[
https://issues.apache.org/jira/browse/KUDU-2653?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16734542#comment-16734542
]
Adar Dembo commented on KUDU-2653:
----------------------------------
I took a look at the logs. Pretty much every test that uses an external mini
cluster generated at least one leak that looked like this:
{noformat}
Direct leak of 16 byte(s) in 1 object(s) allocated from:
#0 0x4f5ec8 in __interceptor_malloc
/mnt/ddb/2/helif/github/kudu/thirdparty/src/llvm-6.0.0.src/projects/compiler-rt/lib/asan/asan_malloc_linux.cc:88
#1 0x7fb53b8ebc77 in glob64 (/lib/x86_64-linux-gnu/libc.so.6+0xbdc77)
#2 0x46bf89 in __interceptor_glob
sanitizer_common/sanitizer_common_interceptors.inc:2222
#3 0x7fb540cd65f7 (/usr/lib/x86_64-linux-gnu/libgssapi_krb5.so.2+0x185f7)
#4 0x7ffcb7e1c5ef ([stack]+0x165ef)
{noformat}
I suspect a known memory leak in some version of libgssapi_krb5.so, which we
may need to suppress.
[~helifu], could you rerun one of the failing tests with
LSAN_OPTIONS=fast_unwind_on_malloc=0 set in your environment? That may produce
a more useful stack trace. See https://github.com/google/sanitizers/issues/870
for more details.
> The ASAN test failed on Debian 8.9
> ----------------------------------
>
> Key: KUDU-2653
> URL: https://issues.apache.org/jira/browse/KUDU-2653
> Project: Kudu
> Issue Type: Bug
> Components: test
> Affects Versions: 1.4.0, 1.7.0, 1.8.0
> Reporter: HeLifu
> Assignee: Alexey Serbin
> Priority: Major
> Attachments: ctest output.txt, test-logs.tar.gz
>
>
> I tried to run ASAN test of branch 1.4.x, 1.7.x, 1.8.x and master on Debian
> 8.9, but all failed. It seems that there is a historical issue there. The
> errors are in the attachments.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
