What is the best recommendation way to recover node which have a failed drive(Contains data file, commit log file ).Binaries are intact on other drive .
Cluster configuration -
-3 Node cluster
- DSE version 5.8
- Data size ~1.5 TB on each node
- RF is 3 or 2 for application key spaces
-RF is 1 for some default key spaces.
What is the best solution to avoid any data loss ? and what will be the performance impact during recovery?
1) Replace the node with brand new node ?
2) Clean the failed drive and start the Cassandra service on failed drive node and run repair.
3) Replace the drive and start the Cassandra service on failed drive node and run repair.