[ 
https://issues.apache.org/jira/browse/MESOS-8887?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16769813#comment-16769813
 ] 

Vinod Kone edited comment on MESOS-8887 at 2/15/19 10:38 PM:
-------------------------------------------------------------

Landed on master:
-------------------------------------------------------------------------------
commit 1a506a4536a4b79dba6634d8dc627eaf2a55caba
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Tue Feb 5 16:55:19 2019 -0600

    Tested unreachable task behavior on agent GC.
    
    Updated `PartitionTest, RegistryGcByCount` test. This test fails
    without the previous patch.
    
    Review: https://reviews.apache.org/r/69909

commit c72a4f909054e5efa75d9e5d8dde71b0083402c1
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Sat Feb 2 10:01:56 2019 -0600

    Removed unreachable tasks from `Master::Framework` on agent GC.
    
    Unreachable tasks are stored in `Slaves` and `Framework` structs of
    the master, but they were only being removed from the former when
    an unreachable agent is GCed from the registry. This patch fixes it
    so that the latter is also cleaned up.
    
    Review: https://reviews.apache.org/r/69908

commit f0cd3b7b62807fe377b1b47bf1bf364b18c4a373
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Sat Feb 2 09:51:09 2019 -0600

    Fixed variable names in `Master::_doRegistryGC()`.
    
    Substituted `slave` with `slaveId` to be consistent with the code base.
    No functional changes.
    
    Review: https://reviews.apache.org/r/69907



-------------------------------------------------------------------------------
Backported to 1.7.x
-------------------------------------------------------------------------------
commit 6fcf70167076bbe6fb10ca04876939fe0e3379d9
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Fri Feb 15 14:33:00 2019 -0600

    Added MESOS-8887 to the 1.7.2 CHANGELOG.

commit 1a506a4536a4b79dba6634d8dc627eaf2a55caba
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Tue Feb 5 16:55:19 2019 -0600

    Tested unreachable task behavior on agent GC.
    
    Updated `PartitionTest, RegistryGcByCount` test. This test fails
    without the previous patch.
    
    Review: https://reviews.apache.org/r/69909

commit c72a4f909054e5efa75d9e5d8dde71b0083402c1
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Sat Feb 2 10:01:56 2019 -0600

    Removed unreachable tasks from `Master::Framework` on agent GC.
    
    Unreachable tasks are stored in `Slaves` and `Framework` structs of
    the master, but they were only being removed from the former when
    an unreachable agent is GCed from the registry. This patch fixes it
    so that the latter is also cleaned up.
    
    Review: https://reviews.apache.org/r/69908

commit f0cd3b7b62807fe377b1b47bf1bf364b18c4a373
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Sat Feb 2 09:51:09 2019 -0600

    Fixed variable names in `Master::_doRegistryGC()`.
    
    Substituted `slave` with `slaveId` to be consistent with the code base.
    No functional changes.
    
    Review: https://reviews.apache.org/r/69907



was (Author: vinodkone):
Landed on master:

commit 1a506a4536a4b79dba6634d8dc627eaf2a55caba
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Tue Feb 5 16:55:19 2019 -0600

    Tested unreachable task behavior on agent GC.
    
    Updated `PartitionTest, RegistryGcByCount` test. This test fails
    without the previous patch.
    
    Review: https://reviews.apache.org/r/69909

commit c72a4f909054e5efa75d9e5d8dde71b0083402c1
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Sat Feb 2 10:01:56 2019 -0600

    Removed unreachable tasks from `Master::Framework` on agent GC.
    
    Unreachable tasks are stored in `Slaves` and `Framework` structs of
    the master, but they were only being removed from the former when
    an unreachable agent is GCed from the registry. This patch fixes it
    so that the latter is also cleaned up.
    
    Review: https://reviews.apache.org/r/69908

commit f0cd3b7b62807fe377b1b47bf1bf364b18c4a373
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Sat Feb 2 09:51:09 2019 -0600

    Fixed variable names in `Master::_doRegistryGC()`.
    
    Substituted `slave` with `slaveId` to be consistent with the code base.
    No functional changes.
    
    Review: https://reviews.apache.org/r/69907


Backported to 1.7.x

commit 6fcf70167076bbe6fb10ca04876939fe0e3379d9
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Fri Feb 15 14:33:00 2019 -0600

    Added MESOS-8887 to the 1.7.2 CHANGELOG.

commit 1a506a4536a4b79dba6634d8dc627eaf2a55caba
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Tue Feb 5 16:55:19 2019 -0600

    Tested unreachable task behavior on agent GC.
    
    Updated `PartitionTest, RegistryGcByCount` test. This test fails
    without the previous patch.
    
    Review: https://reviews.apache.org/r/69909

commit c72a4f909054e5efa75d9e5d8dde71b0083402c1
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Sat Feb 2 10:01:56 2019 -0600

    Removed unreachable tasks from `Master::Framework` on agent GC.
    
    Unreachable tasks are stored in `Slaves` and `Framework` structs of
    the master, but they were only being removed from the former when
    an unreachable agent is GCed from the registry. This patch fixes it
    so that the latter is also cleaned up.
    
    Review: https://reviews.apache.org/r/69908

commit f0cd3b7b62807fe377b1b47bf1bf364b18c4a373
Author: Vinod Kone <vinodk...@gmail.com>
Date:   Sat Feb 2 09:51:09 2019 -0600

    Fixed variable names in `Master::_doRegistryGC()`.
    
    Substituted `slave` with `slaveId` to be consistent with the code base.
    No functional changes.
    
    Review: https://reviews.apache.org/r/69907


> Unreachable tasks are not GC'ed when unreachable agent is GC'ed.
> ----------------------------------------------------------------
>
>                 Key: MESOS-8887
>                 URL: https://issues.apache.org/jira/browse/MESOS-8887
>             Project: Mesos
>          Issue Type: Bug
>          Components: master
>    Affects Versions: 1.4.3, 1.5.2, 1.6.1, 1.7.1
>            Reporter: Gilbert Song
>            Assignee: Vinod Kone
>            Priority: Major
>              Labels: foundations, mesosphere, partition, registry
>             Fix For: 1.7.2, 1.8.0
>
>
> Unreachable agents will be gc-ed by the master registry after 
> `--registry_max_agent_age` duration or `--registry_max_agent_count`. When the 
> GC happens, the agent will be removed from the master's unreachable agent 
> list, but its corresponding tasks are still in UNREACHABLE state in the 
> framework struct (though removed from `slaves.unreachableTasks`). We should 
> instead remove those tasks from everywhere or transition those tasks to a 
> terminal state, either TASK_LOST or TASK_GONE (further discussion is needed 
> to define the semantic).
> This improvement relates to how do we want to couple the update of task with 
> the GC of agent. Right now they are somewhat decoupled.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to