Hi, We are facing this issue on production, We upgraded our cassandra from 3.0.3 to 3.5
When we ran a query with partition key and clustering column(entire primary key specified), we get 16 rows in return. We have 2DC's, each with RF 3 for our keyspace. 1. We connected with cqlsh, and setting consistency to local_one, and tracing on, we saw that, we got correct result on 3, and erroneous results on 3. Correct result : only 1 row Erroneous result : 16 rows 2. we executed the statement while specifying only the clustering column with ALLOW FILTERING, then we got the only one record for that partition key. 3. While upgrading, we dropped key_cache folder on some, not all. What could be the causes and how to fix this issue? We speculate that it might be due to cache. Any help would be appreciated. Thanks Siddharth Verma
