Sudipta Laha created CASSANALYTICS-99:
-----------------------------------------
Summary: Bulk reader node availability comparator arranges
replicas in incorrect order
Key: CASSANALYTICS-99
URL: https://issues.apache.org/jira/browse/CASSANALYTICS-99
Project: Apache Cassandra Analytics
Issue Type: Bug
Reporter: Sudipta Laha
There’s a logic in the bulk reader that organizes replicas for a token range
based on their availability. The sorting order is determined by the status of
the node. This arrangement ensures that the bulk reader reads SS tables from
available replicas first. If the requested consistency level is achieved by
simply reading SS tables from the available replicas, the bulk reader skips
reading from the unavailable replicas. However, due to a bug in the code, the
bulk reader operates contrary to its intended purpose. Consequently, it
attempts to read SS tables from the unavailable replicas first before
successfully reading from the available replicas.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
