MX record for discourse.osgeo.org is not propagating properly

[strk]

As reported in ​https://trac.osgeo.org/osgeo/ticket/2306#comment:54 and in ​https://trac.osgeo.org/osgeo/ticket/2306#comment:55 there are at least 2 name servers from internet service providers (ovh.net and ehiweb.it) that do now know (after 7 days) about the existance of an MX record for discourse.osgeo.org

I suspect there is something to improve from our side to make propagation of these records faster.

I confirm clearing caches (like installing a new nameserver and using it from scratch) is able to see the MX record, so it is a matter of cache lifetimes.

[strk]

My (new) server and my desktop client are still not getting the MX record for discourse.osgeo.org as of today. Was the zone version updated ?

[strk]

It would be nice to put the DNS zone under Ansible too...

[strk]

Even osgeo7 resolver isn't getting the MX record at the moment:

tech_dev@osgeo7:~$ nslookup
> set type=mx
> discourse.osgeo.org
Server:         127.0.0.53
Address:        127.0.0.53#53

Non-authoritative answer:
discourse.osgeo.org     canonical name = web.osgeo9.osgeo.org.

Authoritative answers can be found from:
>

[strk]

I see no output from these commands:

dig @213.186.33.99 -t mx discourse.osgeo.org| grep 'MX' | grep -v '^;'
dig @2001:41d0:3:163::1 -t mx discourse.osgeo.org| grep 'MX' | grep -v '^;'

Both addresses are from ovh.net The ehinet addresses wont' respond from the outside

[strk]

May be cloudflare related: ​https://community.cloudflare.com/t/mx-record-didnt-updating/353209

[robe]

Replying to strk:

May be cloudflare related: ​https://community.cloudflare.com/t/mx-record-didnt-updating/353209

The person on that ticket made two mistakes

1) They tried to put in the name of their domain which ends up doing something silly like

mydomain.com.mydomain.com

2) They used Proxy instead of DNS only

By contrast ours looks like this:

so this has nothing to do with our problem.

I'm also puzzled why it's not every dns in the world having this issue with our setup. Did those admins tell you what the upstream problem is? If they didn't give specifics, I'd assume they are just trying to pass on the problem to someone else cause they have no clue.

[robe]

I forgot to show the config of meet.osgeo.org in that screen shot.

A meet  	140.211.15.5   DNS only

But anyway I can't replicate the issue on any of the servers I have access to or any of the MX checkers I have tried online. I tried the below:

host -t mx discourse.osgeo.org

on an amazon server, atlantic.net, osuosl.org server, my comcast.net

and all report

discourse.osgeo.org mail is handled by 1 meet.osgeo.osuosl.org.
discourse.osgeo.org mail is handled by 10 meet.osgeo.org.

I also tried - ​https://mxtoolbox.com/SuperTool.aspx?action=mx%3adiscourse.osgeo.org&run=toolpage

and it came back with the same answer in addition to whining about not having a DMARC policy which is a separate issue.