i have similar case and set the first relayseever inmy pool as the one on which changes ar eonly allowed. then i do scp + service restart to the others with bash oneliner on demand.
for the future i plan to check in config from any host to central svn/git repo and check frequently for changes from all nodes. svn/git repo can also be a public one for redundancy reasons (github...) if content is not sensitive. Am Mittwoch, 7. November 2018 schrieb K F : > Hi all > I'm contemplating on how I best keep all our relay servers synchronized in their config. > They are set up as round robin servers in the DNS, so they distribute the load pretty ok. > My first idea was to set up some rsync to copy the relevant directories like /etc/postfix and /etc/opendkim. > Then I need a local script on each server looking for changes, and if there is, it should restart the postfix service. > I could also just mount the folders to a specific server, but then I introduce a SPOF, which is not what I would like, and I still need something on each server to restart the postfix service if any changes are detected. > Are there any nice ready made systems for this? > I don't have any mailboxes on the servers, as they are only relays from our internal systems, so there is no need for a big system that tries to synchronize those as well. > Best regards > Kenneth >
