Stop repairs in progress

[Brandon]

In our stage environment, we have a 9 node Cassandra cluster with repairs in progress. We are attempting to replace the cron repair strategy we previously had baked in our AMI with Cassandra Reaper. Since the repairs are still in progress, Reaper returns that it has paused after it repairs a certain number of tables. What is the best course of action to stop repairs on Cassandra nodes so I can run Reaper?

Thanks!

[Reid Pinchback]

Disable the cron jobs if you haven't already.  Do a rolling restart of your C* cluster(s) as that seems to tidy up any repair artifacts.  Then start a repair via reaper.  Then monitor the results as you'll likely have to find tunings to hit the sweet spot on repair throughput that suits your topology and column family count vs storage footprint.