ElasticSearch 1.0 features?

Shay commented on an issue about one possible feature in the 1.0 release:

Which leads me to the question: what is in store for 1.0? There are issues
tagged v1.0.0.Beta1, so 1.0 is being actively developed.

I would love to know the status of some longstanding requests such as:

Also, when will the 0.20 (last Lucene 3.x) and 0.90 branches stop being
worked on? There are a few commits for 0.20.7, which has not been released.

Still on 0.20, waiting for 1.0. If there are no relevant features, I might
switch to 0.90 instead.

Thanks for all the hard work,

Ivan

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Re split brain bug (
https://github.com/elasticsearch/elasticsearch/issues/2117)

Is there a reason why ES doesn't move to ZooKeeper as its default, like
SolrCloud did?

On Wed, May 8, 2013 at 8:28 PM, Ivan Brusic ivan@brusic.com wrote:

Shay commented on an issue about one possible feature in the 1.0 release:
https://github.com/elasticsearch/elasticsearch/issues/1076#issuecomment-17566680

Which leads me to the question: what is in store for 1.0? There are issues
tagged v1.0.0.Beta1, so 1.0 is being actively developed.

I would love to know the status of some longstanding requests such as:

Also, when will the 0.20 (last Lucene 3.x) and 0.90 branches stop being
worked on? There are a few commits for 0.20.7, which has not been released.

Still on 0.20, waiting for 1.0. If there are no relevant features, I might
switch to 0.90 instead.

Thanks for all the hard work,

Ivan

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

There is a plugin to utilize ZooKeeper:

SolrCloud defaults to a single ZooKeeper instance, but in reality you need
a ZooKeeper cluster for production installs. I personally do not want to
manage another cluster if possible.

The only reason why I mentioned the split-brain issue is that since it
affects zen discovery, I can envision the solution requiring some
substantial (and breaking) changes. If something is going to be
re-architected, hopefully it will be in a new version (and not have
breaking changes between minor versions).

Cheers,

Ivan

On Wed, May 8, 2013 at 11:11 AM, Itamar Syn-Hershko itamar@code972.comwrote:

Re split brain bug (
https://github.com/elasticsearch/elasticsearch/issues/2117)

Is there a reason why ES doesn't move to ZooKeeper as its default, like
SolrCloud did?

On Wed, May 8, 2013 at 8:28 PM, Ivan Brusic ivan@brusic.com wrote:

Shay commented on an issue about one possible feature in the 1.0 release:
https://github.com/elasticsearch/elasticsearch/issues/1076#issuecomment-17566680

Which leads me to the question: what is in store for 1.0? There are
issues tagged v1.0.0.Beta1, so 1.0 is being actively developed.

I would love to know the status of some longstanding requests such as:

Also, when will the 0.20 (last Lucene 3.x) and 0.90 branches stop being
worked on? There are a few commits for 0.20.7, which has not been released.

Still on 0.20, waiting for 1.0. If there are no relevant features, I
might switch to 0.90 instead.

Thanks for all the hard work,

Ivan

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Zookeeper is overkill. It comes with a load of features. ES is not
making use of them.

As Igor Motov noted in the elasticsearch-zookeeper plugin, most
split-brains happen due to irregular node outages

https://groups.google.com/d/msg/elasticsearch-zookeeper/49V9baDFoZg/J1sOZKSvLwkJ

and Zookeeper can not fix that outages either.

On the other hand, there are some leader election algorithms who do not
operate with quorums. Hirschberg-Sinclair may be an option, maybe it
behaves better in some network disconnection scenarios.

Jörg

Am 08.05.13 20:11, schrieb Itamar Syn-Hershko:

Re split brain bug
(https://github.com/elasticsearch/elasticsearch/issues/2117)

Is there a reason why ES doesn't move to ZooKeeper as its default,
like SolrCloud did?

On Wed, May 8, 2013 at 8:28 PM, Ivan Brusic <ivan@brusic.com
mailto:ivan@brusic.com> wrote:

Shay commented on an issue about one possible feature in the 1.0
release:
https://github.com/elasticsearch/elasticsearch/issues/1076#issuecomment-17566680

Which leads me to the question: what is in store for 1.0? There
are issues tagged v1.0.0.Beta1, so 1.0 is being actively developed.

I would love to know the status of some longstanding requests such as:
* the referenced facet refactor
* full cluster restarts for upgrades
(https://groups.google.com/forum/#!msg/elasticsearch/p2mu0Tv3VPI/0sXdwuwCKR8J
<https://groups.google.com/forum/#%21msg/elasticsearch/p2mu0Tv3VPI/0sXdwuwCKR8J>)
* better backups/export
* split brain bug
(https://github.com/elasticsearch/elasticsearch/issues/2117)

Also, when will the 0.20 (last Lucene 3.x) and 0.90 branches stop
being worked on? There are a few commits for 0.20.7, which has not
been released.

Still on 0.20, waiting for 1.0. If there are no relevant features,
I might switch to 0.90 instead.

Thanks for all the hard work,

Ivan
-- 
You received this message because you are subscribed to the Google
Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to elasticsearch+unsubscribe@googlegroups.com
<mailto:elasticsearch%2Bunsubscribe@googlegroups.com>.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google
Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send
an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Very early for christmas, but my (never ending) wish list is:

  • pluggable faceting framework: collations for sorting facet entries,
    histograms with equidistant and non-equidistant binning (uniform and
    non-uniform widths), and the longstanding "grouping/collapsing" feature
    (supervised clustering of docs)

  • Websockets as an alternative to REST HTTP, together with an event
    source on node level to pub/sub for ES nodex/index/shard events, and for
    a streaming API for "piping" JSON into ES bulk indexing

  • contineous datacenter to datacenter replication (Websockets might help
    to "tee" an index to multiple targets in the background, without discovery)

  • query estimator to block "evil"/"heavy" queries from executing

  • resource watchdog to detect low resource conditions and execute
    emergency node actions (throttle, admin mail, shutdown)

  • Lucene DocValues for fast key/value store (should be there already,
    haven't checked)

  • classifiers for terms/documents (unsupervised clustering of docs)

  • improving discovery algorithms (leader election)

  • lots of tooling OOTB for the lazy admin (web console, backup/restore
    tool, upgrade tool, hooks for deployment scripts)

... nuff' for ES 2.0, 3.0 and 4.0 if you ask me :slight_smile:

Jörg

Am 08.05.13 19:28, schrieb Ivan Brusic:

I would love to know the status of some longstanding requests such as:

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

My list is based on previous discussions of issues that could potentially
be featured in future releases (1.0). If I had to compile a wish list, mine
would be just as long!

On Wed, May 8, 2013 at 12:57 PM, Jörg Prante joergprante@gmail.com wrote:

Very early for christmas, but my (never ending) wish list is:

  • pluggable faceting framework: collations for sorting facet entries,
    histograms with equidistant and non-equidistant binning (uniform and
    non-uniform widths), and the longstanding "grouping/collapsing" feature
    (supervised clustering of docs)

  • Websockets as an alternative to REST HTTP, together with an event source
    on node level to pub/sub for ES nodex/index/shard events, and for a
    streaming API for "piping" JSON into ES bulk indexing

  • contineous datacenter to datacenter replication (Websockets might help
    to "tee" an index to multiple targets in the background, without discovery)

  • query estimator to block "evil"/"heavy" queries from executing

  • resource watchdog to detect low resource conditions and execute
    emergency node actions (throttle, admin mail, shutdown)

  • Lucene DocValues for fast key/value store (should be there already,
    haven't checked)

  • classifiers for terms/documents (unsupervised clustering of docs)

  • improving discovery algorithms (leader election)

  • lots of tooling OOTB for the lazy admin (web console, backup/restore
    tool, upgrade tool, hooks for deployment scripts)

... nuff' for ES 2.0, 3.0 and 4.0 if you ask me :slight_smile:

Jörg

Am 08.05.13 19:28, schrieb Ivan Brusic:

I would love to know the status of some longstanding requests such as:

)

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@**googlegroups.comelasticsearch%2Bunsubscribe@googlegroups.com
.
For more options, visit https://groups.google.com/**groups/opt_outhttps://groups.google.com/groups/opt_out
.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

+1 for basically everything Jorg mentions with an emphasis on the
faceting improvements, collapsing/grouping, and dc-to-dc replication.

Thanks,
Matt Weber

On Wed, May 8, 2013 at 1:56 PM, Ivan Brusic ivan@brusic.com wrote:

My list is based on previous discussions of issues that could potentially be
featured in future releases (1.0). If I had to compile a wish list, mine
would be just as long!

On Wed, May 8, 2013 at 12:57 PM, Jörg Prante joergprante@gmail.com wrote:

Very early for christmas, but my (never ending) wish list is:

  • pluggable faceting framework: collations for sorting facet entries,
    histograms with equidistant and non-equidistant binning (uniform and
    non-uniform widths), and the longstanding "grouping/collapsing" feature
    (supervised clustering of docs)

  • Websockets as an alternative to REST HTTP, together with an event source
    on node level to pub/sub for ES nodex/index/shard events, and for a
    streaming API for "piping" JSON into ES bulk indexing

  • contineous datacenter to datacenter replication (Websockets might help
    to "tee" an index to multiple targets in the background, without discovery)

  • query estimator to block "evil"/"heavy" queries from executing

  • resource watchdog to detect low resource conditions and execute
    emergency node actions (throttle, admin mail, shutdown)

  • Lucene DocValues for fast key/value store (should be there already,
    haven't checked)

  • classifiers for terms/documents (unsupervised clustering of docs)

  • improving discovery algorithms (leader election)

  • lots of tooling OOTB for the lazy admin (web console, backup/restore
    tool, upgrade tool, hooks for deployment scripts)

... nuff' for ES 2.0, 3.0 and 4.0 if you ask me :slight_smile:

Jörg

Am 08.05.13 19:28, schrieb Ivan Brusic:

I would love to know the status of some longstanding requests such as:

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.