[
https://issues.apache.org/jira/browse/SOLR-14716?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Cassandra Targett updated SOLR-14716:
-------------------------------------
Description:
The effort to remove oppressive terminology in SOLR-14702 led to somewhat
awkward phrasing on how to refer to non-SolrCloud configurations, specifically
"leader/follower mode", which is potentially very confusing since SolrCloud
also has leaders and one could consider replicas to be followers.
I propose that we standardize what we call these two modes as "coordinated
mode" (SolrCloud) and "uncoordinated mode" (or "non-coordinated" if people
prefer). I chose this because in thinking about what really differentiates the
two approaches is the ZooKeeper coordination for requests, configs, etc. There
are other differences too, of course, but that's the biggest one that stuck out
to me as a key differentiator and applicable in the naming.
There are also places in the Ref Guide where we refer to "standalone mode",
which in many cases means "any cluster not running SolrCloud". This has always
been problematic, because the word "standalone" implies a single node, but it's
of course pretty much always been possible to have a cluster of multiple nodes
that don't run SolrCloud/ZK. This issue would address those examples also.
Note that I'm not proposing replacing the word "SolrCloud" throughout the
documentation. Instead I'll augment the use of the word "SolrCloud" with
clarification that this term means "coordinated mode". Later if we ever replace
SolrCloud references in code and fully remove that name, the conceptual
groundwork will have already been laid for users.
was:
The effort to remove oppressive terminology in SOLR-14702 led to somewhat
awkward phrasing on how to refer to non-SolrCloud configurations, specifically
"leader/follower mode", which is potentially very confusing since SolrCloud
also has leaders and one could consider replicas to be followers.
I propose that we standardize what we call these two modes as "coordinated
mode" (SolrCloud) and "uncoordinated mode" (or "non-coordinated" if people
prefer). I chose this because in thinking about what really differentiates the
two approaches is the ZooKeeper coordination for requests. There are other
differences too, of course, but that's the biggest one that stuck out to me as
a key differentiator and applicable in the naming.
There are also places in the Ref Guide where we refer to "standalone mode",
which in many cases means "any cluster not running SolrCloud". This has always
been problematic, because the word "standalone" implies a single node, but it's
of course pretty much always been possible to have a cluster of multiple nodes
that don't run SolrCloud/ZK. This issue would address those examples also.
Note that I'm not proposing replacing the word "SolrCloud" throughout the
documentation. Instead I'll augment the use of the word "SolrCloud" with
clarification that this term means "coordinated mode". Later if we ever replace
SolrCloud references in code and fully remove that name, the conceptual
groundwork will have already been laid for users.
> Ref Guide: update leader/follower terminology
> ---------------------------------------------
>
> Key: SOLR-14716
> URL: https://issues.apache.org/jira/browse/SOLR-14716
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Components: documentation
> Reporter: Cassandra Targett
> Priority: Major
>
> The effort to remove oppressive terminology in SOLR-14702 led to somewhat
> awkward phrasing on how to refer to non-SolrCloud configurations,
> specifically "leader/follower mode", which is potentially very confusing
> since SolrCloud also has leaders and one could consider replicas to be
> followers.
> I propose that we standardize what we call these two modes as "coordinated
> mode" (SolrCloud) and "uncoordinated mode" (or "non-coordinated" if people
> prefer). I chose this because in thinking about what really differentiates
> the two approaches is the ZooKeeper coordination for requests, configs, etc.
> There are other differences too, of course, but that's the biggest one that
> stuck out to me as a key differentiator and applicable in the naming.
> There are also places in the Ref Guide where we refer to "standalone mode",
> which in many cases means "any cluster not running SolrCloud". This has
> always been problematic, because the word "standalone" implies a single node,
> but it's of course pretty much always been possible to have a cluster of
> multiple nodes that don't run SolrCloud/ZK. This issue would address those
> examples also.
> Note that I'm not proposing replacing the word "SolrCloud" throughout the
> documentation. Instead I'll augment the use of the word "SolrCloud" with
> clarification that this term means "coordinated mode". Later if we ever
> replace SolrCloud references in code and fully remove that name, the
> conceptual groundwork will have already been laid for users.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
