You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cassandra.apache.org by Jean Tremblay <je...@zen-innovations.com> on 2015/06/19 14:25:38 UTC

Long nodetool repair use nodetool repair incremental

Hi,

I understand that we must repair the DB on a regular basis.
Now I also see that making a repair is using lots of resources in the cluster so I need to do this during the weekend because I really would like to have high performance at least during the week days.

In the documentation I see the "incremental nodetool repair”. That really seems to be the option for me.

Few questions:

1) Why is it written in the documentation that it is suggested to do incremental repair daily, and full repair on the weekends? Is repair incremental not good enough or not safe?
2) In order to use incremental repair we need to do migration steps. One of these steps is to run the tool sstablerepairedset. What is the parameter <sstables>? Is it the name (without extension) of all *.db files contained in the data/cfdir?
3) Is the sstablerepairedset really needed when making a repair incremental on an empty DB?

Thanks for your help

Jean