You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "Joseph Wu (JIRA)" <ji...@apache.org> on 2015/07/20 22:55:05 UTC
[jira] [Updated] (MESOS-3077) Registry recovery does not recover
the maintenance object.
[ https://issues.apache.org/jira/browse/MESOS-3077?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Joseph Wu updated MESOS-3077:
-----------------------------
Description:
Persisted info is fetched from the registry when a master is elected or after failover. Currently, this process involves 3 steps:
* Fetch the "registry".
* Start an operation to add the new master to the fetched registry.
* Check the success of the operation and finish recovering.
These methods can be found in src/master/registrar.cpp {code}RegistrarProcess::recover, ::_recover, ::__recover{code}
Since the maintenance schedule is stored in a separate key, the recover process must also fetch a new "maintenance" object. This object needs to be passed along to the master along with the existing "registry" object.
Possible test(s):
* src/tests/registrar_tests.cpp
** Change the "Recovery" test to include checks for the new object.
was:
Persisted info is fetched from the registry when a master is elected or after failover. Currently, this process involves 3 steps:
* Fetch the "registry".
* Start an operation to add the new master to the fetched registry.
* Check the success of the operation and finish recovering.
These methods can be found in src/master/registrar.cpp {code}RegistrarProcess::recover, ::_recover, ::__recover{code}.
Since the maintenance schedule is stored in a separate key, the recover process must also fetch a new "maintenance" object. This object needs to be passed along to the master along with the existing "registry" object.
Possible test(s):
* src/tests/registrar_tests.cpp
** Change the "Recovery" test to include checks for the new object.
> Registry recovery does not recover the maintenance object.
> ----------------------------------------------------------
>
> Key: MESOS-3077
> URL: https://issues.apache.org/jira/browse/MESOS-3077
> Project: Mesos
> Issue Type: Task
> Components: master, replicated log
> Reporter: Joseph Wu
> Assignee: Joseph Wu
>
> Persisted info is fetched from the registry when a master is elected or after failover. Currently, this process involves 3 steps:
> * Fetch the "registry".
> * Start an operation to add the new master to the fetched registry.
> * Check the success of the operation and finish recovering.
> These methods can be found in src/master/registrar.cpp {code}RegistrarProcess::recover, ::_recover, ::__recover{code}
> Since the maintenance schedule is stored in a separate key, the recover process must also fetch a new "maintenance" object. This object needs to be passed along to the master along with the existing "registry" object.
> Possible test(s):
> * src/tests/registrar_tests.cpp
> ** Change the "Recovery" test to include checks for the new object.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)