cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "nheerink@hetnet.nl" <nheer...@hetnet.nl>
Subject varying results count query after alter keyspace system_distributed to NetworkTopologyStrategy
Date Wed, 17 Apr 2019 18:35:17 GMT
Hi, we want to add a new DC to our existing 2 DC cluster Cassandra 3.11.4. In our preparations
we noticed the system_distributed keyspace still on SimpleStrategy and decided to do alter
keyspace to change it into NetworkTopologyStrategy according to documentation online on the
existing 2 DC cluster (running already 1 year). We did not create the existing Cassandra ourselves.
we did:
ALTER KEYSPACE system_distributed WITH replication = {
  'class': 'NetworkTopologyStrategy',
  '<dc_name1>': '<replication_factor>',
  '<dc_name2>': '<replication_factor>'
};
This in itself went fine. Then we did nodetool repair -full on each node. Also went fine.
But now, when we do a count query on the system_distributed keyspace parent_repair_history
or repair_history tables, we get varying results everytime we do this query, querying immediately
after each other. Sometimes count is a bigger number, sometimes a smaller number
query:
select count(parent_id) from system_distributed.parent_repair_history
Should we perform other actions as well?
Mime
View raw message