2

Our QA Team has requested that we completely clear all data within the app's keyspace on our Cassandra 2.1.3 server prior to testing. (Cassandra 2.1.3 is running on an Ubuntu 14.04LTS Azure D12 instance [4 cores, 28GB Memory]).

We have attempted to TRUNCATE the column families and had problems with both Cassandra and Stargate index corruption afterwards. (returning incorrect/no data).

We have attempted to DELETE the data from the column families and had the same problem with indexes and tombstoning.

We were told to use DROP KEYSPACE with snapshot turned off; this resulted in Cassandra shutting down with all remote connections forcibly shut down, a partially deleted state on several occasions where we were able to access the keyspace via DevCenter, but it did not appear in the schema_keyspaces table, and/or corrupted indexes.

There are less than 100,000 records across 30 column families, so not a whole lot of data.

We cannot upgrade Cassandra to the latest version because Stargate only supports the C* 2.1.3 version.

Any other recommendations of how we can resolve this problem?

Chris Trout
  • 31
  • 2

1 Answers1

1

We answered the question internally.

Remove StarGate. Once we removed StarGate, the TRUNCATE and DROP KEYSPACE functionalities began to work appropriately again.

We notified StarGate support.

Chris Trout
  • 31
  • 2