yes in 0.8 the compaction thresholds are for buckets of files. Files are put into buckets where the size of each file is +/- half the size of the average size. Files smaller than 50MB (from memory) are all put in the first bucket.
You can change the compaction thresholds via nodetool, and trigger a minor compaction by flushing. Again via node tool. Cheers ----------------- Aaron Morton Freelance Cassandra Developer @aaronmorton http://www.thelastpickle.com On 21/09/2011, at 9:34 AM, Konstantin Naryshkin wrote: > I believe that minor compactions work on tables of same or similar size, so > as long as your tables do not fall within a small range of each other in > terms of size, Cassandra does not see an opportunity to run a minor > compaction. > > ----- Original Message ----- > From: "myreasoner" <myreaso...@gmail.com> > To: cassandra-u...@incubator.apache.org > Sent: Tuesday, September 20, 2011 2:30:26 PM > Subject: Column family has more SSTables than threshold but no > minorcompaction is running > > Column Family: CF_1 > SSTable count: 249 > Space used (live): 15120496328 > Space used (total): 15120496328 > Number of Keys (estimate): 18860800 > Memtable Columns Count: 0 > Memtable Data Size: 0 > Memtable Switch Count: 2483 > Read Count: 442 > Read Latency: NaN ms. > Write Count: 208741548 > Write Latency: NaN ms. > Pending Tasks: 0 > Key cache capacity: 200000 > Key cache size: 2395 > Key cache hit rate: NaN > Row cache capacity: 200000 > Row cache size: 221 > Row cache hit rate: NaN > Compacted row minimum size: 104 > Compacted row maximum size: 73457 > Compacted row mean size: 394 > > Current compaction thresholds: min = 4, max = 32 > > As shown above, the # of SStables exceeds the max, the node load is light > but it's not running minor compaction on this Column family. I don't want > to run major compaction manually from time to time. Is there anything I can > tune? > > -- > View this message in context: > http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Column-family-has-more-SSTables-than-threshold-but-no-minorcompaction-is-running-tp6813230p6813230.html > Sent from the cassandra-u...@incubator.apache.org mailing list archive at > Nabble.com.
