Anti-Entropy Repair, also called Anti-Entropy Service or AES, is a process where nodes compare their data and ensure data is replicated properly and up-to-date. This recipe explains how to do an anti-entropy repair and the conditions for which it should be ran.
Use nodetool
repair
against a node to be repaired.
$<cassandra_home>/bin/nodetool -h 127.0.0.1 -p 8001 repair
Anti-entropy repair is intensive for disk, CPU, and network resources. It is optimal to run this at times of low traffic. It can create excess copies of data on your nodes. If the storage on nodes grows significantly as a result of AES, use nodetool
compact
. Major compaction should remove duplicate data that resulted from the repair.