cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Benedict (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-6995) Execute local ONE/LOCAL_ONE reads on request thread instead of dispatching to read stage
Date Mon, 14 Apr 2014 21:44:19 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-6995?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13968892#comment-13968892
] 

Benedict commented on CASSANDRA-6995:
-------------------------------------

bq. In a separate note shouldn't we throttle on the number of disk read from the disk instead
of concurrent_writers and reads? its un fair on the threads when one request pulls a lot more
data than others....

Agreed, [~vijay2win@yahoo.com]. I would quite like to see the write and read stages dropped
entirely, and introduce a stage purely for processing (both read and write) requests coming
from other nodes. Direct clients can just execute reads/writes directly in their thread. We
can then have a synchronisation primitive on each disk for taking permission to perform a
rebuffer of data in our readers. Ideally this would dovetail with work in CASSANDRA-5863 to
provide an in-process page cache that could be used not just for compressed files. This way
all direct clients could perform in-memory queries without any synchronisation / blocking
/ context switching.

> Execute local ONE/LOCAL_ONE reads on request thread instead of dispatching to read stage
> ----------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-6995
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-6995
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Jason Brown
>            Assignee: Jason Brown
>            Priority: Minor
>              Labels: performance
>             Fix For: 2.0.7
>
>         Attachments: 6995-v1.diff, syncread-stress.txt
>
>
> When performing a read local to a coordinator node, AbstractReadExecutor will create
a new SP.LocalReadRunnable and drop it into the read stage for asynchronous execution. If
you are using a client that intelligently routes  read requests to a node holding the data
for a given request, and are using CL.ONE/LOCAL_ONE, the enqueuing SP.LocalReadRunnable and
waiting for the context switches (and possible NUMA misses) adds unneccesary latency. We can
reduce that latency and improve throughput by avoiding the queueing and thread context switching
by simply executing the SP.LocalReadRunnable synchronously in the request thread. Testing
on a three node cluster (each with 32 cpus, 132 GB ram) yields ~10% improvement in throughput
and ~20% speedup on avg/95/99 percentiles (99.9% was about 5-10% improvement).



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message