lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jack Krupansky" <j...@basetechnology.com>
Subject Re: Terminology question: Core vs. Collection vs...
Date Thu, 03 Jan 2013 14:08:18 GMT
Oops... let me word that a little more carefully:

...we are "replicating the data of each shard".





-- Jack Krupansky
-----Original Message----- 
From: Jack Krupansky
Sent: Thursday, January 03, 2013 9:03 AM
To: solr-user@lucene.apache.org
Subject: Re: Terminology question: Core vs. Collection vs...

No, a shard is a subset (or "slice") of the collection. Sharding is a way of
"slicing" the original data, before we talk about how the shards get stored
and replicated on actual Solr cores. Replicas are instances of the data for
a shard.

Sometimes people may loosely speak of a replica as being "a shard", but
that's just loose use of the terminology.

So, we're not "sharding shards", but we are "replicating shards".

-- Jack Krupansky

-----Original Message----- 
From: Darren Govoni
Sent: Thursday, January 03, 2013 8:51 AM
To: solr-user@lucene.apache.org
Subject: RE: Re: Terminology question: Core vs. Collection vs...

Thanks again. (And sorry to jump into this convo)

But I had a question on your statement:

On 1/3/2013 08:07 AM Jack Krupansky wrote:
   <br>Collection is the more modern term and incorporates the fact that the
<br>collection may be sharded, with each shard on one or more cores, with
each <br>core being a replica of the other cores within that shard of that
<br>collection.

A collection is sharded, meaning it is distributed across cores. A shard
itself is not distributed across cores in the same since. Rather a shard
exist on a single core and is replicated on other cores. Is that right? The
way its worded above, it sounds like a shard can also be sharded...


<br><br><br>------- Original Message -------
On 1/3/2013  08:28 AM Jack Krupansky wrote:<br>A node is a machine in a
cluster or cloud (graph). It could be a real
<br>machine or a virtualized machine. Technically, you could have multiple
<br>virtual nodes on the same physical "box". Each Solr replica would be on
a
<br>different node.
<br>
<br>Technically, you could have multiple Solr instances running on a single
<br>hardware node, each with a different port. They are simply instances of
<br>Solr, although you could consider each Solr instance a node in a Solr
cloud
<br>as well, a "virtual" node. So, technically, you could have multiple
replicas
<br>on the same node, but that sort of defeats most of the purpose of having
<br>replicas in the first place - to distribute the data for performance and
<br>fault tolerance. But, you could have replicas of different shards on the
<br>same node/box for a partial improvement of performance and fault
tolerance.
<br>
<br>A Solr "cloud' is really a cluster.
<br>
<br>-- Jack Krupansky
<br>
<br>-----Original Message----- 
<br>From: Darren Govoni
<br>Sent: Thursday, January 03, 2013 8:16 AM
<br>To: solr-user@lucene.apache.org
<br>Subject: RE: Re: Terminology question: Core vs. Collection vs...
<br>
<br>Good write up.
<br>
<br>And what about "node"?
<br>
<br>I think there needs to be an official glossary of terms that is
sanctioned
<br>by the solr team and some terms still ni use may need to be labeled
<br>"deprecated". After so many years, its still confusing.
<br>
<br><br><br><br>------- Original Message -------
<br>On 1/3/2013  08:07 AM Jack Krupansky wrote:<br>Collection is the more
modern
<br>term and incorporates the fact that the
<br><br>collection may be sharded, with each shard on one or more cores,
with
<br>each
<br><br>core being a replica of the other cores within that shard of that
<br><br>collection.
<br><br>
<br><br>Instance is a general term, but is commonly used to refer to a
running
<br>Solr
<br><br>server, each of which can service any number of cores. A sharded
<br>collection
<br><br>would typically require multiple instances of Solr, each with a
shard of
<br>the
<br><br>collection.
<br><br>
<br><br>Multiple collections can be supported on a single instance of Solr.
They
<br><br>don't have to be sharded or replicated. But if they are, each Solr
<br>instance
<br><br>will have a copy or replica of the data (index) of one shard of each
<br>sharded
<br><br>collection - to the degree that each collection needs that many
shards.
<br><br>
<br><br>At the API level, you talk to a Solr instance, using a host and
port,
<br>and
<br><br>giving the collection name. Some operations will refer only to the
<br>portion
<br><br>of a multi-shard collection on that Solr instance, but typically
Solr
<br>will
<br><br>"distribute" the operation, whether it be an update or a query, to
all
<br>of
<br><br>the shards of the named collection. In the case of update, the
update
<br>will
<br><br>be distributed to all replicas as well, but in the case of query
only
<br>one
<br><br>replica of each shard of the collection is needed.
<br><br>
<br><br>Before SolrCloud we Solr had master and slave and the slaves were
<br>replicas
<br><br>of the master, but with SolrCloud there is no master and all the
<br>replicas of
<br><br>the shard are peers, although at any moment of time one of them will
be
<br><br>considered the "leader" for coordination purposes, but not in the
sense
<br>that
<br><br>it is a master of the other replicas in that shard. A SolrCloud
replica
<br>is a
<br><br>replica of the data, in an abstract sense, for a single shard of a
<br><br>collection. A SolrCloud replica is more of an instance of the
<br>data/index.
<br><br>
<br><br>An index exists at two levels: the portion of a collection on a
single
<br>Solr
<br><br>core will have a Lucene index, but collectively the Lucene indexes
for
<br>the
<br><br>shards of a collection can be referred to the index of the
collection.
<br>Each
<br><br>replica is a copy or instance of a portion of the collection's
index.
<br><br>
<br><br>The term slice is sometimes used to refer collectively to all of the
<br><br>cores/replicas of a single shard, or sometimes to a single replica
as it
<br><br>contains only a "slice" of the full collection data.
<br><br>
<br><br>-- Jack Krupansky
<br><br>
<br><br>-----Original Message----- 
<br><br>From: Alexandre Rafalovitch
<br><br>Sent: Thursday, January 03, 2013 4:42 AM
<br><br>To: solr-user@lucene.apache.org
<br><br>Subject: Terminology question: Core vs. Collection vs...
<br><br>
<br><br>Hello,
<br><br>
<br><br>I am trying to understand the core Solr terminology. I am looking
for
<br><br>correct rather than loose meaning as I am trying to teach an example
<br>that
<br><br>starts from easy scenario and may scale to multi-core, multi-machine
<br><br>situation.
<br><br>
<br><br>Here are the terms that seem to be all overlapping and/or crossing
over
<br>in
<br><br>my mind a the moment.
<br><br>
<br><br>1) Index
<br><br>2) Core
<br><br>3) Collection
<br><br>4) Instance
<br><br>5) Replica (Replica of _what_?)
<br><br>6) Others?
<br><br>
<br><br>I tried looking through documentation, but either there is a
terminology
<br><br>drift or I am having trouble understanding the distinctions.
<br><br>
<br><br>If anybody has a clear picture in their mind, I would appreciate a
<br><br>clarification.
<br><br>
<br><br>Regards,
<br><br>   Alex.
<br><br>
<br><br>Personal blog: http://blog.outerthoughts.com/
<br><br>LinkedIn: http://www.linkedin.com/in/alexandrerafalovitch
<br><br>- Time is the quality of nature that keeps events from happening all
at
<br><br>once. Lately, it doesn't seem to be working.  (Anonymous  - via GTD
<br>book)
<br><br>
<br><br>
<br>
<br> 


Mime
View raw message