You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "adyatlov@gmail.com (JIRA)" <ji...@apache.org> on 2015/04/16 16:10:59 UTC

[jira] [Updated] (MESOS-2503) Check TaskStatus::Reason in tests

     [ https://issues.apache.org/jira/browse/MESOS-2503?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

adyatlov@gmail.com updated MESOS-2503:
--------------------------------------
    Description: 
We have a number of reasons for task status updates, but we do not test all of them. On the other side, some failures that we check in our tests do not have reason field set. It would be nice to have a check for state, reason and source for every failed or lost task. E.g.
{code}
  AWAIT_READY(status4);
  EXPECT_EQ(TASK_LOST, status4.get().state());
  EXPECT_EQ(TaskStatus::SOURCE_SLAVE, status4.get().source());
  EXPECT_EQ(TaskStatus::REASON_EXECUTOR_TERMINATED, status4.get().reason());
{code}

To solve this problem several changes in different parts of the codebase should be made. Hence we make an epic out of it and do it step-by-step.

  was:
We have a number of reasons for task status updates, but we do not test all of them. On the other side, some failures that we check in our tests do not have reason field set. It would be nice to have a check for state, reason and source for every failed or lost task. E.g.
{code}
  AWAIT_READY(status4);
  EXPECT_EQ(TASK_LOST, status4.get().state());
  EXPECT_EQ(TaskStatus::SOURCE_SLAVE, status4.get().source());
  EXPECT_EQ(TaskStatus::REASON_EXECUTOR_TERMINATED, status4.get().reason());
{code}

     Issue Type: Epic  (was: Improvement)

> Check TaskStatus::Reason in tests
> ---------------------------------
>
>                 Key: MESOS-2503
>                 URL: https://issues.apache.org/jira/browse/MESOS-2503
>             Project: Mesos
>          Issue Type: Epic
>          Components: test
>            Reporter: Alexander Rukletsov
>            Assignee: adyatlov@gmail.com
>
> We have a number of reasons for task status updates, but we do not test all of them. On the other side, some failures that we check in our tests do not have reason field set. It would be nice to have a check for state, reason and source for every failed or lost task. E.g.
> {code}
>   AWAIT_READY(status4);
>   EXPECT_EQ(TASK_LOST, status4.get().state());
>   EXPECT_EQ(TaskStatus::SOURCE_SLAVE, status4.get().source());
>   EXPECT_EQ(TaskStatus::REASON_EXECUTOR_TERMINATED, status4.get().reason());
> {code}
> To solve this problem several changes in different parts of the codebase should be made. Hence we make an epic out of it and do it step-by-step.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)