Cassandra only reads a small part of each SSTable during normal operation (not compaction), in fact Datastax recommends lowering readahead - http://www.datastax.com/documentation/cassandra/2.1/cassandra/install/installRecommendSettings.html
There are also blogposts where people have improved their read latency reducing ra. Thanks, Daniel On Wed, Sep 24, 2014 at 4:15 PM, DuyHai Doan <doanduy...@gmail.com> wrote: > "does it typically have to read in the entire SStable into memory > (assuming the bloom filter said yes)?" --> No, it would be perf killer. > > On the read path, after Bloom filter, Cassandra is using the "Partition > Key Cache" to see if the partition it is looking for is present there. > > If yes, it gets the offset (from the beginning of the SSTable) to skip a > lot of data and move the disk head directly there > If not, it then relies on the "Partition sample" to move the disk head to > the nearest location of the sought partition > > If compaction is on (by default), there will be another step before > hitting disk: compression offset. It's a translation table to match > uncompressed file offset / compressed file offset > > > On Wed, Sep 24, 2014 at 10:07 PM, Donald Smith < > donald.sm...@audiencescience.com> wrote: > >> We’re using cassandra as a key-value store; our values are small. So >> we’re thinking we don’t need much disk readahead (e.g., “blockdev –getra >> /dev/sda”). We’re using SSDs. >> >> >> >> When cassandra does disk seeks to satisfy read requests does it typically >> have to read in the entire SStable into memory (assuming the bloom filter >> said yes)? If cassandra needs to read in lots of blocks anyway or if it >> needs to read the entire file during compaction then I'd expect we might as >> well have a big readahead. Perhaps there’s a tradeoff between read >> latency and compaction time. >> >> >> >> Any feedback welcome. >> >> >> Thanks >> >> >> >> *Donald A. Smith* | Senior Software Engineer >> P: 425.201.3900 x 3866 >> C: (206) 819-5965 >> F: (646) 443-2333 >> dona...@audiencescience.com >> >> >> [image: AudienceScience] >> >> >> > >
