Bug #8709
closedCannot update policies on master, server refuse simultaneous connection
Description
Node cannot update it's policies, because server refuses simultaneous connection:
An output of rudder server debug is available here: https://gist.github.com/VinceMacBuche/b16b0123c14c65f87415aca070fbe442
But the error extract (last line of logs) are here:
rudder verbose: 192.168.41.4> Host relay.rudder.local granted access to /usr/share/ncf/tree/60_services/baseline/stable/baseline.cf rudder verbose: Obtained IP address of '192.168.41.4' on socket 8 from accept error: Remote host '192.168.41.4' is not in allowallconnects, denying second simultaneous connection rudder verbose: Obtained IP address of '192.168.41.4' on socket 8 from accept error: Remote host '192.168.41.4' is not in allowallconnects, denying second simultaneous connection rudder verbose: Obtained IP address of '192.168.41.4' on socket 8 from accept error: Remote host '192.168.41.4' is not in allowallconnects, denying second simultaneous connection rudder info: 192.168.41.4> Closing connection, terminating thread
Defining allowallconnects instead of allowconnects in cf-served.cf fix the issue!
Updated by Alexis Mousset over 8 years ago
Did it happen only once? It can be because of a running file copy (started by cf-execd) at the very moment you tried to update.
Updated by Vincent MEMBRÉ over 8 years ago
I had this state for like ... 10 minutes,
Updated by Alexis Mousset over 8 years ago
- Category changed from System techniques to Agent
- Assignee changed from Benoît PECCATTE to Alexis Mousset
Maybe a connection that was not properly closed, seem possible according to the log (but without timestamp, it is not easy to know). I will try to reproduce.
Updated by Nicolas CHARLES about 8 years ago
- Status changed from New to Rejected
i couldn't reproduce, nor on Debian, nor on Centos, with server 4.0 and agent 3.2 and 4.0
Rejecting