[ https://issues.apache.org/jira/browse/CASSANDRA-7040?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13969736#comment-13969736 ]
Benedict commented on CASSANDRA-7040: ------------------------------------- bq. You could add in helpers like mincore (and row cache) to help inform you Or CASSANDRA-5863 :-) As to batching - that's another step further along: it would be interesting to experiment with an intelligent "storage manager" that requests are submitted to, and are coordinated by, but I think that comes after 5863 + this. There's lots of ways we might be able to get improved performance with that approach, but I'm not absolutely sure they'll pan out, and they'll be a non-trivial undertaking. > Replace read/write stage with per-disk access coordination > ---------------------------------------------------------- > > Key: CASSANDRA-7040 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7040 > Project: Cassandra > Issue Type: Improvement > Components: Core > Reporter: Benedict > Labels: performance > Fix For: 3.0 > > > As discussed in CASSANDRA-6995, current coordination of access to disk is > suboptimal: instead of ensuring disk accesses alone are coordinated, we > instead coordinate at the level of operations that may touch the disks, > ensuring only so many are proceeding at once. As such, tuning is difficult, > and we incur unnecessary delays for operations that would not touch the > disk(s). > Ideally we would instead simply use a shared coordination primitive to gate > access to the disk when we perform a rebuffer. This work would dovetail very > nicely with any work in CASSANDRA-5863, as we could prevent any blocking or > context switching for data that we know to be cached. It also, as far as I > can tell, obviates the need for CASSANDRA-5239. -- This message was sent by Atlassian JIRA (v6.2#6252)