[arin-tech-discuss] Issue for Delegated Users within ARIN's RPKI Repository - Outage Report

[Jay Borkenhagen]

Mark Kosters writes:
 > On 11/23/20, 11:12 PM, "Christopher Morrow" <morrowc.lists at gmail.com> wrote:
 >     
 >     Do you all document what is tested? so either RP software folk can
 >     integrate similar tests (or suggest others), and/or so actual ARIN
 >     users of the RPKI can test their (or instrument their) repository
 >     collections?
 >     
 > The validator collection that we test against is documented at https://www.arin.net/resources/manage/rpki/#supported-tested-rpki-validators . There is regression testing done with every release which includes successful runs by the validators. The check that we did not perform is to ensure the validators actually fetched at least one ROA from a delegated repo. That is now a part of the test plan.
 > 
 > Additionally we run a validator every time we do a update to the repository.
 > 

Hi Mark,

Hmm, I think it would be great if ARIN could publish more details of
the regression tests you perform.

For example, I don't think you can expect that each of your supported
validators will produce the same set of VRPs from the objects
descending from ARIN's TAL.  In fact, different versions of the same
validator do not produce identical results, so I do not think you can
expect different validators to do so (at least not now, and maybe
never).  I'm curious what ARIN considers a 'successful' run to be.

Also, is it enough to have fetched at least one ROA from a delegated
repo, or should one or more VRPs have resulted from the retrieved ROA?


By ARIN publishing their regression test details, other RIRs could
learn from ARIN's experience.  Conversely, ARIN could learn from
regression test details the other RIRs would also publish.  Some
details might well be unique to ARIN's specific environment, but many
should be universal.  RP developers and service providers would also
benefit from knowing these test details.

Thanks.

						Jay B.