> Disabling row cache in this case makes sense, but disabling key cache > is probably hurting your performance quite a bit. If you wrote 20GB > of data per node, with narrow rows as you describe, and had default > memtable settings, you now have a huge number of sstables on disk. > You did not indicate you use nodetool compact to trigger a major > compaction, so I'm assuming you did not.
Well, unless this is a matter of background compaction happening more slowly than writes (i.e., that after insertion there is a built-up need for pending compactions that have not yet caught up by the time the read test is done), he should not have a huge number of sstables. If the memtables are flushing due to write count rather than memory, and the resulting memtables vary greatly in size, the result may be more sstables than ideal - but still not a huge number of them. Or am I missing something? -- / Peter Schuller
