With JUnit up to 4.10 there's no way to distinguish between what's shown in IDEs and test ids that need to be unique. A test Description is identified by just a string, that needs to be unique and is shown by IDEs. IDEs have slightly different behaviours when it comes to showing tests and suites titles. Some IDE (e.g. IntelliJ) strips the description on '.', which is why the '.' was replaced in the first place, in order to obtain the same behaviour on all IDEs. On the other hand the information printed out by RestReproduceInfoPrinter was wrong as the file path contained a '_' instead of a '.', which made the string to reproduce a failure useless in some cases. At the end of the day it seems better to just keep the dots and accept slightly different behaviours that are IDE dependent.
If the thread pools of an elasticsearch node cannot be shutdown
immediately, a wait of 10 seconds is added. This clashes with the
RPM scripts, as by default the init functions wait for 3 seconds
for a service to shutdown before a KILL signal is sent, resulting
in an unclean shutdown - not from an elasticsearch point of view,
but from init system point of view, as some lock files are left
around.
In order to prevent this the init script as well as the systemd
configuration now feature the same timeout than the debian package,
which is 20 seconds.
The await statement, which causes the 10 second delay can be found in
InternalNode.close()
Previously, the cluster state before flushing was used to check which
types map the given types pattern. However, this state might not be
up to date. Instead use the more recent cluster state from clusterState.state()
This fixes a test failure of PercolatorTests.testDeletePercolatorType
Other changes:
- use BoolFilter instead of OrFilter, because it is faster
- throw exception immediately when no type matching the given patterns
was found in cluster state
The get field mapping API now includes a mappings element after the index in its JSON
Added more consistent endpoint /{index}/_mapping/{type}/field/{fields}
and added endpoint /_mapping/{type}/field/{fields}
which are also used in tests
Added rest spec tests for wildcards and _all
Relates #4071
NOTE: This is not yet complete for 1.0. We need to return an empty JSON document instead
of a 404 if the field of an existing index and type is not found. However this is not
possible with the current data structure being returned. Needs to be finished for 1.0.
* Made GET mappings consistent, supporting
* /{index}/_mappings/{type}
* /{index}/_mapping/{type}
* /_mapping/{type}
* Added "mappings" in the JSON response to align it with other responses
* Made GET warmers consistent, support /{index}/_warmers/{type} and /_warmer, /_warner/{name}
as well as wildcards and _all notation
* Made GET aliases consistent, support /{index}/_aliases/{name} and /_alias, /_aliases/{name}
as well as wildcards and _all notation
* Made GET settings consistent, added /{index}/_setting/{name}, /_settings/{name}
as well as supportings wildcards in settings name
* Returning empty JSON instead of a 404, if a specific warmer/
setting/alias/type is missing
* Added a ton of spec tests for all of the above
* Added a couple of more integration tests for several features
Relates #4071
See issue #4071
PUT options for _mapping:
Single type can now be added with
`[PUT|POST] {index|_all|*|regex|blank}/[_mapping|_mappings]/type`
and
`[PUT|POST] {index|_all|*|regex|blank}/type/[_mapping|_mappings]`
PUT options for _warmer:
PUT with a single warmer can now be done with
`[PUT|POST] {index|_all|*|prefix*|blank}/{type|_all|*|prefix*|blank}/[_warmer|_warmers]/warmer_name`
PUT options for _alias:
Single alias can now be PUT with
`[PUT|POST] {index|_all|*|prefix*|blank}/[_alias|_aliases]/alias`
DELETE options _mapping:
Several mappings can be deleted at once by defining several indices and types with
`[DELETE] /{index}/{type}`
`[DELETE] /{index}/{type}/_mapping`
`[DELETE] /{index}/_mapping/{type}`
where
`index= * | _all | glob pattern | name1, name2, …`
`type= * | _all | glob pattern | name1, name2, …`
Alternatively, the keyword `_mapings` can be used.
DELETE options for _warmer:
Several warmers can be deleted at once by defining several indices and names with
`[DELETE] /{index}/_warmer/{type}`
where
`index= * | _all | glob pattern | name1, name2, …`
`type= * | _all | glob pattern | name1, name2, …`
Alternatively, the keyword `_warmers` can be used.
DELETE options for _alias:
Several aliases can be deleted at once by defining several indices and names with
`[DELETE] /{index}/_alias/{type}`
where
`index= * | _all | glob pattern | name1, name2, …`
`type= * | _all | glob pattern | name1, name2, …`
Alternatively, the keyword `_aliases` can be used.
When excluding '*.f1' from `{ "obj": { "f1": 1, "f2": 2 } }` XContentMapValues.filter returns `{ "obj": { "f2": 2}}`. When run on `{ "obj": { "f1" : 1 }}` we should return `{ "obj": { }}` to maintain object structure. People currently need to always check whether `obj` is there or not.
Closes#4715Closes#4047
Related to #4491
Fixes#4701. Changes behavior of the snapshot operation. The operation now fails if not all primary shards are available at the beginning of the snapshot operation. The restore operation no longer tries to restore indices with shards that failed or were missing during snapshot operation.
Currently it is possible to index a document as:
```
POST /myindex/mytype/1
{ "foo"...}
```
or as:
```
POST /myindex/mytype/1
{
"mytype": {
"foo"...
}
}
```
This makes indexing non-deterministic and fields can be misinterpreted
as type names.
This changes makes Elasticsearch accept only the first form by default,
ie without the type wrapper. This can be changed by setting
`index.mapping.allow_type_wrapper` to `true`` when creating the index.
Closes#4484
correctly.
InterruptedExceptions should be handled by either rethrowing or
restoring the interrupt state (i.e. calling
`Thread.currentThread().interrupt()`). This is important since the
caller of the is method or subequent method calls might also be
interested in this exception. If we ignore the interrupt state the
caller might be left unaware of the exception and blocks again on
a subsequent method.
Closes#4712
The tribes feature allowed to create a tribe node that can act as a federated client across multiple clusters.
The tribe node configuration looks something like this:
```
tribe.t1.cluster.name: cluster1
tribe.t2.cluster.name: cluster2
```
The configuration above configure connections to 2 clusters, named `t1`, `t2`. It creates a "node" client to each (so by default, above, multicast discovery is used). The settings for each node client is extracted from the `tribe.[tribe_name]` prefix.
The way the tribe node works is by merging the cluster state from each cluster, and creating a merged view of all clusters. This means all operations work the same, distributed search, suggest, percolation, indexing, ... .
The merged view drops conflicted indices and picks one of them if there are 2 indices with the same name across multiple clusters.
By default, read and write operations are allowed. Master level read operations (cluster state for example), require setting the local flag to true (since there is no elected master). Master level write operations are not allowed (create index, ...).
The tribe node can be configured to block write operations `tribe.blocks.write` to `true`, and metadata operations by setting `tribe.blocks.metadata` to `true`.
closes#4708
When set to false a new strict mode of parsing is employed which
a) does not permit numbers to be passed as JSON strings in quotes
b) rejects numbers with fractions that are passed to integer, short or long fields.
Closes#4117
Put mapping now supports either of these formats:
POST foo/doc/_mapping
{
"doc": {
"_routing": {"required": true},
"properties": {
"body": {"type": "string"}
}
}
}
or
POST foo/doc/_mapping
{
"_routing": {"required": true},
"properties": {
"body": {"type": "string"}
}
}
Closes#4483
Java Builder apis drop old “len” methods in favour of new “length”
Rest APIs support both old “len: and new “length” forms using new ParseField class to a) provide compiler-checked consistency between Builder and Parser classes and
b) a common means of handling deprecated syntax in the DSL.
Documentation and rest specs only document the new “*length” forms
Closes#4083
% curl localhost:9200/_cat/indices\?h=health,index,rep,docs.count,ss,cs,fm\&v\&pri
health index rep docs.count ss pri.ss cs pri.cs fm pri.fm
green wiki 1 6401 302.9mb 151.4mb 0b 0b 25.1mb 0b
green twitter 1 11434 64mb 32mb 0b 0b 0b 0b
Closes#4689.
If the path_type is set to `just_name` in a multi_field typed field and that field is updated (for example another multi field is added) then if the path isn't specified again the path_type isn't taken into account and full path names are generated.
`standard_html_strip` and `pattern` analyzer support stopwords which are
set to the default `english` stopwords by default. Those analyzers
should not use stopwords by default since they are language neutral
Closes#4699
`min_doc_count` is the minimum number of hits that a term or histogram key
should match in order to appear in the response.
`min_doc_count=0` replaces `compute_empty_buckets` for histograms and will
behave exactly like facets' `all_terms=true` for terms aggregations.
Close#4662
When upgrading to ES 1.0 the existing mappings with a multi-field type automatically get replaced to a core field with the new `fields` option.
If a `multi_field` type-ed field doesn't have a main / default field, a default field will be chosen for the multi fields syntax. The new main field type
will be equal to the first `multi_field` fields' field or type string if no fields have been configured for the `multi_field` field and in both cases
the default index will not be indexed (`index=no` is set on the default field).
If a `multi_field` typed field has a default field, that field will replace the `multi_field` typed field.
Closes to #4521
No error wa previously thrown when a catch section was specified and a the rest call returned no error
Also, moved the catch:request on top to avoid even sending requests in that case
============
The default unit for measuring distances is *MILES* in most cases. This commit moves ES
over to the *International System of Units* and make it work on a default which relates
to *METERS* . Also the current structures of the `GeoBoundingBox Filter` changed in
order to define the *Bounding* by setting abitrary corners.
Distances
---------
Since the default unit for measuring distances has changed to a default unit
`DistanceUnit.DEFAULT` relating to *meters*, the **REST API** has changed at the
following places:
* `ScriptDocValues.factorDistance()` returns *meters* instead of *miles*
* `ScriptDocValues.factorDistanceWithDefault()` returns *meters* instead of *miles*
* `ScriptDocValues.arcDistance()` returns *meters* instead of *miles*
one might use `ScriptDocValues.arcDistanceInMiles()`
* `ScriptDocValues.arcDistanceWithDefault()` returns *meters* instead of *miles*
* `ScriptDocValues.distance()` returns *meters* instead of *miles*
one might use `ScriptDocValues.distanceInMiles()`
* `ScriptDocValues.distanceWithDefault()` returns *meters* instead of *miles*
one might use `ScriptDocValues.distanceInMilesWithDefault()`
* `GeoDistanceFilter` default unit changes from *kilometers* to *meters*
* `GeoDistanceRangeFilter` default unit changes from *miles* to *meters*
* `GeoDistanceFacet` default unit changes from *miles* to *meters*
Geo Bounding Box Filter
-----------------------
The naming of the GeoBoundingBoxFilter properties allows to set arbitrary corners
(see #4084) namely `top_right`, `top_left`, `bottom_right` and `bottom_left`. This
change also includes the fields `topRight` and `bottomLeft` Also it is be possible to
set the single values by using just `top`, `bottom`, `left` and `right` parameters.
Closes#4515, #4084
