0

DNS Outage with dns1.nextdns.io?

I experienced a dns outage this morning around 6:45am PST on Dec 15. Looks like it is still ongoing and I've failed over to my secondary nextdns host.

ping.nextdns.io shows:

  vultr-sea          1 ms

  do-sfo            22 ms

  vultr-sjc         22 ms

  anexia-lax        32 ms

  vultr-lax         34 ms

■ zepto-lax         38 ms  (secondary)

  vultr-chi         49 ms

  vultr-dal         51 ms

  anexia-dal        60 ms

  anexia-atl        60 ms

dns1.nextdns.io error (primary)

 

Questions:

Was there an outage? Is it ongoing? 

How long does the backend take to failover? In my case, I had an outage for several minutes before it failed over. Though this could be due to my client (dnscrypt-proxy).

How do I get on the closer servers? You can guess that I'm in Seattle, but I'm always tied to the LA servers.

22 replies

null
    • seashell_summer
    • 3 yrs ago
    • Reported - view

    I experienced the same thing. I'm using a DNSStamp via DNSCrypt, and I thought my whole internet was down.

    Turns out, it was just 45.90.28.227, based on my ping tests. 45.90.30.227 was up and running.

    This does seem to point out a problem with using DNSStamps. I decoded my stamp using https://dnscrypt.info/stamps/ and only one DNS server was shown in the list.

    Does there need to be two DNSStamps in order to load balance properly between the two DNS servers provided by NextDNS?

      • seashell_summer
      • 3 yrs ago
      • Reported - view

      JB Looks like according to the DNSStamp spec, only one server is encoded per stamp:

      https://dnscrypt.info/stamps-specifications

      Can NextDNS update the admin page to display a stamp for both servers, so that I can load balance correctly (and ideally not go down when a single server is down)?

      • seashell_summer
      • 3 yrs ago
      • Reported - view

      JB For anyone else looking to do this manually (until NextDNS adds support for generating all DNSStamps), you can use https://dnscrypt.info/stamps/.

      Take your existing stamp, and plug it in. It should populate the IP address, hostname, etc.

      Simply update the IP address with the other addresses listed on your account page, making sure that you stick IPv6 addresses in brackets - ex: "[foo:bar:...]".

      This let me go from my one DNSStamp to 4, which covers both DNS servers now and supports IPv6!

    • mango_squirrel
    • 3 yrs ago
    • Reported - view

    Now my primary has gone to HKG. What's the logic being used here? Shouldn't I hit vultr-sea automatically or vultr-sjc, etc?

     

      vultr-sea     2 ms

      vultr-sjc    23 ms

      do-sfo       24 ms

      vultr-lax    35 ms

    ■ zepto-lax    38 ms  (secondary)

      anexia-lax   39 ms

      vultr-chi    50 ms

      vultr-dal    58 ms

      anexia-atl   60 ms

      anexia-dal   62 ms

    ■ rix-hkg     241 ms  (primary)

    • olivier
    • 3 yrs ago
    • Reported - view

    Please use the https://nextdns.io/diag tool and give me the report id.

    Note there is no outage ongoing, it’s most likely a routing issue.

    • seashell_summer
    • 3 yrs ago
    • Reported - view

    I'm seeing 100% packet loss from the `45.90.28.***` servers.

    I'm running the diag tool as we speak, although it's taking a very long time.

    I've got a ping test from OpenWRT from the past 1 hour, showing that the *28* IPs are down. The second set of IPs for *28* and *30* I added just a few minutes ago, which is why there's little data.

      • Chris_Leidich
      • 3 yrs ago
      • Reported - view

      JB Experiencing a very similar issue staring about 90 mins ago, and another person in Hong Kong is also experiencing it. Looks like they're trying to sort it out, Olivier's been responding, but definitely seems to be a worldwide network of gremlins for this one.

      • seashell_summer
      • 3 yrs ago
      • Reported - view

      Chris Leidich Good to know it's widespread (not a sentence I ever thought I'd type...)

      For what it's worth, the diag report failed to post, but here's the output from my terminal:

      • Chris_Leidich
      • 3 yrs ago
      • Reported - view

      JB Yeah, I work in IT, and find myself saying "well, at least it's broken for everyone!!!" more often than would be allowed in most vocations.  :)

      Your diag's traceroute output for 45.90.28.0 dies at the same place mine does, 85.202.80.3.

    • plttn
    • 3 yrs ago
    • Reported - view

    I'm also seeing this issue. It also looks like YogaDNS is only using 45.90.28.***, which equals complete outage for me. 

    • Neil
    • 3 yrs ago
    • Reported - view

    Yup, seems to be a routing issue to .28.

    • Neil
    • 3 yrs ago
    • Reported - view

    hmm.  Ran the reporting tool Olivier.  It errored out trying to Post the report to you....

    Post unsuccessful: status 400
    {"error":"0: instance requires property \"Primary\"\n"}

    • plttn
    • 3 yrs ago
    • Reported - view

    Posting failed here as well. 

    • ray_toth
    • 3 yrs ago
    • Reported - view

    toll you recamend cant be run on windows 10 64 bit tell me not 64 bit ready and not allowed to run and block it im in desplaines il here a ping test i did

    Pinging 45.90.28.27 with 32 bytes of data:
    Request timed out.
    Request timed out.
    Request timed out.
    Request timed out.

    Ping statistics for 45.90.28.27:
        Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),

    • Qadhi
    • 3 yrs ago
    • Reported - view

    dns.nextdns.io is down and unable to load dns stamps. I checked it from my PC, router and few websites that check uptime status and all showing dns.nextdns.io as down.

      • Shashank
      • 3 yrs ago
      • Reported - view

      Qadhi Same here India servers also not working . Requests going to anexia-lax

      Hope gets resolved soon

      • olivier
      • 3 yrs ago
      • Reported - view

      Kruxie this sound like a different issue. Can you please use https://nextdns.io/diag and send the diag id?

    • olivier
    • 3 yrs ago
    • Reported - view

    We are having an issue with a provider who misconfigured our route which is affecting our primary DNS anycast IP for some locations. We are working with them to solve the issue ASAP. 

    • olivier
    • 3 yrs ago
    • Reported - view

    The issue should be fixed. Those who had issues, please use the https://nextdns.io/diag tool and send me the posted ID (not the output) in private message.

    • seashell_summer
    • 3 yrs ago
    • Reported - view

    My issue appears to have been fixed, sent my diag output via PM.

    • Chris_Leidich
    • 3 yrs ago
    • Reported - view

    Sent my diag info a few minutes ago, working well today.

    • Neil
    • 3 yrs ago
    • Reported - view

    Thanks JB ! 

    I created a new static server and added the corresponding new stamp to the new static section.

Content aside

  • 3 yrs agoLast active
  • 22Replies
  • 601Views
  • 9 Following