This message was deleted.
# rke
a
This message was deleted.
l
Hey @quiet-fountain-61995, does your installation use custom certs by any chance? Or is it using the default certs?
q
Hey @limited-pizza-33551 its the default certs. i figured out what the issue was after two weeks. of pain staking debugging. it was one of the entries in our search domain filed of the
/etc/resolv.conf
file. The file liked like the below file :
nameserver 10.0.0.0
search production.svc.cluster.local svc.cluster.local cluster.local <http://dc.mydomain.com|dc.mydomain.com>
the issue was that for the last entry
<http://dc.mydomain.com|dc.mydomain.com>
when coredns tried to resolve
<https://google.com>
with it it was returning a
NOERROR
instead of a
NXDOMAIN.
and so the search loop was closed and it resulted in a
tls error
because there was no active DNS resolver at the
<http://dc.mydomain.com|dc.mydomain.com>
search entry. the fix was to either remove the search entry form the servers network manager or setup a dns resolver at the endpoint which is how we wanted it to work. and now our dns tls queries are resolved. So looks like its more of a coredns issue than rancher.