DNS Not working?

🛎️ Get Help
1

My Hub is throwing errors for a few hostnames:
Failed to refresh token: java.net.UnknownHostException: wbsapi.withings.net
ava.net.UnknownHostException: platform.cloud.petsafe.net: Temporary failure in name resolution on line 485 (method refreshDevices)
java.net.UnknownHostException: v2.api.whisker.iothings.site: Temporary failure in name resolution

My hub is configured to use 8.8.8.8. I'm not sure what's going on, they all resolve just fine on the same network. See below from a Windows PC on the same network. After I rebooted my hub it's working fine again.

nslookup v2.api.whisker.iothings.site
Server:  dns.google
Address:  8.8.8.8

Non-authoritative answer:
Name:    v2.api.whisker.iothings.site
Addresses:  13.225.230.121
          13.225.230.47
          13.225.230.45
          13.225.230.100

nslookup platform.cloud.petsafe.net
Server:  dns.google
Address:  8.8.8.8

Non-authoritative answer:
Name:    platform.cloud.petsafe.net
Addresses:  52.202.253.74
          3.212.20.237
          107.20.21.169

nslookup wbsapi.withings.net
Server:  dns.google
Address:  8.8.8.8

Non-authoritative answer:
Name:    ws.withings.net
Addresses:  2a01:8200:1001:1:89:30:121:170
          89.30.121.170
Aliases:  wbsapi.withings.net
2

Well, I don’t know about Withings, but I do see this occasionally on my SureFlap Pet Door integration and my Litter Robot (“Whisker”) integration. It doesn’t persist for long, I just shrug my shoulders and assume it’s either a load issue on their servers, or that they are taking backups, or that they are installing new host software.

I use the same 8.8.8.8 DNS that you do.

I only see it on the very few cloud integrations that I have. We have dual internet service (Xfinity Gigabit and AT&T DSL), and I’ve seen it on both. It clears up for me without rebooting.

Now, I’m not doing active development on that hub (C-7) with these cloud integrations like you may be. I do my development on my C-5 that has no Z-Wave or Zigbee devices, radios turned off. My production C-7 only reboots when platform firmware releases occur - last reboot 17 days ago.

3

This wasn’t doing any development. This was for normal use. It lasted for a few hours before I rebooted. Then it was instantly fixed.

4

It may be that there aren’t enough UDP buffers allocated for DNS. It’s only relatively recently that many cloud integrations have become common on Hubitat hubs. Perhaps some configuration tweaks of the TCP/IP and UDP network stacks are needed.

1 Like
5

Maybe? Not too familiar with UDP buffers as it relates to DNS. For me though I’ve had a bunch of cloud integrations for years. Not sure if this is a newer firmware change or something else but I’ve never notified it happen before.

6

DNS uses UDP as its transport protocol. I’ve seen it get bogged down when I was doing Unix internals years ago if number of buffers wasn’t high enough to support the traffic.

It’s configurable, maybe @gopher.ny needs to tweak. Or, perhaps there are some bugs in the protocol stack, causing buffers to disappear. Seen that, too. Very difficult to troubleshoot.

2 Likes