This project has two offerings. They are:
Note: This repository does not need to be cloned. Everything is already included in the pip commands.
Make sure python-setuptools
and python-pip
are installed, then run:
pip install cassandralauncher
pip install --upgrade cassandralauncher
Start the program once to copy the default /usr/local/etc/cassandralauncher/clusterlauncher.conf
to ~/.clusterlauncher.conf
.
Exit the program and edit ~/.clusterlauncher.conf
with your appropriate authentication parameters.
cassandralauncher
imagelauncher
Either run cassandralauncher
or imagelauncher
again.
With cassandralauncher
:
With imagelauncher
:
THIS MUST BE DONE! IF NOT YOUR AWS ACCOUNT WILL GET A HUGE BILL. TAKE IT FROM ME!
Disclaimer: Even though these tools try their best to keep track of launched clusters, it is always best to ensure that all clusters are terminated periodically. This is especially true in cases where AWS/RAX was unable to tag the machine as they were launched. If the tools were unable to tag said machines, they will forever ignore them since we would rather not even present the option to kill a cluster these tools did not launch.
Run:
cassandralauncher -h
to display all options. imagelauncher
does not have this functionality, but is easily scriptable at the Python level calling ec2.py or rax.py. See clusterlauncher.py
for how to do this.
My cluster is not done launching one (or several) of my nodes. What did I do wrong?
Nothing. EC2 and Rackspace do this from time to time. You can either continue on to do
basic testing, or terminate this cluster and try again. Using EC2 and Rackspace off it's
peak hours helps in this scenario, in general.