In the study Elasticsearch time, because each version of the problem, not quite sure, very troublesome, the government has also given the circumstances of each version update, but the English version, the version update information and particularly, the recent study, read a lot of information, none of finishing very clear, and then himself under the consolidation, finishing first statement following are various versions personally think the more important point, because every major version updates the content too much, not one by one example, details need to refer to the official documentation , there are links to the bottom of the article, I am also for their own convenience on the whole, to understand Elasticsearch each version of the iteration, we can better understand and use Elasticsearch product, so with this article.
The initial version 0.7.0
May 14, 2010 released the first version can be sent to query version information and important features:
- Zen Discovery Auto discovery module
- Groovy Client Support
- Simple plug-in management system
- Better support ICU tokenizer
- More Management API
Version of the initialization, a temporary little introduction, let so much.
Upgrade version 1.0.0
February 14, 2014 release, important features: -Snapshot / Restore API API backup and recovery
- Support aggregation analysis Aggregations
- CAT API support
- Support Alliance inquiry
- Breaker support
- Doc values introduced
Version 2.0.0
October 28, 2015 release, important features:
- Increased pipleline Aggregations
- query / filter combined query, the query are combined in different queries depending on context
- Compression can be configured to store
- Rivers module is removed
- Multicast Multicast discovery is removed, as a plug-in, the production environment must be configured unicast address
New features version 5.0.0
October 26, 2016 release, important features:
- Lucene 6.x support, less than half the disk space; less than half the time index; query performance by 25%; support IPV6.
- Internal engine level removed to avoid concurrent updates to the same document lock competition, bringing a 15% -20% performance increase
- Shrink API, the number of fragments that can be contracted to its factor, you are 15 as previously fragments, you can shrink or 3 to 5 or 1 and, as we can imagine a scenario, in the writing pressure very large collection phase, more than enough to set the index, take advantage of parallel shard of writing ability, index shrunk to less shard after writing, improve query performance
- REST client SDK provides the first native Java
- IngestNode, if necessary before processing the data, are handled before the index, such as logstash may be structured and log conversion, it can be directly processed es
- It provides Painless script, instead of Groovy script
- Remove the site plugins, that is head, bigdesk can not be directly mounted inside es, but can be deployed independent sites (anyway, are static files) or plug-ins developed kibana
- New Sliced Scroll type Scroll interfaces now can be complicated for data traversed. Scroll each request, the request can be divided into a plurality of Slice, slices will be appreciated, each independently Slice parallel, or by using Scroll rebuild traverse many times faster.
- Added Profile API
- Added Rollover API
- New Reindex
- Providing a first native Java SDK REST client based on the client HTTP protocol dependent decouple Elasticsearch no conflict jar package, a cluster node automatic discovery, automatic poll request log processing node request fails, full high availability capability to play Elasticsearch
- Introducing a new field type Text / Keyword replaced String
- Index request size limit, to avoid a large number of concurrent requests crushed ES
- Limiting the number of shards single request, a default 1000
New features version 6.0.0
August 31, 2017 release, important features:
- Sparsity Doc Values of support
- Index sorting, that is the sort of indexing phase.
- Support sequence number, the operation of each es has a sequence number (incremental design similar)
- Seamless rolling upgrade
- Removal of types, which at 6.0, does not support multiple type began to exist inside a index
- Index-template inheritance, succession index version of the current index is the template for all matches will merge, this will cause the index template some conflict, will match only a 6.0, will be validated when the index is created
- Load aware shard routing, load-based request routing, the current node is a full polling search request, then the performance of the slowest node often leads to overall increase in the delay, the new implementation of the queue length will be automatically adjusted based on queue time-consuming, queue length high load node will be reduced, so that other nodes share more pressure, and the search index will be based on this mechanism.
- The index has closed will also support the replica of automatic processing to ensure reliable data.
New features version 7.0.0
April 10, 2019 release, important features:
- Cluster connectivity changes: TransportClient that is discarded, java es7 codes, use only restclient. Then, the individual integrated a little, for java programming, we recommend using High-level-rest-client operate in a manner ES cluster
- ES package Default package jdk: 7.x version of that package size 300MB + side suddenly found 6.x contrast, a large pack 200MB +, it is the size of JDK
- Lucene9.0
-
Major improvements - formally abolished under a single index to support multi-Type when es6, official mentions the es7 will delete type, and es6 time has been specified for each index can have only one type. _Doc as the default type in es7, the official said, completely removed type in the 8.x versions. api send request mode changes, such as to obtain a document ID for an index of: GET index / _doc / id and id is the index where a specific value
-
7.1 start, Security features free use
- ECK-ElasticSearch Operator is Kubernetes
- Breakers introduced real memory, it can be more accurately detected that the request can not be processed, and to prevent them from a single node so unstable
- Zen2 is Elasticsearch new cluster coordination layer, to improve the reliability, performance and user experience, faster, more secure, and easier to use
- new function
- New Cluster coordination
- Feature - Complete High Level REST Client
- Script Score Query
- Performance Optimization
- Weak-AND algorithm to improve query performance
- The default number of Primary Shared from 5 to 1, to avoid Over Sharding
- Faster top-k query
- Query interval (Intervals queries) certain search use cases (for example, legal and patent search) introduced a word or phrase to find a certain distance away from the recording of each other's needs. Interval of 7.0 Elasticsearch query introduces a new way of constructing such queries, compared with the previous method (span query span queries), the use and definition easier. Compared with the span query, the query interval adaptability to the edge of the case stronger.
to sum up
By iterative upgrade versions will find, Elasticsearch significant improvement of product experience, to understand the differences between versions, and you will raise a cognitive level, a large article online, sometimes you find that the success of the operation when the author of the article and to you here would have failed, Best not the mystery, or how one of my method or an object is gone, right or wrong, is not conclusive, is the key to understand the essence of things, back to the root of the problem, It is the root of the problem.
Hope Benpian'll get you on the road to learn Elasticsearch more smoothly, and so after you finished school Elasticsearch latest version, go back and look at this article, the feeling is not the same, I feel that learning a skill when my heart for all to have a cognitive profile, and will not got into a space, can not see the whole forest invalid embarrassing situation. Like this article title says, look at the whole forest, go study a lesson trees, will know better.
END
If the harvest, please help forward, there will be follow-up articles contributed better, your encouragement is the author of the greatest power!
I welcome the attention of the public number: architect practice, access to exclusive organize learning resources and push dry goods daily.
Reference article: