[
https://issues.apache.org/jira/browse/HBASE-19867?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16341806#comment-16341806
]
Duo Zhang commented on HBASE-19867:
-----------------------------------
{quote}
Add Timeout rule to enforce category based timeout.
{quote}
We need to add a rule explicitly to enforce the timeout? I used to think it is
enabled by default...
{quote}
Looking at runtimes, it's probably enough to split it in lesser class?
{quote}
For safety I want to split it to more classes... For one of the mid cluster
test we use 100000 regions and it will cost a lot of time to run so I placed
the tests for different cluster sizes.
And for region replica, we use a 2m iteration count in one test and a 10m
interation count in another test which makes the test very slow, so I split
them out...
I think later we should revisit these tests to see whether we really need these
large numbers in UT...
> Split TestStochasticLoadBalancer into several small tests
> ---------------------------------------------------------
>
> Key: HBASE-19867
> URL: https://issues.apache.org/jira/browse/HBASE-19867
> Project: HBase
> Issue Type: Sub-task
> Reporter: Duo Zhang
> Assignee: Duo Zhang
> Priority: Major
> Fix For: 2.0.0-beta-2
>
> Attachments: HBASE-19867.patch, image-2018-01-26-12-56-49-560.png
>
>
> Just like what I have done to TestHCM.
> The test is CPU bound. We set a very large number of regions and also a very
> large number iterations which makes some of the tests execute for a very long
> time. It even make things worse that, our machines to run the jenkins job are
> very weak, so it will take much longer time than running them locally...
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
