We are planning out migration to SmartStore within AWS
Currently, we are running on a multi-site cluster on EC2 instances.
Site 1 is in us-east-1(VA) and site 2 is in us-east-2(OH).
the recommendation is to put the S3 storage in the same AWS region as the indexers. With this being a multi-site cluster that spans 2 regions. What is the best way to configure it?
You can Pick a single region for the s3 bucket, both sites point to it. whichever has more data.
although s3 supports cross region replication would mean that we can try out different nearest s3 site per cluster site,indexes.conf of the bundle that is pushed by master to the slaves would still need to use the same endpoint. editing the slave's indexes.conf manually to have different site is not recommended
Make sure the endpoint includes the region specifier otherwise ec2 instances will use their local regions
@rbal_splunk, thanks for this post, I had a scenario based question.
If the goal/intent is to proof out disaster recovery. In the very unlikely scenario an entire AWS region were to belly up, what would your approach be to ensure failover?
Let's say for a single site cluster that uses S3 replication from Region A to Region B.
What would be a recommended process to failover to the replicated S3 bucket in Region B?