Cassandra

From Mosuma
Jump to: navigation, search

Partioning

  1. Time Series with Cassandra (datastax)
  2. Metric storage and Collection with Cassandra (datastax)
  3. Case against byte-ordered partitioning
  4. Ebay
  5. cql3 and time-series data

Tuning Worker processes to access Cassandra

  1. Start the first worker, run it for 3-5 minutes, Cassandra will optimize the throughput, and when the throughput has stabilized
  2. Start the next worker, wait for a while to measure the consistent increased throughput
  3. Repeat and stop until there is no increase in throughput


If some queues are not producing ACK, it may be time to restart machine and/or rabbitmq


Installing Opscenter

Opscenter

  1. Generate the agent package *after* you setup opscenter

(IMPORTANT)

cd /opt/opscenter
bin/setup.py

Debian

cd /usr/share/opscenter
bin/setup.py
  1. copy the agent.tar.gz to each cassandra node
  2. run opscenter
bin/opscenter -f

Opscenter Agent

  1. install agent to /opt/agent
apt-get install sysstat
cd /opt/agent/
bin/setup hostname_to_opscenter
bin/datastax-agent -f


Data Migration

exporting and recreating SCHEMA using cqlsh

References

Storing Time series data with Cassandra