Instance DNS lookups do not work across subnets

[markdryan]

This can be reproduced by forcing a subnet per instance by setting the cidr-prefix-size to 30. When then create two new instances with different names. The two instances belong to different subnets. We restart the instances (to restart their networking to work around a cloud-init bug which initializes networking too late) and then ssh into one of the instances. The instance is able to resolve its own name via its DNS server, but not that of its sibling.

$ CIAO_CLIENT_CERT_FILE=/etc/pki/ciao/auth-admin.pem ciao-cli tenant update -for-tenant=97f5a4bc-bce5-41b5-a71e-bbb084383cbc -cidr-prefix-size=30
 $ ciao-cli instance add --workload 2a94c223-d8ce-4c35-9daa-3a4c3ef9e79d -name electron
Created new (pending) instance: d18d9131-b778-4715-b676-f8998adacb34
 $ ciao-cli instance add --workload 2a94c223-d8ce-4c35-9daa-3a4c3ef9e79d -name beebon
Created new (pending) instance: cfe9cc58-07d1-4971-87c5-68ad9319ef51
 $ ciao-cli instance stop --instance d18d9131-b778-4715-b676-f8998adacb34
 $ ciao-cli instance stop --instance cfe9cc58-07d1-4971-87c5-68ad9319ef51
 $ ciao-cli instance restart --instance d18d9131-b778-4715-b676-f8998adacb34
 $ ciao-cli instance restart --instance cfe9cc58-07d1-4971-87c5-68ad9319ef51
 $ ciao-cli instance list
# UUID                                 Status Private IP SSH IP        SSH PORT
1 d18d9131-b778-4715-b676-f8998adacb34 active 172.16.0.2 198.51.100.84 33002
2 cfe9cc58-07d1-4971-87c5-68ad9319ef51 active 172.16.0.6 198.51.100.54 33006
 $ ssh -i ~/local/testkey demouser@198.51.100.84 -p 33002
 $ ping electron
PING electron (172.16.0.2) 56(84) bytes of data.
64 bytes from electron (172.16.0.2): icmp_seq=1 ttl=64 time=0.658 ms
64 bytes from electron (172.16.0.2): icmp_seq=2 ttl=64 time=0.034 ms
^C
--- electron ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 0.034/0.346/0.658/0.312 ms
demouser@electron:~$ ping beebon
ping: unknown host beebon