Hey all,
I seem to be seeing unexpected behaviour using dynamic DNS resolution - any help or advice you can offer would be much appreciated.
I keep seeing messages like this in the log:
Server backend_name/server_name is going DOWN for maintenance (DNS timeout status). 0 active and 0 backup servers left. 0 sessions active, 0 requeued, 0 remaining in queue.
Running a tcpdump on the DNS queries haproxy generates, I typically see successful DNS responses less than 3 seconds before seeing this error. Often I will see this message 3s before the one above (even with a hold valid
value of 30s
), I also see it when the server is apparently marked as DOWN:
backend_name/server_name changed its IP from 10.xx.xx.xx to 10.xx.xx.xx by name-dns/name-dns-01
I also often see this message even if the server hasn’t been reported as down:
Server backend_name/server_name administratively READY thanks to valid DNS answer.
DNS responses typically arrive in under 30ms.
The show stat resolvers
command does not show that any timeouts occurred:
sent: 10797
valid: 1034
update: 493
cname: 0
cname_error: 0
any_err: 869
nx: 0
timeout: 0
refused: 0
other: 0
invalid: 0
too_big: 0
truncated: 0
outdated: 260