[
https://issues.apache.org/jira/browse/KUDU-1741?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dinesh Bhat updated KUDU-1741:
------------------------------
Description: Triaging this so that we can attend to this later. MiniCluster
and ExternalMiniCluster utility classes follow opposite semantics for restart
today and that's confusing. For eg, ExternalMiniCluster::Restart() expects that
all the nodes must be shutdown before we do restart, whereas
MiniCluster::Restart() expects that all servers are up and running so that it
can internally execute shutdown and start in that order. We can keep one
semantics here to make it less confusing. We may have to change bunch of
existing tests whichever approach we end up choosing, because tests are also
written following these semantics. (was: Triaging this so that we can attend
to this later. MiniCluster and ExternalMiniCluster utility classes follow
opposite semantics today and that's confusing. For eg,
ExternalMiniCluster::Restart() expects that all the nodes must be shutdown
before we do restart, whereas MiniCluster::Restart() expects that all servers
are up and running so that it can internally execute shutdown and start in that
order. We can keep one semantics here to make it less confusing. We may have to
change bunch of existing tests whichever approach we end up choosing, because
tests are also written following these semantics.)
> Make MiniCluster and ExternalMiniCluster follow one semantics for Restart
> -------------------------------------------------------------------------
>
> Key: KUDU-1741
> URL: https://issues.apache.org/jira/browse/KUDU-1741
> Project: Kudu
> Issue Type: Task
> Reporter: Dinesh Bhat
> Assignee: Dinesh Bhat
> Priority: Trivial
>
> Triaging this so that we can attend to this later. MiniCluster and
> ExternalMiniCluster utility classes follow opposite semantics for restart
> today and that's confusing. For eg, ExternalMiniCluster::Restart() expects
> that all the nodes must be shutdown before we do restart, whereas
> MiniCluster::Restart() expects that all servers are up and running so that it
> can internally execute shutdown and start in that order. We can keep one
> semantics here to make it less confusing. We may have to change bunch of
> existing tests whichever approach we end up choosing, because tests are also
> written following these semantics.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
