This commit is a follow up to the work completed in #27132. Essentially
it transitions two more methods (sendMessage and getLocalAddress) from
Transport to TcpChannel. With this change, there is no longer a need for
TcpTransport to be aware of the specific type of channel a transport
returns. So that class is no longer parameterized by channel type.
The default value for ignore_unavailable did not match what was documented when using the REST APIs for snapshot creation and restore. This commit sets the default value of ignore_unavailable to false, the way it is documented and ensures it's the same when using either REST API or transport client.
Closes#25359
* This change adds a module called `aggs-composite` that defines a new aggregation named `composite`.
The `composite` aggregation is a multi-buckets aggregation that creates composite buckets made of multiple sources.
The sources for each bucket can be defined as:
* A `terms` source, values are extracted from a field or a script.
* A `date_histogram` source, values are extracted from a date field and rounded to the provided interval.
This aggregation can be used to retrieve all buckets of a deeply nested aggregation by flattening the nested aggregation in composite buckets.
A composite buckets is composed of one value per source and is built for each document as the combinations of values in the provided sources.
For instance the following aggregation:
````
"test_agg": {
"terms": {
"field": "field1"
},
"aggs": {
"nested_test_agg":
"terms": {
"field": "field2"
}
}
}
````
... which retrieves the top N terms for `field1` and for each top term in `field1` the top N terms for `field2`, can be replaced by a `composite` aggregation in order to retrieve **all** the combinations of `field1`, `field2` in the matching documents:
````
"composite_agg": {
"composite": {
"sources": [
{
"field1": {
"terms": {
"field": "field1"
}
}
},
{
"field2": {
"terms": {
"field": "field2"
}
}
},
}
}
````
The response of the aggregation looks like this:
````
"aggregations": {
"composite_agg": {
"buckets": [
{
"key": {
"field1": "alabama",
"field2": "almanach"
},
"doc_count": 100
},
{
"key": {
"field1": "alabama",
"field2": "calendar"
},
"doc_count": 1
},
{
"key": {
"field1": "arizona",
"field2": "calendar"
},
"doc_count": 1
}
]
}
}
````
By default this aggregation returns 10 buckets sorted in ascending order of the composite key.
Pagination can be achieved by providing `after` values, the values of the composite key to aggregate after.
For instance the following aggregation will aggregate all composite keys that sorts after `arizona, calendar`:
````
"composite_agg": {
"composite": {
"after": {"field1": "alabama", "field2": "calendar"},
"size": 100,
"sources": [
{
"field1": {
"terms": {
"field": "field1"
}
}
},
{
"field2": {
"terms": {
"field": "field2"
}
}
}
}
}
````
This aggregation is optimized for indices that set an index sorting that match the composite source definition.
For instance the aggregation above could run faster on indices that defines an index sorting like this:
````
"settings": {
"index.sort.field": ["field1", "field2"]
}
````
In this case the `composite` aggregation can early terminate on each segment.
This aggregation also accepts multi-valued field but disables early termination for these fields even if index sorting matches the sources definition.
This is mandatory because index sorting picks only one value per document to perform the sort.
Today if nested docs are used in an index that is split the operation
will only work correctly if the index is not routing partitioned or
unless routing is used. This change fixes the query that selectes the docs
to delete to also select all parents nested docs as well.
Closes#27378
Right now our different transport implementations must duplicate
functionality in order to stay compliant with the requirements of
TcpTransport. They must all implement common logic to open channels,
close channels, keep track of channels for eventual shutdown, etc.
Additionally, there is a weird and complicated relationship between
Transport and TransportService. We eventually want to start merging
some of the functionality between these classes.
This commit starts moving towards a world where TransportService retains
all the application logic and channel state. Transport implementations
in this world will only be tasked with returning a channel when one is
requested, calling transport service when a channel is accepted from
a server, and starting / stopping itself.
Specifically this commit changes how channels are opened and closed. All
Transport implementations now return a channel type that must comply with
the new TcpChannel interface. This interface has the methods necessary
for TcpTransport to completely manage the lifecycle of a channel. This
includes setting the channel up, waiting for connection, adding close
listeners, and eventually closing.
Right now we have unnecessary implementations of `TransportChannel`.
Additionally, there are methods on the interface that are not used. This
commit removes unnecessary implementations and methods.
The field data cache can come under heavy contention in cases when lots
of search threads are hitting it for doc values. This commit reduces the
amount of contention here by using a double-checked locking strategy to
only lock when the cache needs to be initialized.
Relates #27365
The toXContent method for IndexGraveYard (which is a collection of tombstones for explicitly marking indices as deleted in the cluster state) confused timeValue with dateField, resulting in output of the form "delete_date" : "23424.3d" instead of "delete_date":"2017-11-13T15:50:51.614Z".
When the current master node is shutting down, it sends a leave request to the other nodes so that they can eagerly start a fresh master election. Unfortunately, it was still possible for the master node that was shutting down to respond to ping requests, possibly influencing the election decision as it still appeared as an active master in the ping responses. This commit ensures that UnicastZenPing does not respond to ping requests once it's been closed. ZenDiscovery.doStop() continues to ensure that the pinging component is first closed before it triggers a master election.
Closes#27328
We use affix settings to group settings / values under a certain namespace.
In some cases like login information for instance a setting is only valid if
one or more other settings are present. For instance `x.test.user` is only valid
if there is an `x.test.passwd` present and vice versa. This change allows to specify
such a dependency to prevent settings updates that leave settings in an inconsistent
state.
This commit removes an unnecessary logger instance creation from the
constructor for doc values field data. This construction is expensive
for this oft-created class because of a synchronized block in the
constructor for the logger.
Relates #27349
This is the first step to supporting WKT (and other future) format(s). The ShapeBuilders are quite messy and can be simplified by decoupling the parse logic from the build logic. This commit refactors the parsing logic into its own package separate from the Shape builders. It also decouples the GeoShapeType into a standalone enumerator that is responsible for validating the parsed data and providing the appropriate builder. This future-proofs the code making it easier to maintain and add new shape types.
This is a followup to #26521. This commit expands the alias added for
the elasticsearch client codebase to all codebases. The original full
jar name property is left intact. This only adds an alias without the
version, which should help ease the pain in updating any versions (ES
itself or dependencies).
Queries that create a scroll context cannot use the cache.
They modify the search context during their execution so using the cache
can lead to duplicate result for the next scroll query.
This change fails the entire request if the request_cache option is explictely set
on a query that creates a scroll context (`scroll=1m`) and make sure internally that we never
use the cache for these queries when the option is not explicitely used.
For 6.x a deprecation log will be printed instead of failing the entire request and the request_cache hint
will be ignored (forced to false).
The order in which double values are added in java can give different results
for the sum, so we need to allow a certain delta in the test assertions. The
current value was still a bit too low, which manifested itself in occasional
test failures.
This commit adds a parent pipeline aggregation that allows
sorting the buckets of a parent multi-bucket aggregation.
The aggregation also offers [from] and [size] parameters
in order to truncate the result as desired.
Closes#14928
We cut over to internal and external IndexReader/IndexSearcher in #26972 which uses
two independent searcher managers. This has the downside that refreshes of the external
reader will never clear the internal version map which in-turn will trigger additional
and potentially unnecessary segment flushes since memory must be freed. Under heavy
indexing load with low refresh intervals this can cause excessive segment creation which
causes high GC activity and significantly increases the required segment merges.
This change adds a dedicated external reference manager that delegates refreshes to the
internal reference manager that then `steals` the refreshed reader from the internal
reference manager for external usage. This ensures that external and internal readers
are consistent on an external refresh. As a sideeffect this also releases old segments
referenced by the internal reference manager which can potentially hold on to already merged
away segments until it is refreshed due to a flush or indexing activity.
In order to avoid churn when applying a new `ClusterState`, there are some checks that compare parts of the old and new states and, if equal, the new object is discarded and the old one reused. Since `ClusterState` updates are now largely diff-based, this code is unnecessary: applying a diff also reuses any old objects if unchanged. Moreover, the code compares the parts of the `ClusterState` using their `version()` values which is not guaranteed to be correct, because of a lack of consensus.
This change removes this optimisation, and tests that objects are still reused as expected via the diff mechanism.
This commit fixes a comment in an index shard test which was inaccurate
after it was copied from another test and not modified to reflect the
reasoning in the test that it was copied into.
When a primary is promoted, rolling the translog generation here makes
simpler reasoning about the relationship between primary terms and
translog generation. Note that this is not strictly necessary for
correctness (e.g., to avoid duplicate operations with the same sequence
number within a single generation).
Relates #27313
* Fixes#19768: scripted_metric _agg parameter disappears if params are provided
* Test case for #19768
* Compare boolean to false instead of negating it
* Added mocked script in ScriptedMetricIT
* Fix test in ScriptedMetricIT for implicit _agg map
* Add limits for ngram and shingle settings (#27211)
Create index-level settings:
max_ngram_diff - maximum allowed difference between max_gram and min_gram in
NGramTokenFilter/NGramTokenizer. Default is 1.
max_shingle_diff - maximum allowed difference between max_shingle_size and
min_shingle_size in ShingleTokenFilter. Default is 3.
Throw an IllegalArgumentException when
trying to create NGramTokenFilter, NGramTokenizer, ShingleTokenFilter
where difference between max_size and min_size exceeds the settings value.
Closes#25887
The `TemplateUpgradeService` allows plugins to register a call back that mutates index templates upon recovery. This is handy for upgrade logic that needs to make sure that an existing index template is updated once the cluster is upgraded to a new version of the plugin (and ES).
Currently, the service has complicated logic to decide which node should perform the upgrade. It will prefer the master node, if it is of the highest version of the cluster and otherwise it will fall back to one of the non-coordinating nodes which are on the latest version. While this attempts to make sure that new nodes can assume their template version is in place (but old node still need to be able to operate under both old and new template), it has an inherent problem in that the master (on an old version) may not be able to process the put template request with the new template - it may miss certain features.
This PR changes the logic to be simpler and always rely on the current master nodes. This comes at the price that new nodes need to operate both with old templates and new. That price is small as they need to operate with old indices regardless of the template. On the flip side we reduce a lot of complexity in what can happen in the cluster.
If an out of memory error is thrown while merging, today we quietly
rewrap it into a merge exception and the out of memory error is
lost. Instead, we need to rethrow out of memory errors, and in fact any
fatal error here, and let those go uncaught so that the node is torn
down. This commit causes this to be the case.
Relates #27265
Some code-paths use anonymous classes (such as NonCollectingAggregator
in terms agg), which messes up the display name of the profiler. If
we encounter an anonymous class, we need to grab the super's name.
Another naming issue was that ProfileAggs were not delegating to the
wrapped agg's name for toString(), leading to ugly display.
This PR also fixes up the profile documentation. Some of the examples were
executing against empty indices, which shows different profile results
than a populated index (and made for confusing examples).
Finally, I switched the agg display names from the fully qualified name
to the simple name, so that it's similar to how the query profiles work.
Closes#26405
The warnings headers have a fairly limited set of valid characters
(cf. quoted-text in RFC 7230). While we have assertions that we adhere
to this set of valid characters ensuring that our warning messages do
not violate the specificaion, we were neglecting the possibility that
arbitrary user input would trickle into these warning headers. Thus,
missing here was tests for these situations and encoding of characters
that appear outside the set of valid characters. This commit addresses
this by encoding any characters in a deprecation message that are not
from the set of valid characters.
Relates #27269
This change adds a new `_split` API that allows to split indices into a new
index with a power of two more shards that the source index. This API works
alongside the `_shrink` API but doesn't require any shard relocation before
indices can be split.
The split operation is conceptually an inverse `_shrink` operation since we
initialize the index with a _syntetic_ number of routing shards that are used
for the consistent hashing at index time. Compared to indices created with
earlier versions this might produce slightly different shard distributions but
has no impact on the per-index backwards compatibility. For now, the user is
required to prepare an index to be splittable by setting the
`index.number_of_routing_shards` at index creation time. The setting allows the
user to prepare the index to be splittable in factors of
`index.number_of_routing_shards` ie. if the index is created with
`index.number_of_routing_shards: 16` and `index.number_of_shards: 2` it can be
split into `4, 8, 16` shards. This is an intermediate step until we can make
this the default. This also allows us to safely backport this change to 6.x.
The `_split` operation is implemented internally as a DeleteByQuery on the
lucene level that is executed while the primary shards execute their initial
recovery. Subsequent merges that are triggered due to this operation will not be
executed immediately. All merges will be deferred unti the shards are started
and will then be throttled accordingly.
This change is intended for the 6.1 feature release but will not support pre-6.1
indices to be split unless these indices have been shrunk before. In that case
these indices can be split backwards into their original number of shards.
While it's not possible to upgrade the Jackson dependencies
to their latest versions yet (see #27032 (comment) for more)
it's still possible to upgrade to the latest 2.8.x version.
We have an hidden setting called `index.queries.cache.term_queries` that disables caching of term queries in the query cache.
Though term queries are not cached in the Lucene UsageTrackingQueryCachingPolicy since version 6.5.
This makes the es policy useless but also makes it impossible to re-enable caching for term queries.
This change appeared in Lucene 6.5 so this setting is no-op since version 5.4 of Elasticsearch
The change in this PR removes the setting and the custom policy.
Only tests should use the single argument Environment constructor. To
enforce this the single arg Environment constructor has been replaced with
a test framework factory method.
Production code (beyond initial Bootstrap) should always use the same
Environment object that Node.getEnvironment() returns. This Environment
is also available via dependency injection.
If the master disconnects from the cluster after initiating snapshot, but just before the snapshot switches from INIT to STARTED state, the snapshot can get indefinitely stuck in the INIT state. This error is specific to v5.x+ and was triggered by keeping the master node that stepped down in the node list, the cleanup logic in snapshot/restore assumed that if master steps down it is always removed from the the node list. This commit changes the logic to trigger cleanup even if no nodes left the cluster.
Closes#27180
For FsBlobStore and HdfsBlobStore, if the repository is read only, the blob store should be aware of the readonly setting and do not create directories if they don't exist.
Closes#21495
After recent changes in InternalStats#doXContentBody the corresponding xContent
output of the parsed aggregation needed to be changed in a similar way.
* Uses norms for exists query if enabled
This change means that for indexes created from 6.1.0, if normas are enabled we will not write the field name to the `_field_names` field and for an exists query we will instead use the NormsFieldExistsQuery which was added in Lucene 7.1.0. If norms are not enabled or if the index was created before 6.1.0 `_field_names` will be used as before.
* Fixes tests
The uid bytes (as the type#id) were needlessly being created even though
they are no longer needed after the move to single type per index. This
commit avoids creating these when parsed documents are constructed.
Relates #27241
We do some accounting in IndexShard that is not necessarily correct since
we maintain two different index readers. This change moves the accounting under
the engine which knows what reader we are refreshing.
Relates to #26972
This local checkpoint tracker uses collections of bit sets to track
which sequence numbers are complete, eventually removing these bit sets
when the local checkpoint advances. However, these bit sets were eagerly
allocated so that if a sequence number far ahead of the checkpoint was
marked as completed, all bit sets between the "last" bit set and the bit
set needed to track the marked sequence number were allocated. If this
sequence number was too far ahead, the memory requirements could be
excessive. This commit opts for a different strategy for holding on to
these bit sets and enables them to be lazily allocated.
Relates #27179
We added an index-level setting for controlling the size of the bit sets
used to back the local checkpoint tracker. This setting is really only
needed to control the memory footprint of the bit sets but we do not
think this setting is going to be needed. This commit removes this
setting before it is released to the wild after which we would have to
worry about BWC implications.
Relates #27191
* Enhances exists queries to reduce need for `_field_names`
Before this change we wrote the name all the fields in a document to a `_field_names` field and then implemented exists queries as a term query on this field. The problem with this approach is that it bloats the index and also affects indexing performance.
This change adds a new method `existsQuery()` to `MappedFieldType` which is implemented by each sub-class. For most field types if doc values are available a `DocValuesFieldExistsQuery` is used, falling back to using `_field_names` if doc values are disabled. Note that only fields where no doc values are available are written to `_field_names`.
Closes#26770
* Addresses review comments
* Addresses more review comments
* implements existsQuery explicitly on every mapper
* Reinstates ability to perform term query on `_field_names`
* Added bwc depending on index created version
* Review Comments
* Skips tests that are not supported in 6.1.0
These values will need to be changed after backporting this PR to 6.x
This query returns documents that match with at least one ore more
of the provided terms. The number of terms that must match varies
per document and is either controlled by a minimum should match
field or computed per document in a minimum should match script.
Closes#26915
It is required in order to work correctly with bulk scorer implementations
that change the scorer during the collection process. Otherwise sub collectors
might call `Scorer.score()` on the wrong scorer.
Closes#27131
This commit is a minor refactoring of internal engine to move hooks for
generating sequence numbers into the engine itself. As such, we refactor
tests that relied on this hook to use the new hook, and remove the hook
from the sequence number service itself.
Relates #27082
This change is required in order to support a size based check for the
index rollover.
The index size is estimated by sampling the existing segments only. We
prefer using segments to StoreStats because StoreStats is not reliable
if indexing or merging operations are in progress.
Relates #27004
This change makes sure that we track score when sort is set to relevancy only.
In this case we always track max score like normal search does.
Closes#23840
* Apply missing request options to the expand phase
This change adds some missing options to the expand query that builds the inner hits for field collapsing.
The following options are now applied to the inner_hits query:
* post_filters
* preferences
* routing
Closes#27079Closes#26649
The new discovery stats were pushed to the 6.x branch (currently
versioned at 6.1.0) but master was not updated to reflect this. This
impacts the mixed-cluster BWC tests because a 6.1.0 node will be trying
to send a 7.0.0 node the new discovery stats but the 7.0.0 did not yet
understand that it should be reading these when talking to a 6.1.0
node. This commit addresses this, and changes the skip version on the
discovery stats REST tests.
Windows handles trying to read a file that does not exist because a
component of the path is not a directory differently than other OS
handle this situation. This commit adjusts these assertions for Windows.
When executing a cluster settings update that leaves the cluster state
unchanged, we skip validation and this avoids deprecation logging for
deprecated settings in the cluster state. This commit addresses this by
running validation even if the settings are unchanged.
Relates #27017
When a search is executing locally over many shards, we can stack
overflow during query phase execution. This happens due to callbacks
that occur after a phase completes for a shard and we move to the same
phase on another shard. If all the shards for the query are local to the
local node then we will never go async and these callbacks will end up
as recursive calls. With sufficiently many shards, this will end up as a
stack overflow. This commit addresses this by truncating the stack by
forking to another thread on the executor for the phase.
Relates #27069
Finder creates these files if you browse a directory there. These files
are really annoying, but it's an incredible pain for users that these
files are created unbeknownst to them, and then they get in the way of
Elasticsearch starting. This commit adds leniency on macOS only to skip
these files.
Relates #27108
Turns out that `ShardSearchTarget` is nullable, hence its fields may not be printed out as part of `ShardSearchFailure#toXContent`, in which case `fromXContent` cannot parse it back. We would previously try to create the object with all of its fields set to null, but `Index` complains about it in the constructor. Also made sure that this code path is covered by our unit tests in `ShardSearchFailureTests`.
Closes#27055
Introduce minimal thread scheduler as a base class for `ThreadPool`. Such a class can be used from the `BulkProcessor` to schedule retries and the flush task. This allows to remove the `ThreadPool` dependency from `BulkProcessor`, which requires to provide settings that contain `node.name` and also needed log4j for logging. Instead, it needs now a `Scheduler` that is much lighter and gets automatically created and shut down on close.
Closes#26028
It's believed that using diffs obsoletes the other mechanism for reusing the
bits of the ClusterState that didn't change between updates, but in fact we
don't know for sure how often the diff mechanism works successfully. The stats
collected here will tell us.
Right now if the number of shards for a particular index is equal across the
data paths, we tie-break on space. This changes to tie-break first on the total
number of shards for each path, and then, if that is the same, on the usable
bytes.
Relates to #26654 (it's a follow-up)
If timed runnable wraps an abstract runnable, then it should delegate to
the abstract runnable otherwise force execution and handling rejections
is dropped on the floor. Thus, timed runnable should itself be an
abstract runnable delegating all methods to the wrapped runnable in
cases when it is an abstract runnable. This commit causes this to be the
case.
Relates #27095
Memory usage of queries can't be properly accounted, which can be an issue when
large queries are cached since the actual memory usage will be much higher than
what the cache thinks. This problem is very hard if not impossible to fix so as
a workaround I would like to decrease the maximum number of cached queries so
that this problem is less likely to cause trouble in practice.
For the record, this problem is more likely to occur in envirenments that have
small shards or don't give much memory to the JVM.
Closes#26938
Today we internally accumulate elapsed scroll time in nanoseconds. The
problem here is that this can reasonably overflow. For example, on a
system with scrolls that are open for ten minutes on average, after
sixteen million scrolls the largest value that can be represented by a
long will be executed. To address this, we switch to internally
representing scrolls using microseconds as this enables with the same
number of scrolls scrolls that are open for seven days on average, or
with the same average elapsed time sixteen billion scrolls which will
never happen (executing one scroll a second until sixteen billion have
executed would not occur until more than five-hundred years had
elapsed).
Relates #27068
Upgrade to Jackson 2.9.2 and also use a boolean `closed` flag to
indicate that a FastStringReader instance is closed, so that length
is still correctly reported after the reader is closed.
Due to a change happened via #26102 to make the nested source consistent
with or without source filtering, the _source of a nested inner hit was
always wrapped in the parent path. This turned out to be not ideal for
users relying on the nested source, as it would require additional parsing
on the client side. This change fixes this, the _source of nested inner hits
is now no longer wrapped by parent json objects, irregardless of whether
the _source is included as is or source filtering is used.
Internally source filtering and highlighting relies on the fact that the
_source of nested inner hits are accessible by its full field path, so
in order to now break this, the conversion of the _source into its binary
form is performed in FetchSourceSubPhase, after any potential source filtering
is performed to make sure the structure of _source of the nested inner hit
is consistent irregardless if source filtering is performed.
PR for #26944Closes#26944
David Turner
Tim Brooks
kel
Jim Ferenczi
Simon Willnauer
olcbean
Jason Tedor
tinder-xli
Yannick Welsch
Nicholas Knize
Ryan Ernst
Jim Ferenczi
Christoph Büscher
Dimitris Athanasiou
Costin Leau
Boaz Leskes
Igor Motov
Reese Levine
Mayya Sharipova
Zachary Tong
Tanguy Leroux
Nhat
Nhat
David Roberts
Chris Earle
Colin Goodheart-Smithe
Mathias Fußenegger
Martijn van Groningen
Jack Conradson
Shai Erera
javanna
Armin Braun
Adrien Grand
Nhat
Catalin Ursachi
Luca Cavanna
Lee Hinman
Lee Hinman