Reindex, update by query and delete by query would silently disregard
RED/unavailable shards, thus not copying, updating or deleting matching
data in those shards. Now use `allow_partial_search_results=false` to
ensure these operations fail if the search crosses an unavailable chard.
Added the option to explicitly specify `allow_partial_search_results=true`
for reindex only (seemed too strange for update/delete by query).
Relates #45739 and #42612
* [ML] ML Model Inference Ingest Processor (#49052)
* [ML][Inference] adds lazy model loader and inference (#47410)
This adds a couple of things:
- A model loader service that is accessible via transport calls. This service will load in models and cache them. They will stay loaded until a processor no longer references them
- A Model class and its first sub-class LocalModel. Used to cache model information and run inference.
- Transport action and handler for requests to infer against a local model
Related Feature PRs:
* [ML][Inference] Adjust inference configuration option API (#47812)
* [ML][Inference] adds logistic_regression output aggregator (#48075)
* [ML][Inference] Adding read/del trained models (#47882)
* [ML][Inference] Adding inference ingest processor (#47859)
* [ML][Inference] fixing classification inference for ensemble (#48463)
* [ML][Inference] Adding model memory estimations (#48323)
* [ML][Inference] adding more options to inference processor (#48545)
* [ML][Inference] handle string values better in feature extraction (#48584)
* [ML][Inference] Adding _stats endpoint for inference (#48492)
* [ML][Inference] add inference processors and trained models to usage (#47869)
* [ML][Inference] add new flag for optionally including model definition (#48718)
* [ML][Inference] adding license checks (#49056)
* [ML][Inference] Adding memory and compute estimates to inference (#48955)
* fixing version of indexed docs for model inference
The logic for `cleanupInProgress()` was backwards everywhere (method itself and
all but one user). Also, we weren't checking it when removing a repository.
This lead to a bug (in the one spot that didn't use the method backwards) that prevented
the cleanup cluster state entry from ever being removed from the cluster state if master
failed over during the cleanup process.
This change corrects the backwards logic, adds a test that makes sure the cleanup
is always removed and adds a check that prevents repository removal during cleanup
to the repositories service.
Also, the failure handling logic in the cleanup action was broken. Repeated invocation would lead to the cleanup being removed from the cluster state even if it was in progress. Fixed by adding a flag that indicates whether or not any removal of the cleanup task from the cluster state must be executed. Sorry for mixing this in here, but I had to fix it in the same PR, as the first test (for master-failover) otherwise would often just delete the blocked cleanup action as a result of a transport master action retry.
* Better Exceptions on Concurrent Snapshot Operations
It is somewhat tricky to debug test failures from concurrent operations
without having the exact knowledge of what ran concurrently so I added
it to these exceptions in all spots.
The network disruption was acting on node ids and node names
which made reconnects not work. Moved all usages to node names
to fix this. Since the map of all nodes in the test is indexed
by name this was easier to work with.
* Make FsBlobContainer Listing Resilient to Concurrent Modifications
If we list out files in a folder via the lazily computed directory
stream, we have to deal with concurrent deletes when reading the file
attributes since we don't have a lock on the directory in any way.
Closes#37581
Make queries on the “_index” field fast-fail if the target shard is an index that doesn’t match the query expression. Part of the “canMatch” phase optimisations.
Closes#48473
This is intended as a stop-gap solution/improvement to #38941 that
prevents repo modifications without an intermittent master failover
from causing inconsistent (outdated due to inconsistent listing of index-N blobs)
`RepositoryData` to be written.
Tracking the latest repository generation will move to the cluster state in a
separate pull request. This is intended as a low-risk change to be backported as
far as possible and motived by the recently increased chance of #38941
causing trouble via SLM (see https://github.com/elastic/elasticsearch/issues/47520).
Closes#47834Closes#49048
This commit introduces a new class called ESSloppyMath
that is meant to reflect the purpose of Lucene's SloppyMath,
but add additional unimplemented faster alternatives to math functions.
The two that are used by geotile-grid a lot are sinh/atan.
In a quick elasticsearch rally benchmark for geotile-grid on Switzerland
data points, this shows a (1.22x) 22% speed-up over using Math's functions.
closes#41166.
Currently the `_analyze` endpoint doesn't correctly use normalizers specified
in the request. This change fixes that by returning the resolved normalizer from
TransportAnalyzeAction#getAnalyzer and updates test to be able to catch this
in the future.
Closes#48650
Today we wrap exceptions that occur while executing an ingest processor
in an ElasticsearchException. Today, in ExceptionsHelper#unwrapCause we
only unwrap causes for exceptions that implement
ElasticsearchWrapperException, which the top-level
ElasticsearchException does not. Ultimately, this means that any
exception that occurs during processor execution does not have its cause
unwrapped, and so its status is blanket treated as a 500. This means
that while executing a bulk request with an ingest pipeline,
document-level failures that occur during a processor will cause the
status for that document to be treated as 500. Since that does not give
the client any indication that they made a mistake, it means some
clients will enter infinite retries, thinking that there is some
server-side problem that merely needs to clear. This commit addresses
this by introducing a dedicated ingest processor exception, so that its
causes can be unwrapped. While we could consider a broader change to
unwrap causes for more than just ElasticsearchWrapperExceptions, that is
a broad change with unclear implications. Since the problem of reporting
500s on client errors is a user-facing bug, we take the conservative
approach for now, and we can revisit the unwrapping in a future change.
When a node shuts down, `TransportService` moves to stopped state and
then closes connections. If a request is done in between, an exception
was thrown that was not retried in replication actions. Now throw a
wrapped `NodeClosedException` exception instead, which is correctly
handled in replication action. Fixed other usages too.
Relates #42612
Ensures that we always use the primary term established by the primary to index docs on the
replica. Makes the logic around replication less brittle by always using the operation primary
term on the replica that is coming from the primary.
Reverts #48947 and fixes the issue orginally addressed by removing the assertion.
It turns out we can't simply pass empty shard generations to the snapshot finalization in the
BwC case as that results in no indices being added to the meta for the given snapshot since
we take the indices from the shard generations (even in the BwC case the `null` generations work
fine for this).
Closes#48983
Today in 6.x it is possible to add an index tombstone to the graveyard without
deleting the corresponding index metadata, because the deletion is slightly
deferred. If you shut down the node and upgrade to 7.x when in this state then
the node will fail to apply any cluster states, reporting
java.lang.IllegalStateException: Cannot delete index [...], it is still part of the cluster state.
This commit addresses this situation by skipping over any index metadata with a
corresponding tombstone, allowing this metadata to be cleaned up by the 7.x
node.
This commits sends the cluster name and discovery naode in the transport
level handshake response. This will allow us to stop sending the
transport service level handshake request in the 8.0-8.x release cycle.
It is necessary to start sending this in 7.x so that 8.0 is guaranteed
to be communicating with a version that sends the required information.
Currently the BulkProcessor class uses a single scheduler to schedule
flushes and retries. Functionally these are very different concerns but
can result in a dead lock. Specifically, the single shared scheduler
can kick off a flush task, which only finishes it's task when the bulk
that is being flushed finishes. If (for what ever reason), any items in
that bulk fails it will (by default) schedule a retry. However, that retry
will never run it's task, since the flush task is consuming the 1 and
only thread available from the shared scheduler.
Since the BulkProcessor is mostly client based code, the client can
provide their own scheduler. As-is the scheduler would require
at minimum 2 worker threads to avoid the potential deadlock. Since the
number of threads is a configuration option in the scheduler, the code
can not enforce this 2 worker rule until runtime. For this reason this
commit splits the single task scheduler into 2 schedulers. This eliminates
the potential for the flush task to block the retry task and removes this
deadlock scenario.
This commit also deprecates the Java APIs that presume a single scheduler,
and updates any internal code to no longer use those APIs.
Fixes#47599
Note - #41451 fixed the general case where a bulk fails and is retried
that can result in a deadlock. This fix should address that case as well as
the case when a bulk failure *from the flush* needs to be retried.
We were tripping the assertion that the makes sure we only have empty `ShardGenerations` in `RepositoryData` in the BwC case because shard generations were passed to the `Repository` in the BwC case. Fixed by only generating empty shard gen for BwC snapshots in `SnapshotsService`.
Backport of #48450.
Make a number of changes so that code in the `server` directory is more
resilient to automatic formatting. This covers:
* Reformatting multiline JSON to embed whitespace in the strings
* Move some comments around to they aren't auto-formatted to a strange
place. This also required moving some `&&` and `||` operators from the
end-of-line to start-of-line`.
* Add helper method `reformatJson()`, to strip whitespace from a JSON
document using XContent methods. This is sometimes necessary where
a test is comparing some machine-generated JSON with an expected
value.
Also, `HyperLogLogPlusPlus.java` is now excluded from formatting because it
contains large data tables that don't reformat well with the current settings,
and changing the settings would be worse for the rest of the codebase.
The realtime GET API currently has erratic performance in case where a document is accessed
that has just been indexed but not refreshed yet, as the implementation will currently force an
internal refresh in that case. Refreshing can be an expensive operation, and also will block the
thread that executes the GET operation, blocking other GETs to be processed. In case of
frequent access of recently indexed documents, this can lead to a refresh storm and terrible
GET performance.
While older versions of Elasticsearch (2.x and older) did not trigger refreshes and instead opted
to read from the translog in case of realtime GET API or update API, this was removed in 5.0
(#20102) to avoid inconsistencies between values that were returned from the translog and
those returned by the index. This was partially reverted in 6.3 (#29264) to allow _update and
upsert to read from the translog again as it was easier to guarantee consistency for these, and
also brought back more predictable performance characteristics of this API. Calls to the realtime
GET API, however, would still always do a refresh if necessary to return consistent results. This
means that users that were calling realtime GET APIs to coordinate updates on client side
(realtime GET + CAS for conditional index of updated doc) would still see very erratic
performance.
This PR (together with #48707) resolves the inconsistencies between reading from translog and
index. In particular it fixes the inconsistencies that happen when requesting stored fields, which
were not available when reading from translog. In case where stored fields are requested, this
PR will reparse the _source from the translog and derive the stored fields to be returned. With
this, it changes the realtime GET API to allow reading from the translog again, avoid refresh
storms and blocking the GET threadpool, and provide overall much better and predictable
performance for this API.
We should not open new engines if a shard is closed. We break this
assumption in #45263 where we stop verifying the shard state before
creating an engine but only before swapping the engine reference.
We can fail to snapshot the store metadata or checkIndex a closed shard
if there's some IndexWriter holding the index lock.
Closes#47060
This change fixes a poisonous situation where an ongoing recovery was
canceled because a better copy was found on a node that the cluster had
previously tried allocating the shard to but failed. The solution is to
keep track of the set of nodes that an allocation was failed on so that
we can avoid canceling the current recovery for a copy on failed nodes.
Closes#47974
Today if the primary discovers that an indexing request needs a mapping update
then it will send it to the master for validation and processing. If, however,
the put-mapping request is invalid then the master still processes it as a
(no-op) cluster state update. When there are a large number of indexing
operations that result in invalid mapping updates this can overwhelm the
master.
However, the primary already has a reasonably up-to-date mapping against which
it can check the (approximate) validity of the put-mapping request before
sending it to the master. For instance it is not possible to remove fields in a
mapping update, so if the primary detects that a mapping update will exceed the
fields limit then it can reject it itself and avoid bothering the master.
This commit adds a pre-flight check to the mapping update path so that the
primary can discard obviously-invalid put-mapping requests itself.
Fixes#35564
Backport of #48817
This test failure manifests the limitation of the recovery source merge
policy explained in #41628. If we already merge down to a single segment
then subsequent force merges will be noop although they can prune
recovery source. We need to adjust this test until we have a fix for the
merge policy.
Relates #41628Closes#48735
The loading of `RepositoryData` is not an atomic operation.
It uses a list + get combination of calls.
This lead to accidentally returning an empty repository data
for generations >=0 which can never not exist unless the repository
is corrupted.
In the test #48122 (and other SLM tests) there was a low chance of
running into this concurrent modification scenario and the repository
actually moving two index generations between listing out the
index-N and loading the latest version of it. Since we only keep
two index-N around at a time this lead to unexpectedly absent
snapshots in status APIs.
Fixing the behavior to be more resilient is non-trivial but in the works.
For now I think we should simply throw in this scenario. This will also
help prevent corruption in the unlikely event but possible of running into this
issue in a snapshot create or delete operation on master failover on a
repository like S3 which doesn't have the "no overwrites" protection on
writing a new index-N.
Fixes#48122
The problem with wrapping here is that it converts any exception into an
IAE, which we treat as a client error (400 status) whereas the exception
being wrapped here could be a server error (e.g., NPE). This commit
stops wrapping all ingest processor exceptions as IAEs.
* Add ingest info to Cluster Stats (#48485)
This commit enhances the ClusterStatsNodes response to include global
processor usage stats on a per-processor basis.
example output:
```
...
"processor_stats": {
"gsub": {
"count": 0,
"failed": 0
"current": 0
"time_in_millis": 0
},
"script": {
"count": 0,
"failed": 0
"current": 0,
"time_in_millis": 0
}
}
...
```
The purpose for this enhancement is to make it easier to collect stats on how specific processors are being used across the cluster beyond the current per-node usage statistics that currently exist in node stats.
Closes#46146.
* fix BWC of ingest stats
The introduction of processor types into IngestStats had a bug.
It was set to `null` and set as the key to the map. This would
throw a NPE. This commit resolves this by setting all the processor
types from previous versions that are not serializing it out to
`_NOT_AVAILABLE`.
Previous behavior while copying HTTP headers to the ThreadContext,
would allow multiple HTTP headers with the same name, handling only
the first occurrence and disregarding the rest of the values. This
can be confusing when dealing with multiple Headers as it is not
obvious which value is read and which ones are silently dropped.
According to RFC-7230, a client must not send multiple header fields
with the same field name in a HTTP message, unless the entire field
value for this header is defined as a comma separated list or this
specific header is a well-known exception.
This commits changes the behavior in order to be more compliant to
the aforementioned RFC by requiring the classes that implement
ActionPlugin to declare if a header can be multi-valued or not when
registering this header to be copied over to the ThreadContext in
ActionPlugin#getRestHeaders.
If the header is allowed to be multivalued, then all such headers
are read from the HTTP request and their values get concatenated in
a comma-separated string.
If the header is not allowed to be multivalued, and the HTTP
request contains multiple such Headers with different values, the
request is rejected with a 400 status.
Today a couple of allocation deciders iterate through all the shards on a node
to find the `INITIALIZING` or `RELOCATING` ones, and this can slow down cluster
state updates in clusters with very high-density nodes holding many thousands
of shards even if those shards belong to closed or frozen indices. This commit
pre-computes the sets of `INITIALIZING` and `RELOCATING` shards to speed up
this search.
Closes#46941
Relates #48579
Co-authored-by: "hongju.xhj" <hongju.xhj@alibaba-inc.com>
Backport of #48553. Make a number of changes so that JSON in the server
directory is more resilient to automatic formatting. This covers:
* Reformatting multiline JSON to embed whitespace in the strings
* Add helper method `stripWhitespace()`, to strip whitespace from a JSON
document using XContent methods. This is sometimes necessary where
a test is comparing some machine-generated JSON with an expected
value.
Today it is possible that we import a dangling index that was created in a
newer version than one or more of the nodes in the cluster. Such an index would
prevent the older node(s) from rejoining the cluster if they were to briefly
leave it for some reason. This commit prevents the import of such dangling
indices.
Fixes#34264
With this change, we won't warm up searchers until we externally refresh
an engine. We explicitly refresh before allowing reading from a shard
(i.e., move to post_recovery state) and during resetting. These
guarantees that we have warmed up the engine before exposing the
external searcher.
Another prerequisite for #47186.
Fixes the shard snapshot status reporting for failed shards
in the corner case of failing the shard because of an exception
thrown in `SnapshotShardsService` and not the repository.
We were missing the update on the `snapshotStatus` instance in
this case which made the transport APIs using this field report
back an incorrect status.
Fixed by moving the failure handling to the `SnapshotShardsService`
for all cases (which also simplifies the code, the ex. wrapping in
the repository was pointless as we only used the ex. trace upstream
anyway).
Also, added an assertion to another test that explicitly checks this
failure situation (ex. in the `SnapshotShardsService`) already.
Closes#48526
We can run into an already closed store here and hence
throw on trying to increment the ref count => moving to
the guarded ref count increment
closes#48625
Henning Andersen
Benjamin Trent
gpaimla
Armin Braun
Yannick Welsch
Mark Tozzi
markharwood
Jason Tedor
Tal Levy
bellengao
Christoph Büscher
Igor Motov
David Turner
Tim Brooks
Jake Landis
Rory Hunter
Nhat Nguyen
Adrien Grand
Ioannis Kakavas
Zachary Tong
Stéphane Campinas
kkewwei
Arvind Ramachandran
Ignacio Vera