0

we have a master-master (two-way sync) configuration with our cloud and on-site DB. we're using "Server version: 5.7.28-31-57 Percona XtraDB Cluster (GPL)"

Our problem is when the on-site DB dies it causes the cloud DB to die as well. We think it dies because the on-site DB is ahead of the cloud DB so it dies and tries to re-sync.

I found "symmetricds" when looking into how to better handle the data sync between the 2 DB and I'm wondering if this will help with our current problem.

Since our main goal atm is to stop the on-site DB from also killing the cloud DB.

Improve this question
2
  • What are you using for data sync, replication or the actual cluster? If a cluster you should have at least 3 nodes.
    – Jesus Uzcanga
    Commented Apr 6 at 16:40
  • @JesusUzcanga I think we're using a cluster, what do you mean by 3 nodes?
    – magicianiam
    Commented Apr 7 at 7:02

1 Answer 1

Reset to default
0

Percona XtraDB Cluster (PXC) is a "cluster". To work properly, it needs 3 nodes, even if one of them is a "garb". With only 2 nodes each node cannot tell the difference between a node failure and a network failure, so it will stop taking writes.

You may be able to "weight" one of your two nodes as more important. This would say that if the other node dies, the cluster survives. But if the node with more weight dies, the lesser node will stop receiving writes.

Under the covers is a product called Galera; probably anything you can read about Galera applies to PXC.

Improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.