In our production system we have a secondary index we would like to drop. Our application has been changed to no longer use this index, and generally all read activity stopped on this several weeks ago. However, in cfstats we do see a tiny amount of reads still being registered (Local read count), and would like to account for these before going ahead and dropping the index on the live system – just as a precaution really.
Does anyone know if there are any internal Cassandra housekeeping jobs that could be causing this? Interestingly there will be no reads for hours, and then a tiny spike of reads (< 10) across the cluster at once, followed by continued inactivity.
Any suggestions would be much appreciated