read_repair_chance: Probability (0.0-1.0) with which to perform read repairs on CL.ONE reads
Default is 1.0 your CL.ONE reads will trigger read repairs for all reads. You can update your column families definition to reduce that chance. For the CL.ALL writes you can consider turning it off by setting it to 0.0.
Running nodetool repair will reduce the chance of inconsistent data, it does not mean that read repair will not get triggered.
You are reading with ONE and the
On Sat, Jul 30, 2011 at 5:04 AM, Philippe <firstname.lastname@example.org>
I have a 3-node ring at RF=3 that is doing reads & writes. I am using two types of consistency levels.
- write ALL,read ONE for one set of column families
- write QUORUM, read ONE for another set of families
Every day, I have a cron job that runs a nodetool repair on each node. The nodes are not going down and there are no exceptions in the logs.
And yet, nodetool netstats tells me that the ReadRepairStage is increasing during the day, for example between the time I started this email and the time I finished it.
I don't understand how this is possible. Any explanation ?