Replies: 10 comments 1 reply
-
Re PoolAvailableRule, even with the following the pool is stuck in 'down':
|
Beta Was this translation helpful? Give feedback.
-
With a classic healthcheck, the downstream is correctly marked as 'up' after a few seconds. Config snippet:
|
Beta Was this translation helpful? Give feedback.
-
Is there anything in the logs when |
Beta Was this translation helpful? Give feedback.
-
I tried this and there I see checks:
The checks are done wit a period of 30s, the default value of
|
Beta Was this translation helpful? Give feedback.
-
But when I use a non-routable address I see this, note the missing
So the current hypothesis is that errors sending the healthcheck probe are not handled correctly, while timeouts do cause the right state change. |
Beta Was this translation helpful? Give feedback.
-
Adding a call to
|
Beta Was this translation helpful? Give feedback.
-
Clearly a bug then, I'll look into it! |
Beta Was this translation helpful? Give feedback.
-
I'm working on a fix for this in #13834 |
Beta Was this translation helpful? Give feedback.
-
Opened #13837 to track this as an issue. |
Beta Was this translation helpful? Give feedback.
-
Shot in the dark, but the "broken after a reboot but OK after subsequent restart" symptom when paired with IPv6 servers might just be the good old DAD race when dnsdist is started too quickly while the interface has a "tentative" IPv6 address. See here and the original blog. |
Beta Was this translation helpful? Give feedback.
-
I have dnsdist in a setup which some people would probably call "strange". dnsdist sits on my router VM to provide recursive DNS for my LAN, and I want to use these downstreams: nextdns (mostly for tracking blocking), a local pdns-recursor (fallback), and a local dnsmasq (for LAN domains).
This setup generally works, but on a reboot, the nextdns servers are marked down and never make it into the "up" state. Here is a boot log excerpt:
For the servers, I have this config:
Note that my actions have this (excerpt):
Manually restarting dnsdist later makes the nextdns upstreams work.
Obviously, during reboot the WAN interfaces are down, so the upstreams are unreachable. But I was hoping for the lazy healthcheck to recover after some time, but that seems not to happen. Any ideas?
Beta Was this translation helpful? Give feedback.
All reactions