2 thoughts on “Scaling, Distributing and Partitioning Registrations with OpenSIPS 2.4

  1. This article is well summed up and helps me sort out the differences between the various working mode presets. But I have some doubt. What’s the difference between restart persistency and dataset mirroring? What is the purpose of restart persistency? why restart persistency is needed in Federated Cluster(in this mode, every opensips just need keep data of client that registered to it)?

    Like

    1. Hey, @sxtang!

      The purpose of restart persistency is to allow you to “systemctl restart opensips” without losing dialogs, online subscribers, presence subscriptions, etc.

      The wording “dataset mirroring” refers to the fact that two cluster nodes have (and keep sharing changes to) the same data set. Example such data: live dialogs, online users.

      Correct: in federated clustering, each OpenSIPS keeps its client data and does not waste bandwidth, cpu, etc. to broadcast this data to everyone. But what does this have to do with restart persistency? Or even high-availability? We may want to restart this node and still have the data. Or we may want to allow it to crash & burn while still keeping the service up!

      Like

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s