[
https://issues.apache.org/jira/browse/CASSANDRA-14983?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16773237#comment-16773237
]
Marcus Olsson commented on CASSANDRA-14983:
-------------------------------------------
I created CASSANDRA-15022 for the SEP executor and have uploaded some patches
for 3.0, 3.11 and 4.0:
|[3.0|https://github.com/emolsson/cassandra/tree/3.0-CASSANDRA-14983]|[unit
tests|https://circleci.com/gh/emolsson/cassandra/13]|
|[3.11|https://github.com/emolsson/cassandra/tree/3.11-CASSANDRA-14983]|[unit
tests|https://circleci.com/gh/emolsson/cassandra/16]|
|[4.0|https://github.com/emolsson/cassandra/tree/4.0-CASSANDRA-14983]|[unit
tests|https://circleci.com/gh/emolsson/cassandra/19]|
I seem to be having some issues running the tests on 4.0 though. When I tried
to run the tests on a clean trunk branch locally I got similar issues as the
ones in the build. But the 3.0 and 3.11 branches seems to pass the unit tests.
I also looked at the distributed tests but I'm not sure how to delay or lose
the local requests in a good way without changing the actual code base.
> Local reads potentially blocking remote reads
> ---------------------------------------------
>
> Key: CASSANDRA-14983
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14983
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Coordination
> Reporter: Marcus Olsson
> Priority: Minor
> Attachments: graph_local_read.html, graph_local_read_trunk.html,
> local_read_trace.log
>
>
> Since CASSANDRA-4718 there is a fast path allowing local requests to continue
> to [work in the same
> thread|https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/service/reads/AbstractReadExecutor.java#L157]
> rather than being sent over to the read stage.
> Based on the comment
> {code:java}
> // We delay the local (potentially blocking) read till the end to avoid
> stalling remote requests.
> {code}
> it seems like this should be performed last in the chain to avoid blocking
> remote requests but that does not seem to be the case when the local request
> is a data request. The digest request(s) are sent after the data requests are
> sent (and now the transient replica requests as well). When the fast path is
> used for local data/transient data requests this will block the next type of
> request from being sent away until the local read is finished and add
> additional latency to the request.
> In addition to this it seems like local requests are *always* data requests
> (might not be a problem), but the log message can say either ["digest" or
> "data"|https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/service/reads/AbstractReadExecutor.java#L156]
> as the type of request.
> I have tried to run performance measurements to see the impact of this in 3.0
> (by moving local requests to the end of ARE#executeAsync()) but I haven't
> seen any big difference yet. I'll continue to run some more tests to see if I
> can find a use case affected by this.
> Attaching a trace (3.0) where this happens. Reproduction:
> # Create a three node CCM cluster
> # Provision data with stress (rf=3)
> # In parallel:
> ## Start stress read run
> ## Run multiple manual read queries in cqlsh with tracing on and
> local_quorum (as this does not always happen)
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
