mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Meng Zhu <m...@mesosphere.io>
Subject Review Request 65449: Fixed an issue where executor info linger on master if failed to launch.
Date Thu, 01 Feb 2018 02:03:42 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/65449/
-----------------------------------------------------------

Review request for mesos, Benjamin Mahler, Chun-Hung Hsiao, and Vinod Kone.


Bugs: MESOS-1720
    https://issues.apache.org/jira/browse/MESOS-1720


Repository: mesos


Description
-------

Master relies on `ExitedExecutorMessage` from the agent to recycle
executor entry. However, this message won't be sent if the executor
never actually launched (due to transient error), leaving executor
info on the master lingering and resource claimed.
See MESOS-1720.

This patch fixes this issue by sending the `ExitedExecutorMessage`
from the agent if the executor is never launched. And by
setting a new field `launch_executor` in the RunTask(Group)Message,
the master is able to control the executor creation on the agent.


Diffs
-----

  src/master/master.hpp b434d2398b8815811345b6586ca586d2025cb2a2 
  src/master/master.cpp b97ebae6ebfd8ae0f73e617d0c55e140b9c3fce7 
  src/slave/slave.hpp 09c01ebd1b5e8008ba9e7d412042f1db76a2c5a5 
  src/slave/slave.cpp a6a5c93ab2d541c870cb52587495de20ed5ac1f4 
  src/tests/mock_slave.hpp 29ce7140501888d95d5f2d6c26b752ad276b484a 
  src/tests/mock_slave.cpp 8357edc7b3a35624c813eccb9ecca9d3b5dbe07c 
  src/tests/slave_tests.cpp f393a8433a984267adc4db307ef07fcbafd1e62f 


Diff: https://reviews.apache.org/r/65449/diff/1/


Testing
-------


Thanks,

Meng Zhu


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message