[ https://issues.apache.org/jira/browse/CASSANDRA-9960?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14661624#comment-14661624 ]
Jaroslav Kamenik commented on CASSANDRA-9960: --------------------------------------------- I have changed cluster name and paths to data dirs only. The issue doesnt appear imediatelly, but after some time, few release cycles, sometimes hours, sometimes days.. I have around 30 types and 150 tables and push few GBs of data during release, with hundreds of threads parallely. > UDTs still visible after drop/recreate keyspace > ----------------------------------------------- > > Key: CASSANDRA-9960 > URL: https://issues.apache.org/jira/browse/CASSANDRA-9960 > Project: Cassandra > Issue Type: Bug > Components: Core > Reporter: Jaroslav Kamenik > Assignee: Robert Stupp > Priority: Critical > Fix For: 2.2.x > > > When deploying my app from the scratch I run sequence - drop keyspaces, > create keyspaces, create UDTs, create tables, generate lots of data... After > few cycles, randomly, cassandra ends in state, where I cannot see anything in > table system.schema_usertypes, when I select all rows, but queries with > specified keyspace_name and type_name return old values. Usually it helps to > restart C* and old data disapear, sometimes it needs to delete all C* data. -- This message was sent by Atlassian JIRA (v6.3.4#6332)