mesos-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Neil Conway (JIRA)" <j...@apache.org>
Subject [jira] [Created] (MESOS-7596) Multiple registration attempts might result in agent shutdown
Date Wed, 31 May 2017 17:55:04 GMT
Neil Conway created MESOS-7596:
----------------------------------

             Summary: Multiple registration attempts might result in agent shutdown
                 Key: MESOS-7596
                 URL: https://issues.apache.org/jira/browse/MESOS-7596
             Project: Mesos
          Issue Type: Bug
            Reporter: Neil Conway


This sequence of events is possible:

# Agent sends register message M1 to master.
# Agent register timer expires, sends register message M2 to master.
# Master sees M1 and adds agent with ID A1.
# Agent gets SlaveRegisteredMessage with ID A1.
# The master <-> agent socket breaks; the master marks the agent as disconnected.
# Master sees M2; since the agent is currently disconnected, the master removes A1 and adds
the agent with ID A2.
# Agent gets SlaveRegisteredMessage with ID A2. Since this is unexpected, the agent exits
("Registered but got wrong id").

Shutting down the agent is unfortunate, although arguably not catastrophic.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message