Pranay Singh has uploaded this change for review. ( 
http://gerrit.cloudera.org:8080/11021


Change subject: IMPALA-6214: Determine and warn about stuck fragment instances.
......................................................................

IMPALA-6214: Determine and warn about stuck fragment instances.

In order to diagnose query hangs, we need to know the fragment execution
time on a particular exec node. Inspecting the query run time profile
to find the cause of hang does not give much details.

This change helps in finding the problematic 'exec node' where the fragment
execution is not making progress. The change makes use of kudu watchdog that
periodically polls and prints the delay in response from an exec node.

Testing:
--------
a) Added a delay on the sender side as a part of manual test case to notice
   the affect of change. The watchdog prints the detail of fragmentID and
   nodeID when the watchdog timer expires.

b) Ran the core test without failure.

Change-Id: I260a1d0a3477e5c6a46094e664500c3e2ed7de62
---
M be/src/common/global-flags.cc
M be/src/runtime/krpc-data-stream-recvr.cc
2 files changed, 27 insertions(+), 1 deletion(-)



  git pull ssh://gerrit.cloudera.org:29418/Impala-ASF refs/changes/21/11021/1
--
To view, visit http://gerrit.cloudera.org:8080/11021
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: newchange
Gerrit-Change-Id: I260a1d0a3477e5c6a46094e664500c3e2ed7de62
Gerrit-Change-Number: 11021
Gerrit-PatchSet: 1
Gerrit-Owner: Pranay Singh

Reply via email to