[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Assignment of /etc/resolv.conf on pods


due to our application needs, we have a couple of openshift nodes running a BIND nameserver on port 53 which is used instead of the standard dnsmasq. I see that the firs nameserver on /etc/resolv.conf within pods instantiated on these nodes is, where on nodes running dnsmasq the first one is the physical node addresses. And in both cases, there is a second nameserver listed with the physical node addresses.

Our problem is that, while the works well for resolving local addresses, whenever a name not existing within cluster.local is queried it returns ";; Got recursion not available from, trying next server", and although this seems fine on command line, looks like our applications (java ones) are unable to handle switching to the next nameserver. I hope developers will fix this, but in the meantime I would like to try some workaround.

I have two simple solutions in mind

  1. Enable recursion on the OpenShift built-in dns server
  2. Force the first nameserver on pods to be the phyisical host (which is able to forward queries to cluster.local)
is possible (and how) to do any of the above? Also any other alternative solution is welcome.

Javier Palacios

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]