > 
> > 
> > Whilst playing the the PDB API some more I have found another "issue".
> > It's not an issue with the API as such but perhaps a feature request;
> > there seems to be no duplication detection.
> > 
> > I was trying to automate some stuff at this IX: 
> > https://www.peeringdb.com/ix/745
> > 
> > If you look there are two peers listed at that exchange with the same
> > IPv4 address (AS 8468 and AS 60688, both with IP 195.66.246.11). This
> > kind of thing should be easily detectable. Any scope for this sort of
> > function to be added, where the AS number and company are different
> > there shouldn't be a matching IP, the IP can only be announced by one
> > AS number at an exchange.
> 
> Woow ... how could this happen? Afaik it is not possible to add an IP
> twice or even more often.
> 
> Thanks for spotting. We will follow up with LINX to sort out the issue.

It is possible.  We just had someone do this at the SIX in the past week.  
They added our route server addresses, which were registered already in 
PeeringDB.

Chris
___
Pdb-tech mailing list
Pdb-tech@lists.peeringdb.com
http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech

Re: [PDB Tech] PeeringDB API Discrepency

[James Bensley]

On 13 September 2016 at 11:00, Stefan Pratter  wrote:
> Hi,
>
> Thank you for bringing this to our attention, i have opened a github ticket
> for it here:
>
> https://github.com/peeringdb/peeringdb/issues/68
>
> Stefan


Whilst playing the the PDB API some more I have found another "issue".
It's not an issue with the API as such but perhaps a feature request;
there seems to be no duplication detection.

I was trying to automate some stuff at this IX: https://www.peeringdb.com/ix/745

If you look there are two peers listed at that exchange with the same
IPv4 address (AS 8468 and AS 60688, both with IP 195.66.246.11). This
kind of thing should be easily detectable. Any scope for this sort of
function to be added, where the AS number and company are different
there shouldn't be a matching IP, the IP can only be announced by one
AS number at an exchange.


Cheers,
James.
___
Pdb-tech mailing list
Pdb-tech@lists.peeringdb.com
http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech

Re: [PDB Tech] PeeringDB API Discrepency

[Stefan Pratter]

Hi,

Thank you for bringing this to our attention, i have opened a github 
ticket for it here:


https://github.com/peeringdb/peeringdb/issues/68

Stefan

On 09/09/16 17:29, James Bensley wrote:

Hi All,

When using the API to look at a random network I see a possible
difference in behavior between the API and the web interface.

https://www.peeringdb.com/api/net/21

info_prefixes4": null, "info_prefixes6": 0

Here we see it says "0" for both values: https://www.peeringdb.com/net/21

Has the user/network actually set "0" for v6 prefix count and given no
value for the v4 count so the DB is returning 0 and null respectively,
or is it a discrepancy in behavior; for v6 if there is no value is the
DB defaulting to 0 value but for v4 defaulting to null?

Sorry if this is a known issue.

Cheers,
James.
___
Pdb-tech mailing list
Pdb-tech@lists.peeringdb.com
http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech


___
Pdb-tech mailing list
Pdb-tech@lists.peeringdb.com
http://lists.peeringdb.com/cgi-bin/mailman/listinfo/pdb-tech