MID Server clustered approach

[Shashank45]

Hi,

Currently, we have an implemented "Inventory Scanning" Form which is a 'discovery-like' application. On that form, we are requiring someone to input the ip address of the targeted hostname and they will select the nearest location which is basically the mid server location.

Right now, the behavior of the discovery if we have multiple midservers under one location is that it is selecting the available mid server randomly.

Is there a way for us to configure it in such a way that the mid-server will do a clustering model approach rather than selecting randomly?

[Niklas Peterson]

Hi,

Yes, you can set up MID Server clusters to gain functionality like failover or load balancing. It's well documented on docs: https://docs.servicenow.com/bundle/washingtondc-servicenow-platform/page/product/mid-server/task/t_C...

Regards,
Niklas

[Adam Rasmussen]

I think the goal is to allow an end-user to trigger a quick discovery experience but using a cluster instead of a specified or random mid-server.  Assuming you've already built your clusters and have those clusters associated to a discovery schedule.... I suggest looking at the startFromIP API call for Discovery.  startFromIP calls an individual device discovery but uses the Discovery Schedule to determine which mid server cluster (or individual mid server) SHOULD be doing the discovery. 

[Appli]

Hi, just to better understand a concern - what is wrong with an approach where MID server is taken randomly?