[
https://issues.apache.org/jira/browse/IMPALA-11176?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17514999#comment-17514999
]
ASF subversion and git services commented on IMPALA-11176:
----------------------------------------------------------
Commit 38062b4ec0cfdf28474c84d136718510fe712437 in impala's branch
refs/heads/master from Yida Wu
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=38062b4 ]
IMPALA-11196 Fix a ClientCacheTest flaky test
The memory leak detection testcase is flaky, in local environment,
it could fail over 10 times out of 100. The fix changes to use
the ordinary ADD_BE_TEST to replace ADD_UNIFIED_BE_LSAN_TEST,
there could be some elements in the unified test which affect the
process vm size occasionally.
After the fix, the testcase ran 5000 times without a failure, and
can still catch the memory leak issue in IMPALA-11176.
Tests:
Ran Core Tests in DEBUG/RELEASE/ASAN build.
Ran consecutive 5000 times of the testcase and all passed.
Change-Id: I6ee273db4f9b0a518b6438df6982325f5ad1fa0c
Reviewed-on: http://gerrit.cloudera.org:8080/18363
Reviewed-by: Impala Public Jenkins <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>
> Memory leak in ClientCacheHelper
> --------------------------------
>
> Key: IMPALA-11176
> URL: https://issues.apache.org/jira/browse/IMPALA-11176
> Project: IMPALA
> Issue Type: Bug
> Components: Backend
> Affects Versions: Impala 4.0.0, Impala 3.4.0
> Reporter: Yida Wu
> Assignee: Yida Wu
> Priority: Critical
> Fix For: Impala 4.1.0
>
>
> ClientCacheHelper::CreateClient uses a function pointer "factory_method" to
> create a new ThriftClientImpl object and store it in a share pointer of type
> ThriftClientImpl.
> However, in the implementation of the function pointer, which uses
> ClientCache::MakeClient, it creates a derived object of ThriftClientImpl and
> return the pointer as the base object pointer ThriftClientImpl*. When the
> ClientCacheHelper::CreateClient uses the returned pointer to create a base
> shared pointer, it can't recognize that the pointer is from a derived object,
> therefore something goes wrong that the shared pointer couldn't release all
> the memory occupied by the derived object.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
