Description
when running observium with partitioned pollers the generate-rancid script pulls all the devices into the router.db. this causes lockups and timeouts on the pollers.
Also customer X potentially sees hostnames from customer Y so can you please address this?
the script should only generate router.db for devices that are polled by it directly and ignore others.
example from customer yyy - the local poller tried to log in to the switches of customer xxx
causing timeouts and the rancid-run to run for > 2 hours
arudel@srv-yyy:/var/lib/rancid$ sudo tail -f /var/log/rancid/observium.20200609.090001
switch-01-03.lelystad.xxx.nl clogin error: Error: TIMEOUT reached
!
switch-01-04.lelystad.xxx.nl: missed cmd(s): all commands
switch-01-04.lelystad.xxx.nl: End of run not found
switch-01-04.lelystad.xxx.nl clogin error: Error: TIMEOUT reached
!
switch-01-05.lelystad.xxx.nl: missed cmd(s): all commands
switch-01-05.lelystad.xxx.nl: End of run not found
switch-01-05.lelystad.xxx.nl clogin error: Error: TIMEOUT reached
!
^C