1

The Pivotal Running Geode Locator Processes documentation states:

"You can run the locator standalone or embedded within another Geode process. Running your locators standalone provides the highest reliability and availability of the locator service as a whole."

Given a cluster with multiple nodes, with a Geode locator and Geode server running on each node...would it make sense to just embed the locator within the Geode server process.
In this scenario, is there any benefit gained by running the locator in a separate process?

wyck
  • 327
  • 1
  • 8

1 Answers1

1

You need fewer locators as compared to servers in the Geode system. You could be running say 3 locators and 10s or 100s of servers.

Swapnil
  • 1,191
  • 7
  • 8
  • Thanks Swapnil. That makes sense given large numbers of Geode servers. Assuming only a small number of Geode locators and Geode servers (let's say there are 2 of each), would there be any reason to not embed the locators with the servers? – wyck Aug 17 '17 at 14:03
  • 1
    Locators play a crucial part in network partition detection. So, if you were to embed locators, I would recommend that you run a odd number of servers with embedded locators. – Swapnil Aug 17 '17 at 16:46
  • Thanks for the tip on keeping an odd number of servers! Your comment led me to this helpful Geode "[How Network Partitioning Management Works](https://gemfire.docs.pivotal.io/geode/managing/network_partitioning/how_network_partitioning_management_works.html)" link. – wyck Aug 18 '17 at 16:17