[atlas] HTTP server configuration fix deployed to v3 anchors

Hi,

Recently we identified a web server misconfiguration on our v3 anchors, 
such that those anchors would not respond as expected to requests with 
their IP literal in the URL. Requests to their FQDN worked as expected.

In other words:

  * http://nl-ams-as3333-3.anchors.atlas.ripe.net:80/ would respond as
    expected, but
  * http://[2001:67c:2e8:3::c100:a4]:80/ would respond with a 404 status
    code.

The configuration was fixed and deployed by around 0900UTC on 16 
January, 2018. V3 anchors now respond correctly to requests issued with 
their IP literal in the URL.

This affected HTTP measurements to v3 anchors as follows:

  * Before the fix was deployed, HTTP measurements to a v3 anchor with
    'Resolve on probe' set to false will have used the IP literal, and
    thus logged a 404 status code.
  * HTTP measurements as above that were running when the fix was
    deployed will have started returning correct results part-way
    through their run. If the measurement is ongoing, no other action
    needs to be taken.
  * HTTP measurements with 'Resolve on probe' set to true always
    operated as expected.


Cheers,
Stephen Strowes
RIPE NCC


-------------- next part --------------
An HTML attachment was scrubbed...
URL: </ripe/mail/archives/ripe-atlas/attachments/20180122/08040bfa/attachment.html>