cloud controller loses instance, describe instance gives two private addresses
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Eucalyptus |
Fix Released
|
Critical
|
chris grzegorczyk |
Bug Description
I started an instance was it got to a running state with public address 10.249.162.201 and private address 172.30.0.210, capturing what comes down to euca_ipt I can see that the appropriate iptables rules are added, but then removed (last message at the top)
-D PREROUTING -d 10.249.162.201 -j DNAT --to 172.30.0.210
-D POSTROUTING -s 172.30.0.210 -j SNAT --to 10.249.162.201
-D OUTPUT -d 10.249.162.201 -j DNAT --to 172.30.0.210
-A PREROUTING -d 10.249.162.201 -j DNAT --to 172.30.0.210
-A OUTPUT -d 10.249.162.201 -j DNAT --to 172.30.0.210
-A POSTROUTING -s 172.30.0.210 -j SNAT --to 10.249.162.201
While I was looking at that, doing an describe-instances I get this
INSTANCE i-49B00939 emi-F46D117C 172.30.0.210 172.30.0.210 running
Now my instance has lost it's public ipaddress, however SSH is available on the private IP address if you're on the right network. What's interesting is that the private address isn't in the database (euca.addresses) at all, and iptables-save reports that private IP address routes from a different public ip address which also isn't in the DB
-A PREROUTING -d 10.249.162.223 -j DNAT --to-destination 172.30.0.210
-A POSTROUTING -s 172.30.0.210 -j SNAT --to-source 10.249.162.223
-A OUTPUT -d 10.249.162.223 -j DNAT --to-destination 172.30.0.210
I suspect this is closely related to: https:/
Changed in eucalyptus: | |
status: | Confirmed → Fix Committed |
Changed in eucalyptus: | |
status: | Fix Committed → Fix Released |
Please see an update to: https:/ /bugs.launchpad .net/eucalyptus /+bug/403111 for more detail and steps to reproduce