Prefix length for the "pingable" attribute
Hi WG, The documentation unequivocally states the value of the "pingable" attribute in route[6] objects must include a prefix length: $ whois -h whois.ripe.net -- -v route [...] pingable Allows a network operator to advertise an IP address of a node that should be reachable from outside networks. This node can be used as a destination address for diagnostic tests. The IP address must be within the address range of the prefix containing this attribute. An address prefix is represented as an IPv4 address followed by the character slash "/" followed by an integer in the range from 0 to 32. The following are valid address prefixes: 128.9.128.5/32, 128.9.0.0/16, 0.0.0.0/0; and the following address prefixes are invalid: 0/0, 128.9/16 since 0 or 128.9 are not strings containing four integers. $ whois -h whois.ripe.net -- -v route6 [...] pingable Allows a network operator to advertise an IP address of a node that should be reachable from outside networks. This node can be used as a destination address for diagnostic tests. The IP address must be within the address range of the prefix containing this attribute. <ipv6-address>/<prefix> Therefore I added mine as /32 and /128. That ended up in me getting complaints that I broke some people's scripts, which were assuming that value of this attribute should be a "bare" IP address. (For example: http://ip.v6net.ru/archives/1764) Contradictory to what the documentation says, the e-mail robot happily accepted my update to remove the prefix length. So hopefully people's scripts are unbroken now. This sparked my curiosity, though: - What exactly is the purpose of the prefix length? - How should one interpret a value that has a prefix length other than /32 (in "route" objects) or /128 (in "route6" objects)? - Why does the documentation say it is mandatory when it clearly is not? - And finally, am I really the only LIR hostmaster in the RIPE that actually do RTFM? :-) $ curl -s ftp://ftp.ripe.net/ripe/dbase/split/ripe.db.route{,6}.gz | zgrep -c '^pingable:.*/' 0 Tore
On Sat, Mar 15, 2014 at 02:53:09PM +0100, Tore Anderson wrote:
Hi WG,
The documentation unequivocally states the value of the "pingable" attribute in route[6] objects must include a prefix length:
<snip> RFC5943 [1] states that the value of the 'pingable:' attribute is either a regular IPv4 or IPv6 address (rfc2622 / rfc4012), and makes absolutely no mention of including a prefix length. It probably is a good move to update the RIPE Whois Server documentation and ensure the value can only be a valid IPv4 or IPv6 address, without prefix length. Nice catch Tore! Kind regards, Job [1] - https://tools.ietf.org/html/rfc5943 "RPSL Pingable Attribute"
On Sat, Mar 15, 2014 at 03:44:16PM +0100, Job Snijders wrote:
The documentation unequivocally states the value of the "pingable" attribute in route[6] objects must include a prefix length:
<snip>
Nice catch Tore!
When this commit is merged the documentation will be aligned with expected behaviour. https://github.com/RIPE-NCC/whois/pull/189 Have a good weekend, and keep hunting for bugs! :-) Job
Dear Job and Tore Thanks for pointing this out and for the patch. It seems there was some misinterpretation of the RFC when this feature was implemented. The documentation states it must be an address prefix. But the implementation accepted either an address or an address prefix. Thanks for the patch Job. The developers will check it to make sure it now accepts only an address and fails on an address prefix...which is what the RFC states. Then we will update the documentation to reflect this. regards Denis Walker Business Analyst RIPE NCC Database Team On 15/03/2014 21:32, Job Snijders wrote:
On Sat, Mar 15, 2014 at 03:44:16PM +0100, Job Snijders wrote:
The documentation unequivocally states the value of the "pingable" attribute in route[6] objects must include a prefix length: <snip>
Nice catch Tore! When this commit is merged the documentation will be aligned with expected behaviour.
https://github.com/RIPE-NCC/whois/pull/189
Have a good weekend, and keep hunting for bugs! :-)
Job
Hi, On Sat, Mar 15, 2014 at 02:53:09PM +0100, Tore Anderson wrote:
- And finally, am I really the only LIR hostmaster in the RIPE that actually do RTFM? :-)
I think you are :-o (I have used "pingable:" in the past, but have never noticed that there is a *mandatory* prefix length attached to it... and consequently, never added one) Gert Doering -- NetMaster -- have you enabled IPv6 on something today...? SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (0)89/32356-444 USt-IdNr.: DE813185279
participants (4)
-
Denis Walker
-
Gert Doering
-
Job Snijders
-
Tore Anderson