[
https://issues.apache.org/jira/browse/IMPALA-8539?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16838990#comment-16838990
]
Bikramjeet Vig commented on IMPALA-8539:
----------------------------------------
From the logs this seems like it happened in test_shutdown_coordinator based on
the queries that it ran. Also it seems this error is from a backend that was
not being shutdown by the shutdown command and was working perfectly during the
test. It seems like the only time this is possible is when the backend process
was already getting killed, and the heap-use-after-free failure was triggered
after it received the SIGTERM signal.
> ASAN heap-use-after-free failure during TestGracefulShutdown
> ------------------------------------------------------------
>
> Key: IMPALA-8539
> URL: https://issues.apache.org/jira/browse/IMPALA-8539
> Project: IMPALA
> Issue Type: Bug
> Components: Backend
> Affects Versions: Impala 3.3.0
> Reporter: David Rorke
> Priority: Major
>
> I'm seeing an ASAN heap-use-after-free failure with no stack trace in the
> stderr. Here's the full ERROR log:
> {noformat}
> Log file created at: 2019/05/10 06:32:22
> Running on machine:
> impala-ec2-centos74-r4-4xlarge-ondemand-08f1.vpc.cloudera.com
> Log line format: [IWEF]mmdd hh:mm:ss.uuuuuu threadid file:line] msg
> E0510 06:32:22.212716 102164 logging.cc:147] stderr will be logged to this
> file.
> 19/05/10 06:32:23 INFO util.JvmPauseMonitor: Starting JVM pause monitor
> =================================================================
> ==102164==ERROR: AddressSanitizer: heap-use-after-free on address
> 0x603000032668 at pc 0x0000017aa3dc bp 0x7ff292197ae0 sp 0x7ff292197290
> READ of size 4 at 0x603000032668 thread T3
> Picked up JAVA_TOOL_OPTIONS:
> -agentlib:jdwp=transport=dt_socket,address=30000,server=y,suspend=n
> {noformat}
> With no stack trace it's hard to say exactly what triggered it but
> correlating the timing of the error log entry with the Jenkins console logs
> it looks like we were running one of the graceful shutdown tests:
> {noformat}
> 06:32:19
> custom_cluster/test_restart_services.py::TestGracefulShutdown::test_sending_multiple_shutdown_signals
> PASSED
> 06:32:37
> custom_cluster/test_restart_services.py::TestGracefulShutdown::test_shutdown_coordinator
> <- hs2/hs2_test_suite.py PASSED
> 06:32:37
> custom_cluster/test_rpc_exception.py::TestRPCException::test_rpc_send_closed_connection[protocol:
> beeswax | exec_option: {'batch_size': 0, 'num_nodes': 0,
> 'disable_codegen_rows_threshold': 0, 'disable_codegen': False,
> 'abort_on_error': 1, 'exec_single_node_rows_threshold': 0} | table_format:
> text/none] SKIPPED
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
