Can a Select Count(*) Affect Writes in Cassandra -
i experienced scenario select count(*) on table every minute (yes, should avoided) caused huge increase in cassandra writes around 150k writes per second.
can explain weird behavior? why select query increase write count in cassandra?
thanks!
if check
org.apache.cassandra.metrics:type=readrepair,name=repairedbackground
and
org.apache.cassandra.metrics:type=readrepair,name=repairedblocking
metrics can see if read repairs sending mutations. perhaps reading data service count(*) causing lot of read repairs if data inconsistent. if thats case lowering read_repair_chance
, dclocal_read_repair_chance
on table (alter table
) reduce load.
other possibilities are:
- you have tracing enabled (either globally or on table) %.
- or if use dse , have slow query's enabled.
Comments
Post a Comment