impala-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sailesh Mukil (Code Review)" <>
Subject [Impala-ASF-CR] IMPALA-5750: Catch exceptions from boost thread creation
Date Wed, 30 Aug 2017 22:13:32 GMT
Sailesh Mukil has posted comments on this change.

Change subject: IMPALA-5750: Catch exceptions from boost thread creation

Patch Set 6:

File be/src/exec/

PS6, Line 363: pool->ReleaseThreadToken(false, true);
Do you need to drop the lock before calling this? Based on the comment from L405.

Or is that why you added the 'in_callback' parameter?
File be/src/runtime/

PS6, Line 59:   RETURN_IF_ERROR(Thread::Create("query-exec-mgr",
            :       Substitute("start-query-finstances-$0", PrintId(query_id)),
            :           &QueryExecMgr::StartQueryHelper, this, qs, &t, true));
Shouldn't we call ReleaseQueryState(qs); if the thread failed to start here?
File be/src/runtime/

PS6, Line 199:     // If this fragment instance is not prepared, it must be reporting an error.
             :     DCHECK(fis->IsPrepared() || !status.ok());
Is this necessarily true?

We start the report thread in FragmentInstanceState::Prepare(), and we set the prepared promise
after that. What if the report thread asynchronously attempts to send a status report before
we set the prepared promise?
File be/src/service/

PS6, Line 598: Thread::Create("query-exec-state", "wait-thread",
             :       &ClientRequestState::Wait, this, &wait_thread_);
Why not inject the thread failures here as well? I think it's an important test site since
the thread failure will be injected after the query starts executing.

i.e. WaitAsync() fails after a call to Execute().

To view, visit
To unsubscribe, visit

Gerrit-MessageType: comment
Gerrit-Change-Id: I15a2f278dc71892b7fec09593f81b1a57ab725c0
Gerrit-PatchSet: 6
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: Joe McDonnell <>
Gerrit-Reviewer: Dan Hecht <>
Gerrit-Reviewer: Joe McDonnell <>
Gerrit-Reviewer: Sailesh Mukil <>
Gerrit-Reviewer: Tim Armstrong <>
Gerrit-HasComments: Yes

View raw message