This is an intial commit that splits HasChildQueryParser / Builder into
the two seperate steps. This one is particularly nasty since it transports
a pretty wild InnerHits object that needs heavy refactoring. Yet, this commit
has still some nocommits and needs more tests and maybe another cleanup but
it's a start to get the code out there.
Whe we call optimize we ignore Exceptions that indicate a closed shard.
However, when a shard is closed while an optimize request is in flight it
might also trigger an AlreadyClosedException from the IndexWriter when we
get the config or ForceMergeFailedEngineException with the EngineClosedException
wrapped inside. Because these are not identified as exceptions that indicate
a closed shard (TransportActions.isShardNotAvailableException(..)) optimize
would sometimes report failures when shards were relocating while optimize was called
and sometimes not. This caused weird test failures, see #13266 .
Instead, we should let EngineClosedException bubble up and also recognize
AlreadyClosedException as an indicator for a closed shard.
Today we try to allocate primaries first and then replicas
but don't take the index creation date and priority into account
as we do in the GatewayAlloactor.
Closes#13249
Adds a listeners to each of the caches that allows us to remove the dependency on IndexService which is cyclic since
the IndexService depends on both of these caches. This cyclic dependency makes
testing the indiviual parts very hard and is only added for the sake of
incrementing some stats.
Transport clients run embedded within external applications, so
elasticsearch should not be doing anything with the filesystem, as there
is not elasticsearch home.
This change makes a number of cleanups to the internal API for loading
settings and creating an environment. The loadFromConfig option was
removed, since it was always true except for tests. We now always
attempt to load settings from config a file when an environment is
created. The prepare methods were also simplified so there is now
prepareSettingsAndEnvironment which nodes use, and prepareSettings which
the transport client uses. I also attempted to improve the tests, but
there is a still a lot of follow up work to do there.
closes#13155
This commit removes and now forbids all uses of
com.google.common.base.Strings across the codebase.
For uses of com.google.common.base.Strings.isNullOrEmpty, use
org.elasticsearch.common.Strings.isNullOrEmpty.
For uses of com.google.common.base.Strings.padStart use
org.elasticsearch.common.Strings.padStart.
For uses of com.google.common.base.Strings.nullToEmpty use
org.elasticsearch.common.Strings.coalesceToEmpty.
Relates #13224
Before #13068 refresh and flush ignored all exceptions that matched
TransportActions.isShardNotAvailableException(e) and this should not change.
In addition, refresh and flush which are based on broadcast replication
might now get UnavailableShardsException from TransportReplicationAction if a shard
is unavailable and this is not caught by TransportActions.isShardNotAvailableException(e).
This must be ignored as well.
This commit removes and now forbids all uses of
com.google.common.base.Predicate and com.google.common.base.Predicates
across the codebase. This is one of the many steps in the eventual
removal of Guava as a dependency. This was enabled by #13314.
Relates #13224
This commit removes and now forbids all uses of
com.google.common.base.Objects across the codebase. This is a small
step in the eventual removal of Guava as a dependency.
Relates #13224
Previously we skip deleting the index store for indices on a shared
filesystem, because we don't want to delete the data when the shard is
relocating around the cluster. This adds a flag to the
`deleteIndexStore` method signifying that the index is closed and that
we should allow deleting the contents even if it is on a shared
filesystem.
Includes a unit test for the IndicesService.canDeleteIndexContents and
integration tests ensure a closed shadow replica index deletes files
correctly.
Resolves#13297
As a refinement to Project Coin (JEP-213, JDK-8042880), Java 9 is going
to disallow the use of ‘_’ as a one-character identifier. This will be
done by adding ‘_’ as a keyword to the Java language (JDK-8065599).
Currently, uses of ‘_’ as a one-character identifier are warnings in
the Java 8 compiler. This commit removes all uses of ‘_’ as a
one-character identifier from the codebase.
SpanContainingQueryParser and SpanWithinQueryParser always set the boost to the parsed lucene query, even if it is the default one. The default boost of the main query though is the boost coming from the inner little query, value that we end up overriding all the time. We should instead set the boost to the main query only if it differs from the default, to mimic lucene's behaviour.
Relates to #13272Closes#13339
SimpleQueryStringParser applies whatever boost the query holds, even if the default 1, to the query obtained from parsing of the query string. that might contain its boost, for instance if it resolved to a simple query like term (single term query against a single field). We should rather multiply the existing boost with the boost set to the query, same as we do in query_string
Relates to #13272Closes#13331
doParse() was supposed to allow aggs to perform extra parsing. Unfortunately, this forced the
parser to carry instance-level state, which would carry-over and "corrupt" any other aggs of the
same type in the same query.
Instead, we are now collecting all unknown params and pasing them as a Map<String, Object>
to buildFactory(). The agg may then parse them and instantiate a factory. Each param the
agg uses, it should unset from the unusedParams object.
After building the factory, the parser verifies that unusedParams is empty. If it is not empty,
an exception is raised so the user knows they provided unknown params.
Fixes#13337
This changes construction of Phrase and Boolean queries to use the builder,
and replaces BitDocIdSetFilter with BitSetProducer for nested and parent/child
queries. I had to remove the ParentIdsFilter for the case when there was a
single parent as it was using the source of BitSets for parents as a regular
Filter, which is not possible anymore now. I don't think this is an issue since
this case rarely occurs, and the alternative logic for when there are several
matching parent ids should not be much worse.
This pipeline will calculate percentiles over a set of sibling buckets. This is an exact
implementation, meaning it needs to cache a copy of the series in memory and sort it to determine
the percentiles.
This comes with a few limitations: to prevent serializing data around, only the requested percentiles
are calculated (unlike the TDigest version, which allows the java API to ask for any percentile).
It also needs to store the data in-memory, resulting in some overhead if the requested series is
very large.
This commit moves ignore_malformed and coerce options from the GeoPointFieldType to the Builder in GeoPointFieldMapper. This makes these options consistent with other types in 2.0.
This was supposed to just help the user, in case they misconfigured something.
Broadcast is an ipv4 only thing, the only way you can really detect its a broadcast
address, is to look and see if an interface has that address as its broadcast address.
But we cannot trust that container interfaces won't have a crazy setup...
Closes#13327
The match_phrase_prefix query properly parses the boost etc. but it loses it in its rewrite method. Fixed that by setting the orginal boost to the rewritten query before returning it. Also cleaned up some warning in MultiPhrasePrefixQuery.
Closes#13129Closes#13142
Target-type inference has been improved in Java 8. This leads to these
lines now being interpreted as invoking String#valueOf(char[]) whereas
they previously were interpreted as invoking String#valueOf(Object).
This change leads to ClassCastExceptions during test execution. Simply
casting the parameter to Object restores the old invocation.
Closes#13315
We have some optimization in FilteredQueryParser that tries to mimic what the rewrite method in lucene does, based on what gets parsed we return the simplest query possible. That might cause issues with boost values though, if specified in both the main query and the inner query that we shortcut to. We should rather rely on lucene's rewrite method to simplify the lucene representation of the query, and always build a filtered query instead.
relates to #13272Closes#13312
We currently optimize scroll when sort=_doc because docs are returned in order.
But documents are also returned in order when sorting by score and the query
gives constant scores. This optimization has the nice side-effect of also
optimizing scrolls with the default `match_all` query.
Until now we had a cloud-aws plugin which is providing 2 disctinct features:
* discovery on EC2
* snapshot/restore on S3
This commit splits the plugin by feature so people can use either one or the other or both features.
Doc is updated accordingly.
Before this change the check would check that all test classes end in Tests but the message would say they need to end in Test or Tests which was confusing.
Today we always collect in order to compute counts, but some of them can be
easily optimized by using pre-computed index statistics. This is especially
true in the case that there are no deletions, which should be common for the
time-based data use-case.
Counts on match_all queries can always be optimized, so requests like
```
GET index/_search?size=0
GET index/_search
{
"size": 0,
"query" : {
"match_all": {}
}
}
```
should now return almost instantly. Additionally, when there are no deletions,
term queries are also optimized, so the below queries which all boil down to a
single term query would also return almost immediately:
```
GET index/type/_search?size=0
GET index/_search
{
"size": 0,
"query" : {
"match": {
"foo": "bar"
}
}
}
GET index/_search
{
"size": 0,
"query" : {
"constant_score": {
"filter": {
"exists": {
"field": "foo"
}
}
}
}
}
```
Users might specify something like -Des.network.host=0.0.0.0, as that
was the old default with previous versions of elasticsearch. This means
to bind to all interfaces, but it makes no sense as a publish address.
Pick a good one in this case, just like we do in other cases where
publish isn't explicitly specified and we are bound to multiple (e.g.
when configured by interface, or dns hostname with multiple addresses).
However, in this case warn the user about it: since its arbitrarily
picking the first non-loopback address like the old versions
did, thats a little too heuristical, but lets make the cutover easy.
Separately, fail hard if things like multicast or broadcast addresses are
configured as bind or publish addresses, as that is simply invalid.
Closes#13274
The number and distribution of errors in some restore test may cause restore process to continue to fail for a prolong time. This test caps the total number of simulated failures to make sure that restore is guaranteed to eventually succeed after a limited number of retries.
We currently have a small number of test classes with the suffix "Test",
yet most use the suffix "Tests". This change renames all the "Test"
classes, so that we have a simple rule: "Non-inner classes ending with
Tests".
These are not actually tests, but command line applications that must be
run manually. This change removes the entire stresstest package. We can
add back individual tests that we find necessary, and make them real
tests (whether integ or not).
While the list of having exclusions is small, it shouldn't be necessary
at all. Base test cases should be suffixed with TestCase so they are not
picked up by the test class name pattern. This same rule works for
abstract classes as well.
This change renames abstract tests to use the TestCase suffix, adds a
check in naming convention tests, and removes the exclusion from our
test runner configuration. It also excludes inner classes (the only
exclude we should have IMO), so that we have no need to @Ignore the
inner test classes for naming convention tests.
In this test we assume that after waitForRelocation() has returned shards
are no more relocated and optimize will therefore succeed always.
However, because the test does not wait for green status, relocations can
still start after waitForRelocation() has returned successfully.
see #13266 for a detailed explanation
At the moment if an index script is used in a request, the spawned request to get the indexed script from the `.scripts` index does not get the headers and context copied to it from the original request. This change makes the calls to the `ScriptService` pass in a `HasContextAndHeaders` object that can provide the headers and context. For the `search()` method the context and headers are retrieved from `SearchContext.current()`.
Closes#12891
The shaded version of elasticsearch was built at the very beginning to avoid dependency conflicts in a specific case where:
* People use elasticsearch from Java
* People needs to embed elasticsearch jar within their own application (as it's today the only way to get a `TransportClient`)
* People also embed in their application another (most of the time older) version of dependency we are using for elasticsearch, such as: Guava, Joda, Jackson...
This conflict issue can be solved within the projects themselves by either upgrade the dependency version and use the one provided by elasticsearch or by shading elasticsearch project and relocating some conflicting packages.
Example
-------
As an example, let's say you want to use within your project `Joda 2.1` but elasticsearch `2.0.0-beta1` provides `Joda 2.8`.
Let's say you also want to run all that with shield plugin.
Create a new maven project or module with:
```xml
<groupId>fr.pilato.elasticsearch.test</groupId>
<artifactId>es-shaded</artifactId>
<version>1.0-SNAPSHOT</version>
<properties>
<elasticsearch.version>2.0.0-beta1</elasticsearch.version>
</properties>
<dependencies>
<dependency>
<groupId>org.elasticsearch</groupId>
<artifactId>elasticsearch</artifactId>
<version>${elasticsearch.version}</version>
</dependency>
<dependency>
<groupId>org.elasticsearch.plugin</groupId>
<artifactId>shield</artifactId>
<version>${elasticsearch.version}</version>
</dependency>
</dependencies>
```
And now shade and relocate all packages which conflicts with your own application:
```xml
<build>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-shade-plugin</artifactId>
<version>2.4.1</version>
<executions>
<execution>
<phase>package</phase>
<goals>
<goal>shade</goal>
</goals>
<configuration>
<relocations>
<relocation>
<pattern>org.joda</pattern>
<shadedPattern>fr.pilato.thirdparty.joda</shadedPattern>
</relocation>
</relocations>
</configuration>
</execution>
</executions>
</plugin>
</plugins>
</build>
```
You can create now a shaded version of elasticsearch + shield by running `mvn clean install`.
In your project, you can now depend on:
```xml
<dependency>
<groupId>fr.pilato.elasticsearch.test</groupId>
<artifactId>es-shaded</artifactId>
<version>1.0-SNAPSHOT</version>
</dependency>
<dependency>
<groupId>joda-time</groupId>
<artifactId>joda-time</artifactId>
<version>2.1</version>
</dependency>
```
Build then your TransportClient as usual:
```java
TransportClient client = TransportClient.builder()
.settings(Settings.builder()
.put("path.home", ".")
.put("shield.user", "username:password")
.put("plugin.types", "org.elasticsearch.shield.ShieldPlugin")
)
.build();
client.addTransportAddress(new InetSocketTransportAddress(new InetSocketAddress("localhost", 9300)));
// Index some data
client.prepareIndex("test", "doc", "1").setSource("foo", "bar").setRefresh(true).get();
SearchResponse searchResponse = client.prepareSearch("test").get();
```
If you want to use your own version of Joda, then import for example `org.joda.time.DateTime`. If you want to access to the shaded version (not recommended though), import `fr.pilato.thirdparty.joda.time.DateTime`.
You can run a simple test to make sure that both classes can live together within the same JVM:
```java
CodeSource codeSource = new org.joda.time.DateTime().getClass().getProtectionDomain().getCodeSource();
System.out.println("unshaded = " + codeSource);
codeSource = new fr.pilato.thirdparty.joda.time.DateTime().getClass().getProtectionDomain().getCodeSource();
System.out.println("shaded = " + codeSource);
```
It will print:
```
unshaded = (file:/path/to/joda-time-2.1.jar <no signer certificates>)
shaded = (file:/path/to/es-shaded-1.0-SNAPSHOT.jar <no signer certificates>)
```
This PR also removes fully-loaded module.
By the way, the project can now build with Maven 3.3.3 so we can relax a bit our maven policy.
Currently, we do not allow reads on shards which are in POST_RECOVERY which
unfortunately can cause search failures on shards which just recovered if there no replicas (#9421).
The reason why we did not allow reads on shards that are in POST_RECOVERY is
that after relocating a shard might miss a refresh if the node that executed the
refresh is behind with cluster state processing. If that happens, a user might execute
index/refresh/search but still not find the document that was indexed.
We changed how refresh works now in #13068 to make sure that shards cannot miss a refresh this
way by sending refresh requests the same way that we send write requests.
This commit changes IndexShard to allow reads on POST_RECOVERY now.
In addition it adds two test:
- test for issue #9421 (After relocation shards might temporarily not be searchable if still in POST_RECOVERY)
- test for visibility issue with relocation and refresh if reads allowed when shard is in POST_RECOVERY
closes#9421
DateHistogramIT has suite scope and therefore must take care of cleaning up after
each test. Otherwise we cannot run tests in several times with tests.iters.
The implementation this commit replaces was almost k-NN regression with
k=2, but had two bugs: (a) it depends on the empirical raw estimates
being in strictly non-decreasing order for the binary search (which they
are not); and (b) it weights the biases positively with increased
distance from the corresponding raw estimate.
“HyperLogLog in Practice” leaves the choice of exact algorithm here
fairly vague, just noting: “We use k-nearest neighbor interpolation to
get the bias for a given raw estimate (for k = 6).” The majority of
other open source HyperLogLog++ implementations appear to use k-NN
regression with uniform weights (and generally k = 6). Uniform
weighting does decrease variance, but also introduces bias at the domain
extrema. This problem, plus the use of the word “interpolation” in the
original paper, suggests (inverse) distance-weighted k-NN, as
implemented here.
Since #13068 refresh and flush requests go to the primary first and are then replicated.
One difference to before is though that if a shard is not available (INITIALIZING for example)
we wait a little for an indexing request but for refresh we don't and just give up immediately.
Before, refresh requests were just send to the shards regardless of what their state is.
In tests we sometimes create an index, issue an indexing request, refresh and
then get the document. But we do not wait until all nodes know that all primaries have ben assigned.
Now potentially one node can be one cluster state behind and not know yet that
the shards have ben started. If the refresh is executed through this node then the
refresh request will silently fail on shards that are started already because from
the nodes perspective they are still initializing. As a consequence, documents
that expected to be available in the test are now not.
Example test failures are here: http://build-us-00.elastic.co/job/elasticsearch-20-oracle-jdk7/395/
This commit changes the timeout to 1m (default) to make sure we don't miss shards
when we refresh. This will trigger the same retry mechanism as for indexing requests.
We still have to make a decision if this change of behavior is acceptable.
see #13238
From a user perspective, the main benefit from this upgrade is that the new
Lucene53Codec has disk-based norms. The elasticsearch directory has been fixed
to load these norms through mmap instead of nio.
Other changes include the removal of `max_thread_states`, the fact that
PhraseQuery and BooleanQuery are now immutable, and that deleted docs are now
applied on top of the Scorer API.
This change introduces a couple of `AwaitsFix`s but I don't think it should
hold us from merging.
This allows reducing privileges with doPrivileged to work,
otherwise it will fail with NPE.
In general, if some code wants to do that, let it. The null
check is needed, even though ProtectionDomain(CodeSource, PermissionCollection)
is more than a bit misleading: "the current Policy will not be consulted".
Additionally add a defensive check for location, since the docs
there are even more confusing: https://bugs.openjdk.java.net/browse/JDK-8129972
The jdk policy impl has both these checks.
We used to mutate the map of fields and weights while transforming to lucene query (toQuery method), by adding the default field if the map is empty. That is an anti-pattern, we should leave the original query as is instead, but still generate the proper lucene query. Also took the chance to fix a couple of issues in SimpleQueryStringBuilderTest#doAssertLuceneQuery.
Several shard-level operations that previously broadcasted a request
per shard were converted to broadcast a request per node. This commit
converts upgrade action to this new model as well.
Closes#13204
Switch to use HTTPS by default for all hardcoded plugin URLs.
If users want to install via HTTP they can still specify a HTTP
URL manually.
Closes#12748
The current netty multiport tests bind on localhost and then try to connect
to 127.0.0.1, which may fail, if localhost is resolved to ipv6 by default.
This randomly chooses between 127.0.0.1, localhost and ::1 (if available) for
binding and then uses this throughout the test.
The path that a shard is allocated on is not taken into account when
we decide to move a shard away from a node because it passed a watermark.
Even worse we potentially moved away (relocated) a shard that was not even
allocated on that disk but on another on the node in question. This commit
adds a ShardRouting -> dataPath mapping to ClusterInfo that allows to identify
on which disk the shards are allocated on.
Relates to #13106
Added some comments about changes that we might be able to make later on in the refactoring. Also exposed handleTermsLookup in the context directly, removed unused similarityService getter and corrected simpleMatchToIndexNames call according to changes happened on master (the method variant that accept types will go away, types are ignored anyway).
The reason behind this change is that we will soon remove the index member from QueryParseContext, as parsing will be independent of the index it happens against.
DummyQueryParser properly implements now fromXContent and the corresponding builder supports converting to the corresponding lucene query through toQuery method.
This allows us to remove the string argument from the parseQuery method and the need for removing the first few tokens from the query, as the selected parser knows only about his portion of dsl.
The field type tests for mappings had a huge hole: check compatibility
was not tested directly at all! I had meant for this to happen in a
follow up after #8871, and was relying on existing mapping tests.
However, there were a number of issues.
This change reworks the fieldtype tests to be able to check all settable
properties on a field type work with checkCompatibility. It fixes a
handful of small bugs in various field types. In particular, analyzer
comparison was just wrong: it was comparing reference equality for
search analyzer instead of the analyzer name. There was also no check
for search quote analyzer.
closes#13112
```sh
bin/plugin install lmenezes/elasticsearch-kopf/develop
-> Installing lmenezes/elasticsearch-kopf/develop...
Trying http://download.elastic.co/lmenezes/elasticsearch-kopf/elasticsearch-kopf-develop.zip ...
Trying http://search.maven.org/remotecontent?filepath=lmenezes/elasticsearch-kopf/develop/elasticsearch-kopf-develop.zip ...
Trying https://oss.sonatype.org/service/local/repositories/releases/content/lmenezes/elasticsearch-kopf/develop/elasticsearch-kopf-develop.zip ...
Trying https://github.com/lmenezes/elasticsearch-kopf/archive/develop.zip ...
Downloading .................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................DONE
Verifying https://github.com/lmenezes/elasticsearch-kopf/archive/develop.zip checksums if available ...
Trying https://github.com/lmenezes/elasticsearch-kopf/archive/master.zip ...
Downloading ....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................DONE
Verifying https://github.com/lmenezes/elasticsearch-kopf/archive/master.zip checksums if available ...
```
This happens because we don't have anymore ElasticsearchWrapperException here but standard java exceptions.
Closes#13196.
Currently, many shard-level operations are transported with a request
per shard via TransportBroadcastAction. These shard-level requests are
then submitted to unbounded execution queues for asynchronous execution
on the receiving node. This transport mechanism and stuffing of the
execution queues can be problematic on large clusters. A better
mechanism would be to aggregate the shard-level requests, transport
them via a single request per node, and execute the shard-level
operations serially on the receiving node.
This commit introduces TransportNodeBroadcastAction which is the
high-level mechanism for transporting the shard-level operations in a
single request per node. The shard-level operations are executed
serially on the receiving node and per-node shard-level results are
aggregated into a single response per node. These node-level results
are then aggregated into a single response to the initial request.
One item of note is a new mechanism for registering request handlers.
This mechanism enables registrants to provide a callback for
instantiating new instances of the request class. Doing this enables
the inner class to be instantiated with the context of its outer class.
This is done so that a single NodeRequest class can be defined rather
than defining a class per operation.
Closes#7990
Today we sum up the disk usage for the allocation decider which is broken since
we don't stripe across multiple data paths. Each shard has it's own private path
now but the allocation deciders still treat all paths as one big disk. This commit
adds allows allocation deciders to access the least used and most used path to make
better allocation decidsions upon canRemain and canAllocate calls.
Yet, this commit doesn't fix all the issues since we still can't tell which shard
can remain and which can't. This problem is out of scope in this commit and will be solved
in a followup commit.
Relates to #13106
Previously multiple clusters in the same JVM reused the same port ranges, leading to potential big gaps in port selection, which in turns causes unicast based discovery to fail, missing to find another node in the default 5 port range.
Also the previous logic had http use a range that is assigned to another JVMs.
There are some leftovers around usage of NamedWriteableRegistry in our branch compared to master, just because the registry was added first here, then backported to master, and something didn't go completely right during the final merge.
We default the value to be 20x the value of a ping timeout, however we only use the legacy ping timeout settings value for the calculation.
Closes#13162
This commit removes and now forbids all uses of
com.google.common.collect.Lists across the codebase. This is the first
of many steps in the eventual removal of Guava as a dependency.
lessThanOrEqualTo is more appropriate when comparing _ttl than lessThan
because in rare cases, when tests run very fast, the ttl you fetch will
still equal the one you sent.
Some listeners may need to do work before a shard's path is
accessed (such as creating the directory in a plugin), so the listener
should be called before anything happens (as its name implies).
detect_noop is pretty cheap and noop updates compartively expensive so this
feels like a sensible default.
Also had to do some testing and documentation around how _ttl works with
detect_noop.
Closes#11282
As discussed in #11744 this is the last step to unify parsing of boost and _name. Those fields are supported only in long version of queries, while we sometimes parse them when wwe shouldn't, inconsistently.
Closes#11744Closes#12966
Since we now don't stripe shards across data paths we need a way
to access the information on which path a shard is allocated to
eventually do better allocation decisions based on disk usage etc.
This commit exposes the shard paths as part of the shard stats.
Relates to #13106
Also clean up static method in MoreLikeThisFetchService, fix randomly failing TermsQueryBuilderTest#testNullValues and rename mappedFieldNamesSmall to MAPPED_LEAF_FIELD_NAMES.
Until a couple of hours ago we expected the position_offset_gap to default
to 0 in 2.0 and 100 in 2.1. We decided it was worth backporting that new
default to 2.0. So now that its backported we need to teach 2.1 that 2.0
also defaults to 100.
Closes#7268
This is much more fiddly than you'd expect it to be because of the way
position_offset_gap is applied in StringFieldMapper. Instead of setting
the default to 100 its simpler to make sure that all the analyzers default
to 100 and that StringFieldMapper doesn't override the default unless the
user specifies something different. Unless the index was created before
2.1, in which case the old default of 0 has to take.
Also postition_offset_gaps less than 0 aren't allowed at all.
New tests test that:
1. the new default doesn't match phrases across values with reasonably low
slop (5)
2. the new default doest match phrases across values with reasonably high
slop (50)
3. you can override the value and phrases work as you'd expect
4. if you leave the value undefined in the mapping and define it on a
custom analyzer the the value from the custom analyzer shines through
Closes#7268
This commit changes the startup behavior of Elasticsearch to throw an
exception if duplicate settings keys are detected in the Elasticsearch
configuration file.
Closes#13079
Currently when an entire type is disabled, our document parser will end
parsing on the first field of the document. This blows up the recently
added check that parsing did not silently skip any tokens (ie whether
there was garbage leftover).
This change fixes the parser to correctly skip the entire document when
the type is disabled.
closes#13017
This commit consolidates the logic in
RoutingTable#allActiveShardsGrouped and
RoutingTable#allAssignedShardsGrouped into a single method that merely
applies a predicate to each ShardRouting.
Closes#13081
The `rewrite` option has been removed from the parser with
commit da5fa6c4 and won't parse anymore, however we still
have a setter for it in the builder that gets rendered out
when used and potentially leads to parsing errors. This PR
removes the setter for the unsupported `rewrite` option.
The `rewrite` option has been removed from the parser with
commit da5fa6c4 and won't parse anymore, however we still
have a setter for it in the builder that gets rendered out
when used and potentially leads to parsing errors. This PR
removes the setter for the unsupported `rewrite` option.
This change makes modules added by plugins come before others, as it was
before #12783. The order of configuration, and thereby binding, happens
in the order modules are received, and without this change, some plugins
can get *insane* guice errors (500mb stack trace).
The setting `plugin.types` is currently used to load plugins from the
classpath. This is necessary in tests, as well as the transport client.
This change removes the setting, and replaces it with the ability to
directly add plugins when building a transport client, as well as
infrastructure in the integration tests to specify which plugin classes
should be loaded on each node.
* makes most classes final and package private
* removes duplicate and confusing multiple entry points
* adds javadocs to some classes like JarHell,Security
* adds a public class BootStrapInfo that exposes any stats
needed by outside code.
When we create a plugin's classloader, we should allow it to register things in
the lucene SPI (registry of Tokenizers, TokenFilters, CharFilters, Codec,
PostingsFormat, DocValuesFormat).
Plugins should be able to do this so they can extend Lucene.
We do check if a directory is present and then open a dir stream on
it. Yet the file can be concurrrently deleted which is OK but we fail
with a hard exception. This change tries to open the dir directly (listing via stream)
and catches NoSuchFileEx | FNFEx.
This commit enforces that at most a single settings file is found. If
multiple settings files are found, a SettingsException will be thrown
Closes#13042
This commit fixes an issue that was causing Elasticsearch to silently
ignore settings files that contain garbage. The underlying issue was
swallowing an SettingsException under the assumption that the only
reason that an exception could be throw was due to the settings file
not existing (in this case the IOException would be the cause of the
swallowed SettingsException). This assumption is mistaken as an
IOException could also be thrown due to an access error or a read
error. Additionally, a SettingsException could be thrown exactly
because garbage was found in the settings file. We should instead
explicitly check that the settings file exists, and bomb on an
exception thrown for any reason.
Closes#13028
We currently test that our query parsers can parse the format that our query builder outputs in XContent format, but in some cases the parser supports more than that, hence we need more specific tests otherwise we have no coverage for alternate formats.
Today we are very verbose when rendering exceptions on the rest layer.
Yet, this isn't necessarily very easy to read and way too much infromation most
of the time. This change suppresses the stacktrace rendering by default but instead
adds a `rest.suppressed` logger that logs the suppressed stacktrace or rather the entire
exception on the node that renderes the exception
The log message looks like this:
```
[2015-08-19 16:21:58,427][INFO ][rest.suppressed ] /test/_search/ Params: {index=test}
[test] IndexNotFoundException[no such index]
at org.elasticsearch.cluster.metadata.IndexNameExpressionResolver$WildcardExpressionResolver.resolve(IndexNameExpressionResolver.java:551)
```
Currently we implicitly enforce a version format on the java.version
property for plugins via JarHell.checkJavaVersion. We should explicitly
enforce this version format and specify it in the documentation.
This commit adds an explicit enforcement of the version format on the
java.version property for plugins and updates the documentation for
plugin-descriptor.properties accordingly.
Closes#13009
This commit improves Java version comparison in JarHell.
The first improvement is the addition of a method to check the version
format of a target version string. This method will reject target
version strings that are not a sequence of nonnegative decimal integers
separated by “.”s, possibly with leading zeros (0*[0-9]+(\.[0-9]+)?).
This version format is the version format used for Java specification
versioning (cf. Java Product Versioning, 1.5.1 Specification Versioning
and the Javadocs for java.lang.Package.)
The second improvement is a clean method for checking that a target
version is compatible with the runtime version of the JVM. This is done
using the system property java.specification.version and comparing the
versions lexicograpically. This method of comparison has been tested on
JDK 9 builds that include JEP-220 (the Project Jigsaw JEP concerning
modular runtime images) and JEP-223 (the version string JEP). The class
that encapsulates the methods for parsing and comparing versions is
written in a way that can easily be converted to use the Version class
from JEP-223 if that class is ultimately incorporated into mainline JDK
9.
Closes#12441
This commit changes the behavior when validating index templates to
accumulate all validation errors before reporting failure to the user.
This addresses a usability issue when creating index templates.
Closes#12900
Aggregation.subPath() always threw an ArrayStoreException because we were trying to pass a List into System.arraycopy(). This change fixes that bug and adds a test to prevent regression
Today we only guess how big the shard will be that we are allocating on a node.
Yet, we have this information on the master but it's not available on the data nodes
when we pick a data path for the shard. We use some rather simple heuristic based on
existing shard sizes on this node which might be complete bogus. This change adds
the expected shard size to the ShardRouting for RELOCATING and INITIALIZING shards
to be used on the actual node to find the best data path for the shard.
Closes#11271
Try to strike a balance between usability and debuggability.
2MB stacktrace is not gonna work... we print message and find
"likely root cause" and identify it as such, limit it to 30 stack
frames, and filter out guice. When frames are truncated we identify
that too, and always inform the user full exception is available from
logs.
Closes#13029
Multicast has known issues (see #12999 and #12993). This change moves
multicast into a plugin, and deprecates it in the docs. It also allows
for plugging in multiple zen ping implementations.
closes#13019
If the machine doesn't support IPv6, or if the user disabled it
with -Djava.net.preferIPv4Stack, or if the user disabled it with
ES_USE_IPV4, we will still try to bind a socket to ::1 out of box,
and we will see lots of exceptions logged to the console.
It is harmless noise but not a great experience.
this commit removes all support for reverse host name resolving from
InetSocketTransportAddress. This class now only returns IP addresses.
Closes#13014
Fix unicast discovery to work when a host has multiple addresses.
Ban dangerous methods in java.net with forbidden APIs.
Fix ipv6 bugs and formatting of network addresses everywhere.
Closes#12999Closes#12993
Squashed commit of the following:
commit 6c1aa001d091c5cf25212a53dc701fb704337f1e
Author: Robert Muir <rmuir@apache.org>
Date: Thu Aug 20 14:25:43 2015 -0400
Fix these to be correct with addresses just in case
commit 648215627e84abf58a71400e7dc9ae775efb71d6
Merge: d00561b 41d8fbe
Author: Robert Muir <rmuir@apache.org>
Date: Thu Aug 20 13:23:09 2015 -0400
Merge branch 'master' into unicast_all_the_way_down
commit d00561b76fd1aa5850699f7901f3dae3d4d402b7
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Aug 20 16:38:50 2015 +0200
limit local ports to 5 in UnicastZenPing
commit e2e15c594006746cbe24432694294a71cc99deb8
Author: Robert Muir <rmuir@apache.org>
Date: Thu Aug 20 10:32:47 2015 -0400
fix port limiting
commit 10153cb7adadda81a1f482445e703836b65cf5e2
Author: Robert Muir <rmuir@apache.org>
Date: Thu Aug 20 10:18:37 2015 -0400
don't serialize scopeids: that's broken
commit 2aa63d43db2baec68a2e9bc227cfeb85dfeb4f83
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Aug 20 16:06:51 2015 +0200
restore @Network
commit c840f1d1ef438826ae1ecfd5e45942a0e30dc9c0
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Aug 20 16:02:30 2015 +0200
Use NetworkAddress.formatAddress where applicable in plugins
commit 374ce878852b35d626b7a29c8c4773545b0e9ddd
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Aug 20 15:34:06 2015 +0200
Use NetworkAddress.formatAddress where applicable
commit e7a606d63f1bc43c1b62b6e17adf707c76d43a15
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Aug 20 10:17:57 2015 +0200
Add @Multicast annotation to disable multicast tests by default.
We only run multicast tests now when we explicitly state it. A working
multicast env is required which is not always the case.
commit 2d7d2d0347179696ab41f71f048b13305014c85b
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Aug 20 09:51:28 2015 +0200
Remove extra check for local mode in InternalTestCluster
commit dda59ac39aa136d4687b9274c2692cd77f8b8f66
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Aug 20 09:37:03 2015 +0200
Handle node mode across entire test cluster
We used static methods reading sys properties to define the node mode
per cluster. this had lots of problems when tests couldn't cope with
mixed or only local mode. Now we are passing it down to the cluster from the test
which allows to @SuppressNetworkMode / @SupressLocalMode on the test to force
consistent node configurations.
commit 058197b7a408318995c88ce7f6762e32348de0de
Author: Robert Muir <rmuir@apache.org>
Date: Thu Aug 20 03:19:14 2015 -0400
really ban InetSocketAddress's trappy method and break build and go to sleep, sorry
commit ac8779185aee1e17e6f5a81766290fdfc9c603ba
Author: Robert Muir <rmuir@apache.org>
Date: Thu Aug 20 03:16:52 2015 -0400
Ban methods that might surprisingly cause DNS lookups
commit e64fe3dff2b11503e5f2831eb9863d64f56c5538
Author: Robert Muir <rmuir@apache.org>
Date: Thu Aug 20 02:59:05 2015 -0400
Add unit test
commit f15434f20fb1a3691b1cc16028597d8fae937e05
Author: Robert Muir <rmuir@apache.org>
Date: Thu Aug 20 02:39:02 2015 -0400
fix ipv6 formatting bugs
commit 05c2c74098052c75fbb79ea1818a295ef2e03e30
Author: Robert Muir <rmuir@apache.org>
Date: Thu Aug 20 02:12:05 2015 -0400
format addresses correctly so I can actually read what comes out of our logs and stats apis
commit 4f9389dcf1e8925f23153c5eb271b4ce2294dbaf
Author: Robert Muir <rmuir@apache.org>
Date: Wed Aug 19 21:26:52 2015 -0400
ban dangerous methods in java.net
commit 6aacd4d9925f324903d1d099a6cf5f862aeaf677
Author: Robert Muir <rmuir@apache.org>
Date: Wed Aug 19 20:59:24 2015 -0400
ban lenient method
commit f466a842c60163d1f4554bdce8a4163edb534c2c
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Aug 20 00:29:00 2015 +0200
fix tests to not mix local transport and zen unicast disco
commit 0de007a33b33fb68cf85cd86db4ca4f8ce10bbc9
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Aug 20 00:10:07 2015 +0200
fix tests to not mix local transport and zen unicast disco
commit 539f6ca6e5137e0d496239adc8684688dedcc824
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Aug 20 00:02:01 2015 +0200
fix tests to not mix local transport and zen unicast disco
commit 004c2881b25467f332acc8c9f9e92b1f0f9d314e
Author: Robert Muir <rmuir@apache.org>
Date: Wed Aug 19 17:51:45 2015 -0400
Fix multinode
commit 54113af325ce31571811c49fdaae89d5687be4ba
Author: Robert Muir <rmuir@apache.org>
Date: Wed Aug 19 17:36:45 2015 -0400
fix integration tests
commit 0156a77a56319d6b9737ec6a531992052e50bd59
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Aug 19 23:32:18 2015 +0200
enable multicast in MulticastZenPingIT.java
commit 1791caa35da853ce0122485fa3fd4674c671ec6e
Author: Robert Muir <rmuir@apache.org>
Date: Wed Aug 19 17:23:16 2015 -0400
Fix constant
commit 22820b53e0b2dc9fd47145c2bc29ce912a8fd484
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Aug 19 22:59:09 2015 +0200
give it some extra ids for local transport crazyness
commit b2138fafa94a8a085813fd48356df63e57ade5b3
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Aug 19 22:51:42 2015 +0200
pass on local addresses from configured transport rather than hard code IP addresses
commit 1bf5de1f457b081e0ce262b57d2b55d39c434156
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Aug 19 22:04:31 2015 +0200
fix PluggableTransportModuleIT.java to use local disco and detach port limit for node local disco
commit b6706eddfa04c43947c16551359ae98a463d34aa
Author: Robert Muir <rmuir@apache.org>
Date: Wed Aug 19 14:16:03 2015 -0400
Default to unicast discovery, with default host list of 127.0.0.1, [::1]
In some cases this may contain duplicates, although its a misconfiguration,
lets not bind to multiple ports. Its no problem for us to dedup, this code
doesn't need to be huper-duper fast since its used only for logic around bind/publish
The cardinality aggregation is a metric aggregation and therefore cannot accept sub-aggregations. It was previously possible to create a rest request with a cardinality aggregation that had sub-aggregations. Now such a request will throw an error in the response.
Close#12988
We have some settings that prevent host name resolution which should
be repected by InetSocketTransportAddress#getHost() to only resolve if
allowed or desired.
At the moment, when installing from an url, a user provides the plugin name on
the command line like:
* bin/plugin install [plugin-name] --url [url]
This can lead to problems when picking an already existing name from another
plugin, and can potentially overwrite plugins already installed with that name.
This, this PR introduces a mandatory `name` property to the plugin descriptor
file which replaces the name formerly provided by the user.
With the addition of the `name` property to the plugin descriptor file, the user
does not need to specify the plugin name any longer when installing from a file
or url. Because of this, all arguments to `plugin install` command are now
either treated as a symbolic name, a URL or a file without the need to specify
this with an explicit option.
The new syntax for `plugin install` is now:
bin/plugin install [name or url]
* downloads official plugin
bin/plugin install analysis-kuromoji
* downloads github plugin
bin/plugin install lmenezes/elasticsearch-kopf
* install from URL or file
bin/plugin install http://link.to/foo.zip
bin/plugin install file:/path/to/foo.zip
If the argument does not parse to a valid URL, it is assumed to be a name and the
download location is resolved like before. Regardless of the source location of
the plugin, it is extracted to a temporary directory and the `name` property from
the descriptor file is used to determine the final install location.
Relates to #12715
it's an important exception to serialize and we see it often in tests
etc. but then it's wrapped in NotSerializableExceptionWrapper which is
odd. This commit adds native support for this exception.
The cluster configuration allows to setup a cluster for use with unicast discovery. This means that nodes have to have pre-calculated known addresses which can be used to poplulate the unicast hosts setting. Despite of repeated attempts to select unused ports, we still see test failures where the node can not bind to it's assigned port due to it already being in use (most on CentOS). This commit changes it to allow each node to have a pre-set mutual exclusive range of ports and add all those ports to the unicast hosts list. That's OK because we know the node will only bind to one of those.
We are in the process of getting rid of guava, and this removes a major
use. The replacement is mostly Collections.emptyList(), Arrays.asList
and Collections.unmodifiableList. While it is questionable whether we
need the last one (as these are usually placed in final members), we can
continue to refactor later by removing unnecessary wrappings.
Dynamic date fields mapped from dates of the form "yyyy-MM-dd"
automatically receive the millisecond paresr epoch_millis as an
alternative parsing format. However, dynamic date fields mapped from
dates of the form "yyyy/MM/dd" do not. This is a bug since the migration
documentation currently specifies that a dynamically added date field,
by default, includes the epoch_millis format. This commit adds
epoch_millis as an alternative parser to dynamic date fields mapped from
dates of the form "yyyy/MM/dd".
Closes#12873
commons-lang really is only used by some core classes to join strings or modiy arrays.
It's not worth carrying the dependency. This commit removes the dependency on commons-lang
entirely.
This causes a FileSystemException when trying to retrieve FileStore for a path,
and falsely returns false for Files.isWritable
Squashed commit of the following:
commit d2cc0d966f3bc94aa836b316a42b3c5724bc01ef
Author: Robert Muir <rmuir@apache.org>
Date: Tue Aug 18 15:49:48 2015 -0400
fixes for the non-bogus comments
commit 6e0a272f5f8ef7358654ded8ff4ffc31831fa5c7
Author: Robert Muir <rmuir@apache.org>
Date: Tue Aug 18 15:30:43 2015 -0400
Fix isWritable too
commit 2a8764ca118fc4c950bfc60d0b97de873e0e82ad
Author: Robert Muir <rmuir@apache.org>
Date: Tue Aug 18 14:49:50 2015 -0400
try to workaround filestore bug
This brings the Lucene query assertions back to the querybuilder test that were
removed in the ancient past when we tested Lucene queries through their
inherent equals method. As we no longer do that it makes sense to do at least
coarse sanity checking on the generated Lucene query. More such checks are
being added as part of this commit.
Relates to #10217
this method is very confusing and if it's used it's likely the wrong thing
with respect to the actual bound / published address. This change discourages
it's use and removes all useage. It's replaced with the actual published address
most of the time.
With #12921 we refactored IndicesModule but we forgot to make sure we create IndicesQueriesRegistry once. IndicesQueriesModule used to do `bind(IndicesQueriesRegistry.class).asEagerSingleton();` otherwise we get multiple instances of the registry. This needs to be ported do the IndicesModule.
It was already disabled, but during tests the test framework enabled the query cache by setting the static default query cache. The caching behaviour should be the same in production and in tests.
This move the `murmur3` field to the `mapper-murmur3` plugin and fixes its
defaults so that values will not be indexed by default, as the only purpose
of this field is to speed up `cardinality` aggregations on high-cardinality
string fields, which only requires doc values.
I also removed the `rehash` option from the `cardinality` aggregation as it
doesn't bring much value (rehashing is cheap) and allowed to remove the
coupling between the `cardinality` aggregation and the `murmur3` field.
Close#12874
Some of our next queries to refactor rely on some state taken from the cluster state. That is why we need to mock cluster service and inject an index to it, the index that we simulate the execution of the queries against. The best would be to have multiple indices actually, but that would make our setup a lot more complicated, especially given that IndexQueryParseService is still per index. We might be able to improve that in the future though, for now this is as good as it gets.
The Plugin interface currently contains 6 different methods for
adding modules. Elasticsearch has 3 different levels of injectors,
and for each of those, there are two methods. The first takes no
arguments and returns a collection of class objects to construct. The
second takes a Settings object and returns a collection of module
objects already constructed. The settings argument is unecessary because
the plugin can already get the settings from its constructor. Removing
that, the only difference between the two versions is returning an
already constructed Module, or a module Class, and there is no reason
the plugin can't construct all their modules themselves.
This change reduces the plugin api down to just 3 methods for adding
modules. Each returns a Collection<Module>. It also removes the
processModule method, which was unnecessary since onModule
implementations fullfill the same requirement. And finally, it renames
the modules() method to nodeModules() so it is clear these are created
once for each node.
java.net.preferIPv6Addresses is a better choice. preferIPv4Stack is a nuclear option
and you just won't even bind to any IPv6 addresses. This reduces confusion.
The restore portion of some snapshot/restore test is failing randomly due to #9421. This change suspends rebalance during snapshot/restore operations until #9421 is fixed.
Closes#12855
Plugins can preovide additional settings to be added to the settings
provided in elasticsearch.yml. However, if two different plugins supply
the same setting key, the last one to be loaded wins, which is
indeterminate. This change enforces plugins cannot have conflicting
settings, at startup time. As a followup, we should do this when
installing plugins as well, to give earlier errors when two plugins
collide.
Custom repository types are registered through the RepositoriesModule.
Later, when a specific repository type is used, the RespositoryModule
is installed, which in turn would spawn the module that was
registered for that repository type. However, a module is not needed
here. Each repository type has two associated classes, a Repository and
an IndexShardRepository.
This change makes the registration method for custom repository
types take both of these classes, instead of a module.
See #12783.
When elasticsearch is configured by interface (or default: loopback interfaces),
bind to all addresses on the interface rather than an arbitrary one.
If the publish address is not specified, default it from the bound addresses
based on the following sort ordering:
* ipv4/ipv6 (java.net.preferIPv4Stack, defaults to true)
* ordinary addresses
* site-local addresses
* link local addresses
* loopback addresses
One one address is published, and multicast is still always over ipv4: these
need to be future improvements.
Closes#12906Closes#12915
Squashed commit of the following:
commit 7e60833312f329a5749f9a256b9c1331a956d98f
Author: Robert Muir <rmuir@apache.org>
Date: Mon Aug 17 14:45:33 2015 -0400
fix java 7 compilation oops
commit c7b9f3a42058beb061b05c6dd67fd91477fd258a
Author: Robert Muir <rmuir@apache.org>
Date: Mon Aug 17 14:24:16 2015 -0400
Cleanup/fix logic around custom resolvers
commit bd7065f1936e14a29c9eb8fe4ecab0ce512ac08e
Author: Robert Muir <rmuir@apache.org>
Date: Mon Aug 17 13:29:42 2015 -0400
Add some unit tests for utility methods
commit 0faf71cb0ee9a45462d58af3d1bf214e8a79347c
Author: Robert Muir <rmuir@apache.org>
Date: Mon Aug 17 12:11:48 2015 -0400
localhost all the way down
commit e198bb2bc0d1673288b96e07e6e6ad842179978c
Merge: b55d092 b93a75f
Author: Robert Muir <rmuir@apache.org>
Date: Mon Aug 17 12:05:02 2015 -0400
Merge branch 'master' into network_cleanup
commit b55d092811d7832bae579c5586e171e9cc1ebe9d
Author: Robert Muir <rmuir@apache.org>
Date: Mon Aug 17 12:03:03 2015 -0400
fix docs, fix another bug in multicast (publish host = bad here!)
commit 88c462eb302b30a82585f95413927a5cbb7d54c4
Author: Robert Muir <rmuir@apache.org>
Date: Mon Aug 17 11:50:49 2015 -0400
remove nocommit
commit 89547d7b10d68b23d7f24362e1f4782f5e1ca03c
Author: Robert Muir <rmuir@apache.org>
Date: Mon Aug 17 11:49:35 2015 -0400
fix http too
commit 9b9413aca8a3f6397b5031831f910791b685e5be
Author: Robert Muir <rmuir@apache.org>
Date: Mon Aug 17 11:06:02 2015 -0400
Fix transport / interface code
Next up: multicast and then http
EsExecutorsTests had a test that was failing spuriously due to threadpools
being threadpools. This weakens the assertions that the test makes to what
should always be true.
It might turn out to be useful to have the actual commit hash of the version we are
looking for if our download manager can just redirect to the right staging repository.
This is purely for maintainance reasons since it easier to see if we can drop
certain stageing urls if we have the version next to the hash.
I also removed the gpg passphrase from the example URL since it's better to get prompted?
This is caused by sending the same file to the chunk handler with offset
`0` which in-turn opens a new outputstream and waits for bytes. But the next round
will send 0 bytes again with offset 0. This commit adds some checks / validators that those
settings are positive byte values and fixes the RecoveryStatus to throw an IAE if the same file
is opened twice.
What is the problem we are trying to solve ?
===========================================
When we are doing aggregations against a field name as shown in
https://github.com/HarishAtGitHub/elasticsearch-tester/blob/master/12135.py#L37-L46
search = {
"aggs": {
"NAME": {
"terms": {
"field": "ip_str",
"size": 10
}
}
}
}
and when the field "ip_str" has values of different types in different indices
. say one is of type StringTerms type and other is of IP(LongTerms type) then
the aggregation fails as the types do not match(incompatible).
The failure throws a class cast exception as follows:
{
"error": {
"root_cause": [],
"type": "reduce_search_phase_exception",
"reason": "[reduce] ",
"phase": "query",
"grouped": true,
"failed_shards": [],
"caused_by": {
"type": "class_cast_exception",
"reason": "org.elasticsearch.search.aggregations.bucket.terms.LongTerms$Bucket cannot be cast to org.elasticsearch.search.aggregations.bucket.terms.StringTerms$Bucket"
}
},
"status": 503
}
which is hard to understand . User cannot infer anything about the cause of the problem and what he should do from seeing the
class cast exception.
What can be the possible solution ?
===================================
Make the exception more readable by showing him the root cause of the problem so that he can
understand which area actually caused the problem, so that he can take necessary steps further.
Code Analysis
=============
Debugging code shows that:
the query /{indices}/_search?search_type=count involves two phases
1) search phase
***************
searchService.sendExecuteQuery(...) [Ref: TransportSearchCountAction]
what happens here ?
the phase 1, which is the search phase goes without error.
In this phase the shards for the given indexes are collected and the search is done on all asynchronously
and finally collected in the variable "firstResults" and given to meger phase.
[Flow: .... -> TransportSearchTypeAction -> method performFirstPhase]
2) merge phase
**************
searchPhaseController.merge(...firstResults...) [Ref: TransportSearchCountAction]
what happens here ?
the "firstresults" QuerySearchResults are now to be aggregated and combined.
[Flow: SearchPhaseController.merge(...) -> ..... -> InternalTerms.doReduce(...)]
the phase 1, which is the search phase goes without error.
The problem comes in phase 2, which is merge phase.
Now the individual term buckets are available.
As per the test case , there are two indices cast and cast2, so by default 10 shards.
cast has ip_str of type StringTerms
cast2 has ip_str of type ip which is actually LongTerms
so here two types of Buckets exist. StringTerms_Bucket and LongTerms_Bucket.
Now the aggregation is to be put inside the BucketPriorityQueue(size 2: as out of 10, 2 has hits) finally.
(docs of PriorityQueue: https://lucene.apache.org/core/4_4_0/core/org/apache/lucene/util/PriorityQueue.html#insertWithOverflow(T))
Now first the LongTerms$Bucket is put inside.
then the StringTerms$Bucket is to be put in.
This is the area where exception is thrown. What happens is when adding the StringTerms$Bucket now it has to
goes through the code "lessThan(element, heap[1])"
which finally calls
---------------------------------------------------------------------------------------------
| StringTerms$Bucket.compareTerms(other) <---------------- Area of exception |
| |
--------------------------------------------------------------------------------------------
where when comparing one to other a type cast is done and it fails as StringTerms$Bucket and LongTerms$Bucket are
incompatible.
Approach to solve:
==================
The best way is to make user understand that the problem is when reducing/merging/aggregating the buckets which came as a result of
querying different shards, so that this will make them infer that the problem is because the values of the fields are of different types.
The message is also user friendly and much better than the indecipherable classcastexception.
The only place to infer correctly that the aggregation has failed is in the place where aggregations take place.
so
at InternalTerms.java -> (BucketPriorityQueue)ordered.insertWithOverflow(b);
so here I can throw AggregationExecutionException saying it is because the buckets are of different
types.
But when can I infer at this point that the failure is due to mismatch of types of buckets ???
it can be possible only if at this point it is informed that the problem which occurred deep inside
is due to buckets that were incomparable.
so from just a classCastException we cannot make such a pointed exact inference, because
as class cast exception can be due to a number of scenarios and at a number of places.
so unless we inform the exact problem to InternalTerms it will not be able to infer properly.
so infer the classCastException at the compareTerms function itself that it is a IncomparableTermBucktesTypeException.
This is the best place to infer classCastException as this the place which generated the exception.
Best inference of exceptions can be done only at the source/origin of the exception.
so IncomparableTermBucktesTypeException to InternalTerms-> will make it infer and conclude on why
aggregation failed and give best information to user.
Close#12821
The IndicesModule was made up of two submodules, one which
handled registering queries, and the other for registering
hunspell dictionaries. This change moves those into
IndicesModule. It also adds a new extension point type,
InstanceMap. This is simply a Map<K,V>, where K and V are
actual objects, not classes like most other extension points.
I also added a test method to help testing instance map extensions.
This was particularly painful because of how guice binds the key
and value as separate bindings, and then reconstitutes them
into a Map at injection time. In order to gain access to the
object which links the key and value, I had to tweak our
guice copy to not use an anonymous inner class for the Provider.
Note that I also renamed the existing extension point types, since
they were very redundant. For example, ExtensionPoint.MapExtensionPoint
is now ExtensionPoint.ClassMap.
See #12783.
shard variable dependency from processFirstPhaseResults as shard is no more
needed here . it only deals with the results obtained from the synchronous search on each shard.
The ClusterModule contained a couple submodules. This moves the
functionality from those modules into ClusterModule. Two of those
had to do with DynamicSettings. This change also cleans up
how DynamicSettings are built, and enforces they are added, with
validators, in ClusterModule.
See #12783.
Before:
PluginInfo{name='cloud-aws', description='The Amazon Web Service (AWS) Cloud plugin allows to use AWS API for the unicast discovery mechanism and add S3 repositories.', site=false, jvm=true, classname=org.elasticsearch.plugin.cloud.aws.CloudAwsPlugin, isolated=true, version='2.1.0-SNAPSHOT'}
After:
- Plugin information:
Name: cloud-aws
Description: The Amazon Web Service (AWS) Cloud plugin allows to use AWS API for the unicast discovery mechanism and add S3 repositories.
Site: false
Version: 2.1.0-SNAPSHOT
JVM: true
* Classname: org.elasticsearch.plugin.cloud.aws.CloudAwsPlugin
* Isolated: true
Previously settings specified in index templates were not validated upon
template creation. Creating an index from an index template with invalid
settings could lead to cluster stability issues because creation of such
indexes would bypass index settings validation.
This commit adds validation of settings specified in index templates at
template creation time. This works by routing the index template
settings through the index settings validation mechanism.
Closes#12865
Users with IPv6 preferred over IPv4 may have `localhost` resolve to
`::1` instead of `127.0.0.1`, so we should be explicit so they don't run
into issues.
Today on a failure the reproduce line printed out by the test framework
will build all projects and might fail if the test class is not present.
This commit adds a reactor filter to the reproduction line to ensure
unrelated projects are skipped.
Closes#12838
In order to match the paths of official plugins, we need to fix
the broken test by removing the elasticsearch prefix from the official
plugin names before testing.
This PR adds a timezone field to ValueParser.DateMath that is
set to UTC by default but can be set using the existing constructors.
This makes it possible for extended bounds setting in DateHistogram
to also use date math expressions that e.g. round by day and apply
this rounding in the time zone specified in the date histogram
aggregation request.
Closes#12278
In order to create releases without actually changing the version
as part of a commit, we also need to reflect the path of the potentially
changing S3 repo.
This method has multiple modes of resolving config files by
first looking in the config directory, then on the classpath,
and finally by prefixing with "config/" on the classpath.
Most of the places taking advantage of this were tests, so they
did not have to setup a real home dir with config. The only place
that was really relying on it was the code which loads names.txt
to randomly choose a node name.
This change fixes test to setup fake home dirs with their config
files. It also makes the logic for finding names.txt explicit:
look in config dir, and if it doesn't exist, load /config/names.txt
from the classpath.
TermsQueryParser still parses those values although deprecated. These need to be present in the java api as well to get ready for the query refactoring, where the builders are the intermediate query format that we parse our json queries into. Whatever the parser supports need to be supported by the builder as well.
Closes#12870
TermsQueryParser doesn't support the cache field anymore, so if it gets set through java api, the subsequent parsing of that query will throw error
Relates to #12870
Refactored a part out of the release script, so the user can
change the version locally as well as move the documentation
and change the Version.java
The background of this change is to have a very simple release
process that puts stuff into a staging environment, so the beta
release can be tested, before it is officially released.
This means the build_release script can be removed soon.
Settings currently has a classloader member, which any user (plugin
or core ES code) can access to load classes/resources. This is extremely
error prone as setting the classloder on the Settings instance is a
public method. Furthermore, it is not really necessary. Classes that
need resources should load resources using normal means
(getClass().getResourceAsStream). Those that need classes
should use Class.forName, which will load the class with the
same classloader as the calling class. This means, in the few
places where classes are loaded by string name, they will use
the appropriate loader: either the default classloader which loads
core ES code, or a child classloader for each plugin.
This change removes the classloader member from Settings, as
well as other classloader related uses (except for a handful
of cases which must use a classloader, at least for now).
We previous used something like Class.forName to load mock classes,
where tests would set a setting that was *supposed* to only be used by
tests. This change make these impls package private so that only tests
can change out these implementations, through test plugins.
closes#12784
No need to load catch this query since it's cheap and not reused.
If we cache it, it can cause assertions to be tripped since this
method is executed during postRecovery phase and might still run while
nodes are shutdown in tests.
Leftover after adding the query builder type to QueryParser. getBuilderPrototype is better typed now and doesn't require unchecked cast anymore. Fixed also some Tuple usage without types.
This commit tries to add some infrastructure to streamline how extension
points should be strucutred. It's a simple approache with 4 implementations
for `highlighter`, `suggester`, `allocation_decider` and `shards_allocator`.
It simplifies adding new extension points and forces to register classes instead
of strings.
When parsing inner array of filters, AndQueryParser seems to
check for correct "filters" field name but then does the same
kind of operation in the `else` branch of the stament. This
seems like it can be removed.
Build fails with maven 3.3.1 and 3.3.3. To reproduce, install one of the 3.3.x versions of maven and run `mvn clean verify` in the root directory of the project. The build will fail in the QA: Smoke Test Shaded Jar module with the following error:
```
Started J0 PID(99979@flea.local).
Suite: org.elasticsearch.shaded.test.ShadedIT
2> NOTE: reproduce with: ant test -Dtestcase=ShadedIT -Dtests.method=testJodaIsNotOnTheCP -Dtests.seed=2F4D23A7462CF921 -Dtests.locale= -Dtests.timezone=Asia/Baku -Dtests.asserts=true -Dtests.file.encoding=UTF-8
FAILURE 0.06s | ShadedIT.testJodaIsNotOnTheCP <<<
> Throwable #1: junit.framework.AssertionFailedError: Expected an exception but the test passed: java.lang.ClassNotFoundException
> at __randomizedtesting.SeedInfo.seed([2F4D23A7462CF921:3A9404F1F69FD80]:0)
> at junit.framework.Assert.fail(Assert.java:57)
> at java.lang.Thread.run(Thread.java:745)
2> NOTE: reproduce with: ant test -Dtestcase=ShadedIT -Dtests.method=testGuavaIsNotOnTheCP -Dtests.seed=2F4D23A7462CF921 -Dtests.locale= -Dtests.timezone=Asia/Baku -Dtests.asserts=true -Dtests.file.encoding=UTF-8
FAILURE 0.01s | ShadedIT.testGuavaIsNotOnTheCP <<<
> Throwable #1: junit.framework.AssertionFailedError: Expected an exception but the test passed: java.lang.ClassNotFoundException
> at __randomizedtesting.SeedInfo.seed([2F4D23A7462CF921:C2502FD54D83433D]:0)
> at junit.framework.Assert.fail(Assert.java:57)
> at java.lang.Thread.run(Thread.java:745)
2> NOTE: reproduce with: ant test -Dtestcase=ShadedIT -Dtests.method=testjsr166eIsNotOnTheCP -Dtests.seed=2F4D23A7462CF921 -Dtests.locale= -Dtests.timezone=Asia/Baku -Dtests.asserts=true -Dtests.file.encoding=UTF-8
FAILURE 0.01s | ShadedIT.testjsr166eIsNotOnTheCP <<<
> Throwable #1: junit.framework.AssertionFailedError: Expected an exception but the test passed: java.lang.ClassNotFoundException
> at __randomizedtesting.SeedInfo.seed([2F4D23A7462CF921:35593286F4269392]:0)
> at junit.framework.Assert.fail(Assert.java:57)
> at java.lang.Thread.run(Thread.java:745)
2> NOTE: leaving temporary files on disk at: /Users/Shared/Jenkins/Home/workspace/elasticsearch-master/qa/smoke-test-shaded/target/J0/temp/org.elasticsearch.shaded.test.ShadedIT_2F4D23A7462CF921-001
2> NOTE: test params are: codec=CheapBastard, sim=DefaultSimilarity, locale=, timezone=Asia/Baku
2> NOTE: Mac OS X 10.10.4 x86_64/Oracle Corporation 1.8.0_25 (64-bit)/cpus=8,threads=1,free=482137936,total=514850816
2> NOTE: All tests run in this JVM: [ShadedIT]
Completed [1/1] in 6.61s, 5 tests, 3 failures <<< FAILURES!
Tests with failures:
- org.elasticsearch.shaded.test.ShadedIT.testJodaIsNotOnTheCP
- org.elasticsearch.shaded.test.ShadedIT.testGuavaIsNotOnTheCP
- org.elasticsearch.shaded.test.ShadedIT.testjsr166eIsNotOnTheCP
```
Please note that build doesn't fail with maven 3.2.x and it doesn't fail if mvn command is executed inside the qa/smoke-test-shaded directory. Only when the build is started from the root directory the error above can be observed.
The reason is because of the shaded version which depends on elasticsearch core.
When Maven build the module only, then elasticsearch core is not added to the dependency tree.
```sh
mvn dependency:tree -pl :smoke-test-shaded
```
```
[INFO] --- maven-dependency-plugin:2.8:tree (default-cli) @ smoke-test-shaded ---
[INFO] org.elasticsearch.qa:smoke-test-shaded:jar:2.0.0-beta1-SNAPSHOT
[INFO] +- org.elasticsearch.distribution.shaded:elasticsearch:jar:2.0.0-beta1-SNAPSHOT:compile
[INFO] | +- org.apache.lucene:lucene-core:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-backward-codecs:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-analyzers-common:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-queries:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-memory:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-highlighter:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-queryparser:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-sandbox:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-suggest:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-misc:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-join:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-grouping:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-spatial:jar:5.2.1:compile
[INFO] | \- com.spatial4j:spatial4j:jar:0.4.1:compile
[INFO] +- org.hamcrest:hamcrest-all:jar:1.3:test
[INFO] \- org.apache.lucene:lucene-test-framework:jar:5.2.1:test
[INFO] +- org.apache.lucene:lucene-codecs:jar:5.2.1:test
[INFO] +- com.carrotsearch.randomizedtesting:randomizedtesting-runner:jar:2.1.16:test
[INFO] +- junit:junit:jar:4.11:test
[INFO] \- org.apache.ant🐜jar:1.8.2:test
```
But if shaded plugin is involved during the build, it modifies the `projectArtifactMap`:
```sh
mvn dependency:tree -pl org.elasticsearch.distribution.shaded:elasticsearch,:smoke-test-shaded
```
```
[INFO] --- maven-dependency-plugin:2.8:tree (default-cli) @ smoke-test-shaded ---
[INFO] org.elasticsearch.qa:smoke-test-shaded:jar:2.0.0-beta1-SNAPSHOT
[INFO] +- org.elasticsearch.distribution.shaded:elasticsearch:jar:2.0.0-beta1-SNAPSHOT:compile
[INFO] | \- org.elasticsearch:elasticsearch:jar:2.0.0-beta1-SNAPSHOT:compile
[INFO] | +- org.apache.lucene:lucene-backward-codecs:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-analyzers-common:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-queries:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-memory:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-highlighter:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-queryparser:jar:5.2.1:compile
[INFO] | | \- org.apache.lucene:lucene-sandbox:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-suggest:jar:5.2.1:compile
[INFO] | | \- org.apache.lucene:lucene-misc:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-join:jar:5.2.1:compile
[INFO] | | \- org.apache.lucene:lucene-grouping:jar:5.2.1:compile
[INFO] | +- org.apache.lucene:lucene-spatial:jar:5.2.1:compile
[INFO] | | \- com.spatial4j:spatial4j:jar:0.4.1:compile
[INFO] | +- com.google.guava:guava:jar:18.0:compile
[INFO] | +- com.carrotsearch:hppc:jar:0.7.1:compile
[INFO] | +- joda-time:joda-time:jar:2.8:compile
[INFO] | +- org.joda:joda-convert:jar:1.2:compile
[INFO] | +- com.fasterxml.jackson.core:jackson-core:jar:2.5.3:compile
[INFO] | +- com.fasterxml.jackson.dataformat:jackson-dataformat-smile:jar:2.5.3:compile
[INFO] | +- com.fasterxml.jackson.dataformat:jackson-dataformat-yaml:jar:2.5.3:compile
[INFO] | | \- org.yaml:snakeyaml:jar:1.12:compile
[INFO] | +- com.fasterxml.jackson.dataformat:jackson-dataformat-cbor:jar:2.5.3:compile
[INFO] | +- io.netty:netty:jar:3.10.3.Final:compile
[INFO] | +- com.ning:compress-lzf:jar:1.0.2:compile
[INFO] | +- com.tdunning:t-digest:jar:3.0:compile
[INFO] | +- org.hdrhistogram:HdrHistogram:jar:2.1.6:compile
[INFO] | +- org.apache.commons:commons-lang3:jar:3.3.2:compile
[INFO] | +- commons-cli:commons-cli:jar:1.3.1:compile
[INFO] | \- com.twitter:jsr166e:jar:1.1.0:compile
[INFO] +- org.hamcrest:hamcrest-all:jar:1.3:test
[INFO] \- org.apache.lucene:lucene-test-framework:jar:5.2.1:test
[INFO] +- org.apache.lucene:lucene-codecs:jar:5.2.1:test
[INFO] +- org.apache.lucene:lucene-core:jar:5.2.1:compile
[INFO] +- com.carrotsearch.randomizedtesting:randomizedtesting-runner:jar:2.1.16:test
[INFO] +- junit:junit:jar:4.11:test
[INFO] \- org.apache.ant🐜jar:1.8.2:test
```
A fix could consist of fixing something on Maven side. Probably something changed in a recent version and introduced this "issue" but it might be not really an issue. More a fix.
There are two workarounds:
1) exclude manually elasticsearch core from shaded version in smoke-test-shaded module and add manually each lucene lib needed by elasticsearch
2) add a new `elasticsearch-lucene` (lucene) POM module which simply declares all needed lucene libs in subprojects (such as the smoke tester one).
I choose the later.
Closes#12791.
This allows `path.shared_data` to be added to the security manager while
still allowing a custom `data_path` for indices using shadow replicas.
For example, configuring `path.shared_data: /tmp/foo`, then created an
index with:
```
POST /myindex
{
"index": {
"number_of_shards": 1,
"number_of_replicas": 1,
"data_path": "/tmp/foo/bar/baz",
"shadow_replicas": true
}
}
```
The index will then reside in `/tmp/foo/bar/baz`.
`path.shared_data` defaults to `null` if not specified.
Resolves#12714
Relates to #11065
The upper bound must be 0-based since we are corrupting an offset into
the file but it can be a 1-based length of the file which results in an
uncorrupted file.
There were two submodules of AllocationModule. This combines them into a
single module, adds a base test case for module testing, and adds back
the ability for plugins to provide custom ShardsAllocators.
closes#12781
Simon Willnauer
Britta Weber
javanna
Britta Weber
Jason Tedor
Boaz Leskes
Ryan Ernst
Adrien Grand
Lee Hinman
Martijn van Groningen
Zachary Tong
Tanguy Leroux
Nicholas Knize
David Pilato
Robert Muir
Robert Muir
Colin Goodheart-Smithe
Igor Motov
Stephen Samuel
Simon Willnauer
Ivannikov Kirill
Isabel Drost-Fromm
Marshall Bockrath-Vandegrift
Isabel Drost-Fromm
xuzha
Michael McCandless
Alexander Reelsen
Michael McCandless
Jim Hooker
Nik Everett
André Carvalho
Alex Ksikes
Christoph Büscher
Harish Kayarohanam
Andrey Fadeyev
saschamarkus
Konrad Beiske