Thank you,

Regarding the first section of question.

If you can't say to the particular Poller which devices to POLL,  how every poller chooses which devices to POLL ? Do they do it round-robin ?

On Thu, Nov 30, 2017 at 3:22 PM, Tom Laermans <tom.laermans@powersource.cx> wrote:
On 30/11/2017 10:38, Edvinas K wrote:
So, i've some questions:

do it's possible to scale the already running setup ?

1) Is it possible to make already running single server as a "master" where would RRD/MYSQL databases reside and that "master" server also would be as a one of the pollers.
And addition to that add two other pollers.

So in summary I just want to add two additional pollers to existing Observium poller, while all the device ID's and all the history would be kept.
Up to here, that is completely correct - adding extra pollers will keep all your current devices set up, with ids and all...
Our DC is not big ~200 devices. We're considering this option to allow us to see what's going on when there problems in the network occur - when setup with single poller cannot reach the network node which is in other continent - all the data is lost. As I understand in distributed poller setup the poller which would reside in other continent would be able to 'grap' the monitoring data and (when the connection would be available again) it would send the information from other continent 'local' poller to the 'main; server. Do i'm correctly understand distributed poller principle ? Do the data saved in one (distant) poller could be resented when connection between two Pollers is up again ?
... but it writes directly to the same database and there is no "backing up the data until the main controller becomes available again".

Tom

_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium