Rudder

Hum. Not good.

When you say "rename", you mean you changed the hostname (just to be sure it was only a hostname change) ?

I'm sure you are aware of that, but for other people finding that ticket, a hostname change does not trigger a policy regeneration, and one must be called manually, see: http://www.rudder-project.org/redmine/issues/1411

We will try to reproduce it, given what you are saying, a simple "grep" on the old hostname after the rename should give us all the places which weren't correctly updated.

Hi,

yes, just changed the hostname.

that's open for 5 years? you need to put that somewhere more prominent.
anyway. I've just had another case of this, and at least clear caches and the following rebuild did not have effect.

as for testing:
I fix the file using sed, so it should be really easy to reproduce.
(fix via sed worked + HUP cf-served worked)

Hi Florian,

1. accept a new node
2. apply policy on it
3. checked it could get policies, and apply it
4. update hostname (hostname <HOST>, edit /etc/sysconfig/network and /etc/hosts
5. ran rudder agent inventory on the node
6. ran rudder agent run on the server
7. checked that the node name was changed
8. correct /etc/hosts on server
9. click on "status/Update"
10. policies are regenerated
11. hostname in /var/rudder/cfengine-community/inputs/common/1.0/cf-served.cf is correct
12. ran rudder agent update on the node - it worked
13. changed policies, it worked