That's only a small part of the problem. A larger issue is that what
follows the index read is basically a multiget within that node, which
generates a lot of non-sequential I/O. The second issue is that nodes are
queried sequentially to get more result rows until the requested row count
has been
Hi All,
Correct me if I am wrong, but I know that secondary indexes are stored
to local column families on each node. Previously where the default
cache key value was 200,000 rows, and you couldn't really tune the
local index column family, that posed a limitation on low cardinality
of the possibl
