Migrating to Incremental Repair

[Kathiresan S <ka...@gmail.com>]

Hi,

We have a 12 node cassandra 3.0.4 cluster (not DSE) and recently upgraded
from 2.0.5 to 3.0.4. We have a scheduled nightly repair on each node with
*sequence* *full* options and would like to start running the incremental
repairs.

As per below documentation, it looks like we need to follow the migration
steps in order to migrate to incremental repair (because we have tables
with LCS)

http://docs.datastax.com/en/cassandra/3.0/cassandra/operations/opsRepairNodesMigration.html

I've few questions around this:

   - Do we need *the migration steps* in first place to switch to
   incremental repair or can we just switch incremental repair?, as we have
   run full repair so far all these days
   - If we need to do, as per migration steps, can some one please confirm
   if below are the steps (one node at a time)?


   1. nodetool disableautocompaction
      2. nodetool repair --sequential --full
      3. stop cassandra
      4. sstablerepairedset *--is-repaired* -f *list_of_sstable_names.txt*
      (Also, the doc says, this is the list of sstables before autocompaction,
      thats bit confusing, could some one please explain if there is any step
      before disabling auto-compaction?)
      5. start cassandra
   - And, whats the proper way to get all the sstable names for the above
   steps?

Thanks,
​​Kathir