[
https://issues.apache.org/jira/browse/HIVE-18899?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16390420#comment-16390420
]
Gopal V edited comment on HIVE-18899 at 3/7/18 11:36 PM:
---------------------------------------------------------
Tested this with 1000 concurrent users and we're doing really well on cache
hit-rates.
{code}
summary = 83681 in 00:10:51 = 128.6 qps
{code}
LGTM - +1
was (Author: gopalv):
Tested this with 1000 concurrent users and we're doing really well on cache
hit-rates.
{code}
summary = 83681 in 00:10:51 = 128.6/s
{code}
LGTM - +1
> Separate FetchWork required for each query that uses the results cache
> ----------------------------------------------------------------------
>
> Key: HIVE-18899
> URL: https://issues.apache.org/jira/browse/HIVE-18899
> Project: Hive
> Issue Type: Sub-task
> Reporter: Jason Dere
> Assignee: Jason Dere
> Priority: Major
> Attachments: HIVE-18899.1.patch
>
>
> [~gopalv] found issues when running lots of concurrent queries against HS2
> with the query cache. Looks like the FetchWork held by the results cache
> cannot be shared between multiple queries because it contains a
> ListSinkOperator that is used to hold the results of a fetch.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
