Mx Record always get lost

My MX Record get lost some times. Is there something wrong?

have the same problem

I’m also losing records at random. All types of entries that use “@” for the name. Records of subdomains don’t get deleted.

I just collected some timestamped data about this bug. Hope this helps.

I’m setting my record with dynv6.sh afterward in querying the API every 15 seconds to narrow down when the AAAA record disappears. After 40 seconds the record is gone.

$ rm -f .dynv6.addr6 && token=${YOUR_TOKEN} \
   /usr/local/bin/dynv6.sh mydomain.net; echo; date \
   && while sleep 10; \
      do date; \
      curl -s \
      -H "Authorization: Bearer ${YOUR_TOKEN}" \
      -H "Accept: application/json" https://dynv6.com/api/v2/zones/[redacted]/records \
      | jq '.[] | .type + " " + .data'; \
     done
addresses updatedaddresses unchanged
Thu Oct 29 13:57:10 CET 2020
Thu Oct 29 13:57:20 CET 2020
"CNAME mydomain.net."
"TXT [redacted]"
"AAAA 2001:....redacted...."
Thu Oct 29 13:57:30 CET 2020
"CNAME mydomain.net."
"TXT [redacted]"
"AAAA 2001:....redacted...."
Thu Oct 29 13:57:40 CET 2020
"CNAME mydomain.net."
"TXT [redacted]"
"AAAA 2001:....redacted...."
Thu Oct 29 13:57:51 CET 2020
"CNAME mydomain.net."
"TXT [redacted]"
"AAAA 2001:....redacted...."
Thu Oct 29 13:58:01 CET 2020
"CNAME mydomain.net."
"TXT [redacted]"
"AAAA 2001:....redacted...."        #RECORD STILL THERE
Thu Oct 29 13:58:11 CET 2020
"CNAME mydomain.net."
"TXT [redacted]"                    #RECORD IS GONE :(
Thu Oct 29 13:58:21 CET 2020
"CNAME mydomain.net."
"TXT [redacted]"

Same problem here. Confirming it only happens to entries that use “@” for the name. All other records are stable.
Is there any progress on this annoying bug? It seems to have occurred first about two weeks ago. Before that, those records were completely stable. Has dynv6 put any process in place that could be responsible for deleting those records?

Thank you for the data. I’ve run this on my machine - and can’t currently reproduce this issue.

I did throw the update against all servers, and validated that all servers respond with the correct answer. You can substitute https://dynv6.com for http://nsX.dynv6.com, but you need to waive a secure connection (so don’t do this with your usual tokens - create a new one and throw it away afterwards).

My best guess is that there was a synchronization error in our DB cluster. We did update/reboot the machines in the meantime a few times, so maybe whatever caused the issue is now gone.

No. The only automatic process which could lead to deleted records is our abuse system. But that would delete whole zones and suspend the user account - so the expected complaint would not be “some records get lost”…

Indeed, the problem appears to have gone away. My records have been stable for several days now, nothing disappeared. You may be right that the reboots have solved the issue.