|
There are thousands of sites and services on the ‘net that offer domain name whois lookup services. As of last night, many of them may have stopped working.
Why?
Many of them rely on fairly rudimentary software that parses the whois from Verisign (for .com and .net) and then relays the query to the registrar whois. The site or service then displays the whois output from the registrar’s whois server to you.
For “thick” registries, like .biz or any new TLD, the whois is always served directly by the registry whois server, so there’s no “referral” or extra parsing. Additional fields in thick whois, therefore, shouldn’t have much impact on most 3rd party whois lookup services.
So what happened?
Verisign, along with most other gTLD registries, updated their whois output last night to include some new fields namely the registrar abuse contacts:
Registrar Abuse Contact Email: [email protected]
Registrar Abuse Contact Phone: +xxx.xxxxx
As a result of the changes, a lot of software is currently “broken” as it simply cannot cope with the new output.
Is whois broken?
No. It’s working fine. The issue is with the software clients and how they are written. A lot of them were written years ago and have hardcoded in certain settings.
From what I can see the changes in whois output only seem to be impacting .com and .net whois lookups and only with *some* software. Doing whois lookups from my Mac’s command line, for example, still works fine for “thick” registries, but is failing miserably for .com and .net.
Will this impact registrars?
That depends.
Registrars generally do NOT use whois to check if domain names are available. Registrars tend to use EPP checks, zone files or other tools to see if a domain is taken or not. (And no, using DNS checks would be a terrible idea!)
There is a possible impact on *some* registrars when it comes to domain name transfers of .com and .net domain names.
However any impact is going to be short lived, as registrars will be aware of the issues and will update their software to handle the changes.
UPDATE: Here’s a fix for jwhois and similar software from Chris Pelling
In /etc/jwhois.conf (the location might vary)
find the line referencing “verisign-grs”:
}<br /> “.*.verisign-grs.com” {<br /> whois-redirect = ".*Whois Server: (.*)";<br /> }
and replace it with:
}<br /> “.*.verisign-grs.com” {<br /> whois-redirect = ".*Registrar WHOIS Server: (.*)";<br /> }
The above changes *should* help, though it will depend on which software you are using.
Thanks to Paul Goldstone for mentioning the issue to me first :)
Sponsored byRadix
Sponsored byVerisign
Sponsored byCSC
Sponsored byDNIB.com
Sponsored byWhoisXML API
Sponsored byVerisign
Sponsored byIPv4.Global
So in summary, the issue was not with rudimentary software not being able to cater for additional data, rather a crucial key in the whois output from Verisign that changed. Much like renaming a column in a database and not communicating it. We need to remember not all parties with a legitimate interest in whois will have access to registrar tools.
From where AA419 sits, we saw the failure, debugged the core issue, also as posted up at ALAC. What is rather bizarre, is that this change was not properly communicated and ICANN distancing themselves from it:
http://blog.aa419.org/2017/07/26/security-broken-whois-it/
From what I can see it’s not those two added fields that’s the problem, it’s the changed name of the field that tells WHOIS software what server to query for more information from the registrar. Changing that without notification and lead time was a dumb-as-a-rock move.