thanks Martin. 99 percentile of all tables are even size. Max is always higher in all tables.
The question is, How do I identify, which table is throwing this "Maximum memory usage reached (512.000MiB)" usage message? On Mon, Jun 11, 2018 at 5:59 AM, Martin Mačura <m.mac...@gmail.com> wrote: > Hi, > we've had this issue with large partitions (100 MB and more). Use > nodetool tablehistograms to find partition sizes for each table. > > If you have enough heap space to spare, try increasing this parameter: > file_cache_size_in_mb: 512 > > There's also the following parameter, but I did not test the impact yet: > buffer_pool_use_heap_if_exhausted: true > > > Regards, > > Martin > > > On Tue, Jun 5, 2018 at 3:54 PM, learner dba > <cassandra...@yahoo.com.invalid> wrote: > > Hi, > > > > We see this message often, cluster has multiple keyspaces and column > > families; > > How do I know which CF is causing this? > > Or it could be something else? > > Do we need to worry about this message? > > > > INFO [CounterMutationStage-1] 2018-06-05 13:36:35,983 > NoSpamLogger.java:91 > > - Maximum memory usage reached (512.000MiB), cannot allocate chunk of > > 1.000MiB > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org > For additional commands, e-mail: user-h...@cassandra.apache.org > >
