[ https://issues.apache.org/jira/browse/CASSANDRA-12219?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-12219: ----------------------------------------- Reviewer: Aleksey Yeschenko Fix Version/s: 3.x 3.0.x Status: Patch Available (was: Open) Indeed. The problem is that the counter mechanism in {{PartitionUpdate}} was forgetting to deal with the static row, and the compact table is internally using a static row. The fix is pretty trivial otherwise. | [12219-3.0|https://github.com/pcmanus/cassandra/commits/12219-3.0] | [utests|http://cassci.datastax.com/job/pcmanus-12219-3.0-testall] | [dtests|http://cassci.datastax.com/job/pcmanus-12219-3.0-dtest] | (Note that I'll include the attached dtest on commit as test coverage) > Cluster + Compact Storage loses counter writes > ---------------------------------------------- > > Key: CASSANDRA-12219 > URL: https://issues.apache.org/jira/browse/CASSANDRA-12219 > Project: Cassandra > Issue Type: Bug > Reporter: Brian Wawok > Assignee: Sylvain Lebresne > Fix For: 3.0.x, 3.x > > Attachments: 0001-add-failing-counter-dtest.patch > > > When we have both multiple nodes, a counter column, and compact storage - > some writes are lost. > Example given in dtest below. > In Cassandra 3.0 does not work. Before 3.0 they seem to have. -- This message was sent by Atlassian JIRA (v6.3.4#6332)