DNS: TXT Records created but never propigate
Hi, I'm creating new txt dns records in my domains and as of recent they never propigate. I can login and create one (or use the api) and when I query cloudflare dns none of the new txt records come back. Existing ones will return, but new ones do not. I am using the free plan, clearly I will never need real help so I can't contact cloudflare. Anyone have any advice for addressing what's going on? Maybe there is an outage? I'm not seeing any configuration issues and it's occuring with all 5 of our domains. The SOA for all 5 domains are with Cloudflare and this was working at least a couple of weeks back.
See attached screenshots from cloudflare dns management panel and results from dns.inbrowser.app (I get the same results from google dig, nslookup against 1.1.1.1, etc).
10 Replies
can you share the domain name?
I'm not super jazzed about sharing it publicly, but apparently that's what is required. Check out eonmn.com - I just created a new test record there (my other test is in one of our other domains).
fwiw I tried making a txt on one of my zones and instantly propagated, I'd guess your nameservers changed or you're checking against cache, but a lot easier to debug if I can query the zone directly
You can delete the message now if you want. What was the record you created?
testtesttest content is "test"
I see it querying auth
and I see it in that tool, even
what the heck - ok let me try again
Ok thank you for checking, I'm going to go through the whole thing start to finish again with my other domain where I have it all scripted and verify nothing else is broken. I'll let you know ASAP - thx!
Worth noting the best way to check is against your auth. nameservers, skipping all cache
Look them up via dig
dig domain.com ns, see ns results, or find them in dash under dns, then query direct, ex: dig cloudflare.com @ns3.cloudflare.com txt
any web tool or your local dns is going to have cache in front. CF DNS Propagation takes ~5-10s or so, so if you beat it to the punch you'll be stuck behind cacheOk I think found the problem, it looks like the process tests for the record and it only retries for a short-ish period of time. What I didn't catch is that it purges the name once it gives up trying - I think maybe the propagation time has slowed down beyond the threshold where it was working and, now, it's cleaning up it's tracks before I can debug, for now I'll see if I can boost the delay for checks and see if that addresses the issue. Apologies for crying wolf on this one, thanks for the really quick help @Chaika!