Charles Allen created MESOS-6859:
------------------------------------
Summary: Document HA behavior during mesos master replacement
Key: MESOS-6859
URL: https://issues.apache.org/jira/browse/MESOS-6859
Project: Mesos
Issue Type: Documentation
Components: documentation, master
Reporter: Charles Allen
In a discussion in https://mesos.slack.com/archives/general/p1483637159001494
the question was brought up when a "new" master is really fully ready.
Specifically, in the case where new masters can spin up faster than masters can
sync their logs, it is unclear from the HA docs at
http://mesos.apache.org/documentation/latest/high-availability/ how to ensure a
freshly spawned master is ready to take over leadership.
There is documentation at
http://mesos.apache.org/documentation/latest/monitoring/ about using
{{registrar/log/recovered}} to gather this kind of information, but such
information is very easy to overlook.
This ask is that the HA docs be amended to include more information about how
to use {{registrar/log/recovered}} properly.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
