Hello. This is my first post here. I've setup hbase with 2 column families 2GB heap size with memstore upper and lower set to 0.5 . I have a couple of wide rows with 200K to 500K columns.
Upon querying the second time for the row, the memstoreSizeMB is still 0. The block cache hit ratio is around 70% .. what doesn't make sense is why this row is not entirely cached in memory. Basically it is taking around 200ms to get via the scanner. Is this time in the expected range? Also, why is the memstore not caching this query result? Thanks so much for any help in tracking this down. If there is additional information that would be helpful, please do tell and I'll post back ASAP. Thanks! Cheers Ash -- View this message in context: http://old.nabble.com/memstoreSizeMB%3D0-tp34007166p34007166.html Sent from the HBase User mailing list archive at Nabble.com.