Hi, This is very interesting that there is a difference between vanilla Cassandra and DSC, I thought it was simply the same thing provided with a setup/package.
-- Jacques-Henri Berthemet From: Eduardo Alonso [mailto:eduardoalo...@stratio.com] Sent: vendredi 6 mai 2016 07:36 To: user@cassandra.apache.org Subject: Re: [C*3.0.3]lucene indexes not deleted and nodetool repair makes DC unavailable Hi Siddaharth: I have tested with apache cassandra 3.0.3 and cassandra-lucene-index-3.0.3.1 and works well but with dsc-cassandra-3.0.3 it does not delete the lucene files. Please, can you open an issue<https://github.com/stratio/cassandra-lucene-index/issues>? Refering to your second question, depending on the consistency level used for writes the repairs in a DC could be very resource-intensive, I have some questions: 1º- How do you run nodetool repair? Do you run it in every machine at the same time or waits for finishing in one machine to run it in the next? Do you use primary range repair (-pr argument) in nodetool repair? 2º- Which is the replication factor for that keyspace?, Which consistency level do you use in writes? Eduardo Alonso [Image removed by sender.] Vía de las dos Castillas, 33, Ática 4, 3ª Planta 28224 Pozuelo de Alarcón, Madrid Tel: +34 91 828 6473 // www.stratio.com<http://www.stratio.com/> // @stratiobd<https://twitter.com/StratioBD> 2016-05-06 9:43 GMT+02:00 Siddharth Verma <verma.siddha...@snapdeal.com<mailto:verma.siddha...@snapdeal.com>>: Hi, I have 2 queries. We are using cassandra dsc 3.0.3 and stratio lucene indexes on tables. 1. when table is truncated, lucene index is not cleared for the same. we see that it still occupied space on disk. 2. when we run nodetool repair, all node are up (nodetool status) but we can't connect to either of the nodes in the same DC. Any help would be appreciated Thanks Siddharth Verma
