[ https://issues.apache.org/jira/browse/CLOUDSTACK-9357?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15584810#comment-15584810 ]
Javier Ayllon commented on CLOUDSTACK-9357: ------------------------------------------- Hi, I've found that even with useextdns=true the local IP is sent as a dns server. It seems that the DNS option 6 configuration is broken. > DHCP DNS option is incorrect for Redundant Router config > -------------------------------------------------------- > > Key: CLOUDSTACK-9357 > URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9357 > Project: CloudStack > Issue Type: Bug > Security Level: Public(Anyone can view this level - this is the > default.) > Components: SystemVM > Affects Versions: 4.8.0 > Reporter: Aaron Brady > Priority: Minor > > With two redundant system routers, my guests are given DNS option 6 > containing the *system* IP (not the virtual IP) of whichever router is master > as their first DNS server entry. > This means that if one router is down or stopped, DNS requests are slowed > until it moves on to the external secondaries I've supplied. > It looks like the `cloud-early-config` script does the right thing with > dnsmasq.conf, but then the cloud.conf put in /etc/dnsmasq.d/ is incorrect. > I've had a look through the code and that appear to be written by > `systemvm/patches/debian/config/opt/cloud/bin/cs/CsDhcp.py`, but I've been > unable to find where it's being passed the incorrect, non-redundant, IP > information. -- This message was sent by Atlassian JIRA (v6.3.4#6332)