Hello,
I plan to run 389ds in containers (docker image on kubernetes), with a multi-master
replication setup, as described below.
Setup:
- Deployment: I'm using kubernetes statefulsets so the containers get created with the
same name
- Name resolution: A headless service is created, so the containers can talk to each other
using hostnames
- Data: Persistent volumes, auto-created by using a storage class, mounted using
persistent volume claims
- replication:
- replica id: extracted from the hostname
- replica host: I'm looking for containers in the same stateful set and extracting
their names
I have a few questions about replication in this setup. When a container is destroyed, and
replaced with a new one
i. should I disable changelog and re-enable it?
ii. should I delete the replication agreements and recreate them?
iii. should I re-initialize the ds instance in the newly created container?
iv. are there any known conditions that can break replication or corrupt the ds instance
if the new container still reads data from the same volume?
Thanks,
Aravind