[ https://issues.apache.org/jira/browse/CASSANDRA-11154?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15224916#comment-15224916 ]
Aleksey Yeschenko commented on CASSANDRA-11154: ----------------------------------------------- Thanks for the stack trace. You almost guessed the issue, but not quite. It's just that old schema tables (system.schema_*) are unloaded from memory by then. The easiest fix would be to check if system.schema_* tables are in {{Schema}} and not attempt if they aren't. None of it is a problem for Cassandra-proper, it really works as it's designed to. C* is not *really* built for being embeddable, so there have to be many issues like that in embedded mode. That said, I can put in an easy fix for just this one problem (but you'll have to test the patch). > CassandraDaemon in Managed mode fails to be restartable > ------------------------------------------------------- > > Key: CASSANDRA-11154 > URL: https://issues.apache.org/jira/browse/CASSANDRA-11154 > Project: Cassandra > Issue Type: Bug > Reporter: Achim Nierbeck > Fix For: 3.x > > Attachments: CASSANDRA-11154_patch.txt > > > Restarting the CassandraDeamon in managed mode fails to restart due to > duplicate migration of already migrated keyspaces. > To reproduce this, just do something like in this test class: > https://github.com/ANierbeck/Karaf-Cassandra/blob/master/Karaf-Cassandra-Embedded/src/test/java/de/nierbeck/cassandra/embedded/TestEmbedded.java -- This message was sent by Atlassian JIRA (v6.3.4#6332)