EC2 stop and start causes Chef nodes to have wrong data


#1

Hi –

I’ve noticed that halting and restarting an EC2 instance causes it to
get new ip addresses and names.
The Chef server doesn’t know about this, and so it retains the old ec2
node information.
This really screws things up if you are using search to get dns names to
connect things together.

Re-running chef-client doesn’t seem to help, and also I can’t do it like
this

knife ssh ‘chef_environment:staging AND role:mongo_rs_member’ ‘sudo
chef-client’

… because of the very problem I’m trying to correct.

So … aside from just adjusting my expectations, is there any way to
update the nodes with the new correct ec2 information?

Liam


Liam Kirsher
PGP: http://liam.numenet.com/pgp/


#2

re running chef client should fix this. this common. you should invoke chef
client soon after reboot. either manually or via rc.d scripts

On Thu, Jul 18, 2013 at 7:02 PM, Liam Kirsher liamk@numenet.com wrote:

Hi –

I’ve noticed that halting and restarting an EC2 instance causes it to
get new ip addresses and names.
The Chef server doesn’t know about this, and so it retains the old ec2
node information.
This really screws things up if you are using search to get dns names to
connect things together.

Re-running chef-client doesn’t seem to help, and also I can’t do it like
this

knife ssh ‘chef_environment:staging AND role:mongo_rs_member’ ‘sudo
chef-client’

… because of the very problem I’m trying to correct.

So … aside from just adjusting my expectations, is there any way to
update the nodes with the new correct ec2 information?

Liam


Liam Kirsher
PGP: http://liam.numenet.com/pgp/