search

ietf-tools / datatracker

The day-to-day front-end to the IETF database for people who work on IETF standards.
https://datatracker.ietf.org
BSD 3-Clause "New" or "Revised" License
578 stars 349 forks source link

New RFCs are not shown in Datatracker even though they are available #3487

Open ietf-svn-bot opened 2 years ago

ietf-svn-bot commented 2 years ago

type_defect | by paul.hoffman@icann.org

An RFC announcement was sent out over half an hour ago. The RFC is available on the RFC Editor's site, but when accessing the URL in the Datatracker, it says "This RFC is not currently available online."

Proposal: mirror the RFC Editor's RFC repo every five minutes so that people going to the Datatracker actually get the RFC, not just the status.

Issue migrated from trac:3487 at 2022-03-04 09:22:12 +0000

ietf-svn-bot commented 2 years ago

@rjsparks@nostrum.com changed priority from n/a to medium

ietf-svn-bot commented 2 years ago

@rjsparks@nostrum.com changed status from new to accepted

ietf-svn-bot commented 2 years ago

@rjsparks@nostrum.com commented

You've made us guess at where the problem actually lies. I'm assuming RFC 9157. If that's not true, let me know what you were actually looking at so I can analyze with better information. In the future, please watch for opportunities to be more specific.

There's a mistaken assumption here that the datatracker learns about RFCs by polling.

Instead, it relies on the rfc-editor's systems doing a push (with a safety net of a daily sync in case a push is missed).

The timestamps on the pushes received for this are interesting in that we are still on a timezone-naive database (this is scheduled to be changed) and the different things pushed (or received) assumed a different timezone, or more likely a date bug.

>>> DocEvent.objects.filter(doc__name='draft-ietf-dnsop-dnssec-iana-cons').order_by('-time').values_list('time',flat=True)[:3]
<QuerySet [datetime.datetime(2021, 12, 1, 16, 32, 56), datetime.datetime(2021, 11, 30, 16, 33, 3), datetime.datetime(2021, 11, 29, 8, 5, 53)]>

(these correspond to what you would see at https://datatracker.ietf.org/doc/rfc9157/history/, and at the time I'm composing this it is ~1935 Pacific on the 30th).

There is no reality in which the first timestamp makes sense, so we have a bug to chase there - lets assume for the moment (subject to research) that it's an off-by-one-day error, and that the timestamp is otherwise in pacific time, which would line up with the "over half an hour ago" description in the report. I'll work with the RPC to find out where this happened.

So, if the two events most recent events really came together (likely) the push was roughly 4:33pm Pacific. The ietf-announce list sent this (to me anyhow) at 4:32, so the datatracker knew about this in a very timely matter.

Perhaps you were getting a page with the metadata, but at the place where it would show text, you were getting the message you describe? I think I can infer that on a re-read of your post. That would point to a different place in the infastructure to look for an issue.

ietf-svn-bot commented 2 years ago

@rjsparks@nostrum.com changed _comment0 which not transferred by tractive

ietf-svn-bot commented 2 years ago

@rjsparks@nostrum.com changed _comment1 which not transferred by tractive

ietf-svn-bot commented 2 years ago

@rjsparks@nostrum.com commented

outside the datatracker the rfc texts are rsynced hourly. I'll look later to see if we are fetching specific RFC content on the push, and if not, it's something we could reasonably add.