Description of problem: collectd uses the udp protocol, the dns service records are _collectd._tcp though. Version-Release number of selected component (if applicable): 0.94, both server appliance (announce) and node (search).
Not fully sure I understand the issue (if it is still an issue). Which components running on which machines are failing to communicate due to the transport protocol mismatch? Is this an issue w/ collectd itself or how we are configuring it in ovirt? What relies on this that's being affected? If you could elaborate, I'd appreciate it.
It's just a misnomer, collectd uses UDP. It doesn't break anything, but should be fixed regardless, for correctness sake: just change default dnsmasq config and lookups in the ovirt-node scripts.
Sorry this took so long, was thrown off by the project the fix should be in. Shouldn't the change be in the server project? As the server installer sets up /etc/dnsmasq.d/ovirt-dns.conf which contains the incorrect collectd entry. In any case, I just send this out which I believe resolves the issue: https://www.redhat.com/archives/ovirt-devel/2009-May/msg00222.html
Changes to the node to use the new dns entry, based on alan's feedback https://www.redhat.com/archives/ovirt-devel/2009-May/msg00238.html
Committed and resolved