[
https://issues.apache.org/jira/browse/CASSANDRA-14696?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16604998#comment-16604998
]
Ariel Weisberg commented on CASSANDRA-14696:
--------------------------------------------
This is used to fix a more general issue with read and write latency tracking
where when we have no data we end up using 0 as the value for speculation
timeouts. This is different from using the default (1/2 the timeout) or the
value from back when we had data.
> DecayingEstimatedHistogramReservoir.EstimatedHistogramReservoirSnapshot
> returns wrong value for size() and incorrectly calculates count
> ----------------------------------------------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-14696
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14696
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Reporter: Ariel Weisberg
> Assignee: Ariel Weisberg
> Priority: Major
> Fix For: 4.0
>
>
> It appears to return the wrong value for no reason. There isn't a single
> instance in the code base where we use size. There is an internal count that
> is calculated, but it is calculated over an empty array.
> Fix both of these bugs and then use the size to only populate read and write
> latency values used for speculation when there are actual samples being used.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
