Current Clustering Use Cases
Some of the use cases below have been mined from this email thread on fedora-tech
High Availability / Fail-over
- Alexander:
- Two node cluster, secondary node for fail-over
- Full replication
- Possible load balancing
- Sufia
- Deploying on AWS (Ubuntu)
- Installation scripts below, #1
- DRI-P:
- Two node cluster, secondary node for fail-over
- With this configuration, experiencing a magnitude worse write performance roadblock
- Configuration details
- Benjamin P:
- Investigating cluster for replication (failover?)
Bulk Ingest
- Art Institute of Chicago:
- Possible future interest in investigating if clustering could help with:
- write performance
- read of large binary files
- Possible future interest in investigating if clustering could help with:
Resources
- Installations scripts targeting AWS deployment
- Ansible scripts for two-node cluster
Noted Issues
- Write performance of cluster
- Out of date installation documentation
- Clustering works on single machine, but not when deployed on separate machines (Note: this is not a generally experienced issue)
Opportunities to Explore
- https://github.com/infinispan/infinispan/wiki/RAC:-Reliable-Asynchronous-Clustering
- https://github.com/infinispan/infinispan/wiki/Handling-cluster-partitions
- https://github.com/infinispan/infinispan/wiki/Clustered-listeners
- https://github.com/infinispan/infinispan/wiki/Design-For-Cross-Site-Replication
- https://github.com/infinispan/infinispan/wiki/Consistency-guarantees-in-Infinispan