I would suggest enabling tracing in cqlsh and see what it has to say. There are many things which could cause this, but I'm thinking in particular you may have a lot of tombstones which get lifted when you read the whole row, and are missed when you read just one column.
On Fri, Dec 26, 2014 at 6:05 AM, Jack Krupansky <jack.krupan...@gmail.com> wrote: > > What do your CQL queries look like? > > -- Jack Krupansky > > On Fri, Dec 26, 2014 at 8:00 AM, <yhq...@sina.com> wrote: > >> Hi, all: >> >> In my cf, each row has two column, one column is the timestamp(64bit), >> another column is data which may be 500k about. >> >> >> I read row, the qps is about 30. >> >> I read that data column, the qps is about 500. >> >> >> Why read performance is so slow where add a so small column in read?? >> >> >> Thanks. >> >> >> >