DNS Failure -
Nameservers for www.newsblur.com:
ns2.dynadot.com returned (SERVFAIL)
ns1.dynadot.com returned (SERVFAIL)
* This is probably temporary, but I thought I’d report it in case it’s helpful.
1 Like
For now, you can get around it by hitting http://199.15.250.228/. I’m switching away from Dynadot the second they’re back up. They are my DNS and my domain registrar. Search Twitter for “dynadot” for kicks.
1 Like
As bad as the DNS failure is, it seems to have one interesting side effect, the frontend UI is way more responsive than usual. Is this because I’m hitting the resources directly or because of reduced load on your servers?
Well, there is rarely any load on the servers. I put the average response time graphs on the dashboard to show the past 24 hours, usually with very little load difference. In fact, the reason that there is a 50% curve to response times is that the feed fetchers are handling the majority of new stories coming in during daylight hours.
But DNS is usually given once per 30 minutes, depending on the TTL (time-to-live) set. It can be every 5 minutes, but it’s still going to be cached for a period of time. DNS has only a tiny overhead to the response time, and it’s distributed, so it’s very rarely my DNS server that you’re hitting.
If somebody knows more about this subject and would like to weigh in, it would be much appreciated.
yeah, if DNS is taking a long time to resolve you have other problems. (of course, you do have other DNS problems, so… maybe?)
At my university, we have a very aggressive proxy server and cache, problem is, it seems to be very slow most of the time. Maybe because I went directly to the IP address, I bypassed the proxy, resulting in better performance?