Check it out:
```
$ curl -u elastic:password -HContent-Type:application/json -XPOST localhost:9200/test/_update/foo?pretty -d'{
"dac": {}
}'
{
"error" : {
"root_cause" : [
{
"type" : "x_content_parse_exception",
"reason" : "[2:3] [UpdateRequest] unknown field [dac] did you mean [doc]?"
}
],
"type" : "x_content_parse_exception",
"reason" : "[2:3] [UpdateRequest] unknown field [dac] did you mean [doc]?"
},
"status" : 400
}
```
The tricky thing about implementing this is that x-content doesn't
depend on Lucene. So this works by creating an extension point for the
error message using SPI. Elasticsearch's server module provides the
"spell checking" implementation.
s
* Track Snapshot Version in RepositoryData (#50930)
Add tracking of snapshot versions to RepositoryData to make BwC logic more efficient.
Follow up to #50853
There is a JVM bug causing `Thread#suspend` calls to randomly take
multiple seconds breaking these tests that call the method numerous times
in a loop. Increasing the timeout would will not work since we may call
`suspend` tens if not hundreds of times and even a small number of them
experiencing the blocking will lead to multiple minutes of waiting.
This PR detects the specific issue by timing the `Thread#suspend` calls and
skips the remainder of the test if it timed out because of the JVM bug.
Closes#50047
* Move metadata storage to Lucene (#50907)
Today we split the on-disk cluster metadata across many files: one file for the metadata of each
index, plus one file for the global metadata and another for the manifest. Most metadata updates
only touch a few of these files, but some must write them all. If a node holds a large number of
indices then it's possible its disks are not fast enough to process a complete metadata update before timing out. In severe cases affecting master-eligible nodes this can prevent an election
from succeeding.
This commit uses Lucene as a metadata storage for the cluster state, and is a squashed version
of the following PRs that were targeting a feature branch:
* Introduce Lucene-based metadata persistence (#48733)
This commit introduces `LucenePersistedState` which master-eligible nodes
can use to persist the cluster metadata in a Lucene index rather than in
many separate files.
Relates #48701
* Remove per-index metadata without assigned shards (#49234)
Today on master-eligible nodes we maintain per-index metadata files for every
index. However, we also keep this metadata in the `LucenePersistedState`, and
only use the per-index metadata files for importing dangling indices. However
there is no point in importing a dangling index without any shard data, so we
do not need to maintain these extra files any more.
This commit removes per-index metadata files from nodes which do not hold any
shards of those indices.
Relates #48701
* Use Lucene exclusively for metadata storage (#50144)
This moves metadata persistence to Lucene for all node types. It also reenables BWC and adds
an interoperability layer for upgrades from prior versions.
This commit disables a number of tests related to dangling indices and command-line tools.
Those will be addressed in follow-ups.
Relates #48701
* Add command-line tool support for Lucene-based metadata storage (#50179)
Adds command-line tool support (unsafe-bootstrap, detach-cluster, repurpose, & shard
commands) for the Lucene-based metadata storage.
Relates #48701
* Use single directory for metadata (#50639)
Earlier PRs for #48701 introduced a separate directory for the cluster state. This is not needed
though, and introduces an additional unnecessary cognitive burden to the users.
Co-Authored-By: David Turner <david.turner@elastic.co>
* Add async dangling indices support (#50642)
Adds support for writing out dangling indices in an asynchronous way. Also provides an option to
avoid writing out dangling indices at all.
Relates #48701
* Fold node metadata into new node storage (#50741)
Moves node metadata to uses the new storage mechanism (see #48701) as the authoritative source.
* Write CS asynchronously on data-only nodes (#50782)
Writes cluster states out asynchronously on data-only nodes. The main reason for writing out
the cluster state at all is so that the data-only nodes can snap into a cluster, that they can do a
bit of bootstrap validation and so that the shard recovery tools work.
Cluster states that are written asynchronously have their voting configuration adapted to a non
existing configuration so that these nodes cannot mistakenly become master even if their node
role is changed back and forth.
Relates #48701
* Remove persistent cluster settings tool (#50694)
Adds the elasticsearch-node remove-settings tool to remove persistent settings from the on
disk cluster state in case where it contains incompatible settings that prevent the cluster from
forming.
Relates #48701
* Make cluster state writer resilient to disk issues (#50805)
Adds handling to make the cluster state writer resilient to disk issues. Relates to #48701
* Omit writing global metadata if no change (#50901)
Uses the same optimization for the new cluster state storage layer as the old one, writing global
metadata only when changed. Avoids writing out the global metadata if none of the persistent
fields changed. Speeds up server:integTest by ~10%.
Relates #48701
* DanglingIndicesIT should ensure node removed first (#50896)
These tests occasionally failed because the deletion was submitted before the
restarting node was removed from the cluster, causing the deletion not to be
fully acked. This commit fixes this by checking the restarting node has been
removed from the cluster.
Co-authored-by: David Turner <david.turner@elastic.co>
* fix tests
Co-authored-by: David Turner <david.turner@elastic.co>
* Fix Snapshot Repository Corruption in Downgrade Scenarios (#50692)
This PR introduces test infrastructure for downgrading a cluster while interacting with a given repository.
It fixes the fact that repository metadata in the new format could be written while there's still older snapshots in the repository that require the old-format metadata to be restorable.
It's impossible to tell why #50754 fails without this change.
We're failing to close the `exchange` somewhere and there is no
write timeout in the GCS SDK (something to look into separately)
only a read timeout on the socket so if we're failing on an assertion without
reading the full request body (at least into the read-buffer) we're locking up
waiting forever on `write0`.
This change ensure the `exchange` is closed in the tests where we could lock up
on a write and logs the failure so we can find out what broke #50754.
This solves half of the problem in #46813 by moving the S3
tests to using the shared minio fixture so we at least have
some non-3rd-party, constantly running coverage on these tests.
This PR adds per-field metadata that can be set in the mappings and is later
returned by the field capabilities API. This metadata is completely opaque to
Elasticsearch but may be used by tools that index data in Elasticsearch to
communicate metadata about fields with tools that then search this data. A
typical example that has been requested in the past is the ability to attach
a unit to a numeric field.
In order to not bloat the cluster state, Elasticsearch requires that this
metadata be small:
- keys can't be longer than 20 chars,
- values can only be numbers or strings of no more than 50 chars - no inner
arrays or objects,
- the metadata can't have more than 5 keys in total.
Given that metadata is opaque to Elasticsearch, field capabilities don't try to
do anything smart when merging metadata about multiple indices, the union of
all field metadatas is returned.
Here is how the meta might look like in mappings:
```json
{
"properties": {
"latency": {
"type": "long",
"meta": {
"unit": "ms"
}
}
}
}
```
And then in the field capabilities response:
```json
{
"latency": {
"long": {
"searchable": true,
"aggreggatable": true,
"meta": {
"unit": [ "ms" ]
}
}
}
}
```
When there are no conflicts, values are arrays of size 1, but when there are
conflicts, Elasticsearch includes all unique values in this array, without
giving ways to know which index has which metadata value:
```json
{
"latency": {
"long": {
"searchable": true,
"aggreggatable": true,
"meta": {
"unit": [ "ms", "ns" ]
}
}
}
}
```
Closes#33267
Introduce a new static setting, `gateway.auto_import_dangling_indices`, which prevents dangling indices from being automatically imported. Part of #48366.
* Faster and Simpler GCS REST Mock
I reworked the GCS mock a little to use less copying+allocation,
log the full request body on failure to read a multi-part request
and generally be a little simpler and easy to follow to track down
the remaining issues that are causing almost daily failures from this
class's multi-part request parsing that can't be reproduced locally.
This test seems to be bogus as it was confusing a nominal execution time with a
delay (i.e. an elapsed time). This commit reworks the test to address this.
Fixes#50650
* Fix GCS Mock Broken Handling of some Blobs
We were incorrectly handling blobs starting in `\r\n` which broke
tests randomly when blobs started on these.
Relates #49429
* Adds JavaDoc to `AbstractWireTestCase` and
`AbstractWireSerializingTestCase` so it is more obvious you should prefer
the latter if you have a choice
* Moves the `instanceReader` method out of `AbstractWireTestCase` becaue
it is no longer used.
* Marks a bunch of methods final so it is more obvious which classes are
for what.
* Cleans up the side effects of the above.
The additional change to the original PR (#49657), is that `org.elasticsearch.client.cluster.RemoteConnectionInfo` now parses the initial_connect_timeout field as a string instead of a TimeValue instance.
The reason that this is needed is because that the initial_connect_timeout field in the remote connection api is serialized for human consumption, but not for parsing purposes.
Therefore the HLRC can't parse it correctly (which caused test failures in CI, but not in the PR CI
:( ). The way this field is serialized needs to be changed in the remote connection api, but that is a breaking change. We should wait making this change until rest api versioning is introduced.
Co-Authored-By: j-bean <anton.shuvaev91@gmail.com>
Co-authored-by: j-bean <anton.shuvaev91@gmail.com>
Today, the replica allocator uses peer recovery retention leases to
select the best-matched copies when allocating replicas of indices with
soft-deletes. We can employ this mechanism for indices without
soft-deletes because the retaining sequence number of a PRRL is the
persisted global checkpoint (plus one) of that copy. If the primary and
replica have the same retaining sequence number, then we should be able
to perform a noop recovery. The reason is that we must be retaining
translog up to the local checkpoint of the safe commit, which is at most
the global checkpoint of either copy). The only limitation is that we
might not cancel ongoing file-based recoveries with PRRLs for noop
recoveries. We can't make the translog retention policy comply with
PRRLs. We also have this problem with soft-deletes if a PRRL is about to
expire.
Relates #45136
Relates #46959
We need to make sure that the global checkpoints and peer recovery
retention leases were advanced to the max_seq_no and synced; otherwise,
we can risk expiring some peer recovery retention leases because of the
file-based recovery threshold.
Relates #49448
* Add ILM histore store index (#50287)
* Add ILM histore store index
This commit adds an ILM history store that tracks the lifecycle
execution state as an index progresses through its ILM policy. ILM
history documents store output similar to what the ILM explain API
returns.
An example document with ALL fields (not all documents will have all
fields) would look like:
```json
{
"@timestamp": 1203012389,
"policy": "my-ilm-policy",
"index": "index-2019.1.1-000023",
"index_age":123120,
"success": true,
"state": {
"phase": "warm",
"action": "allocate",
"step": "ERROR",
"failed_step": "update-settings",
"is_auto-retryable_error": true,
"creation_date": 12389012039,
"phase_time": 12908389120,
"action_time": 1283901209,
"step_time": 123904107140,
"phase_definition": "{\"policy\":\"ilm-history-ilm-policy\",\"phase_definition\":{\"min_age\":\"0ms\",\"actions\":{\"rollover\":{\"max_size\":\"50gb\",\"max_age\":\"30d\"}}},\"version\":1,\"modified_date_in_millis\":1576517253463}",
"step_info": "{... etc step info here as json ...}"
},
"error_details": "java.lang.RuntimeException: etc\n\tcaused by:etc etc etc full stacktrace"
}
```
These documents go into the `ilm-history-1-00000N` index to provide an
audit trail of the operations ILM has performed.
This history storage is enabled by default but can be disabled by setting
`index.lifecycle.history_index_enabled` to `false.`
Resolves#49180
* Make ILMHistoryStore.putAsync truly async (#50403)
This moves the `putAsync` method in `ILMHistoryStore` never to block.
Previously due to the way that the `BulkProcessor` works, it was possible
for `BulkProcessor#add` to block executing a bulk request. This was bad
as we may be adding things to the history store in cluster state update
threads.
This also moves the index creation to be done prior to the bulk request
execution, rather than being checked every time an operation was added
to the queue. This lessens the chance of the index being created, then
deleted (by some external force), and then recreated via a bulk indexing
request.
Resolves#50353
Co-authored-by: Daniel Huang <danielhuang@tencent.com>
This is a spinoff of #48130 that generalizes the proposal to allow early termination with the composite aggregation when leading sources match a prefix or the entire index sort specification.
In such case the composite aggregation can use the index sort natural order to early terminate the collection when it reaches a composite key that is greater than the bottom of the queue.
The optimization is also applicable when a query other than match_all is provided. However the optimization is deactivated for sources that match the index sort in the following cases:
* Multi-valued source, in such case early termination is not possible.
* missing_bucket is set to true
Avoid backwards incompatible changes for 8.x and 7.6 by removing type
restriction on compile and Factory. Factories may optionally implement
ScriptFactory. If so, then they can indicate determinism and thus
cacheability.
**Backport**
Relates: #49466
Cache results from queries that use scripts if they use only
deterministic API calls. Nondeterministic API calls are marked in the
whitelist with the `@nondeterministic` annotation. Examples are
`Math.random()` and `new Date()`.
Refs: #49466
We need to read in a loop here. A single read to a huge byte array will
only read 16k max with the S3 SDK so if the blob we're trying to fully
read is larger we close early and fail the size comparison.
Also, drain streams fully when checking existence to avoid S3 SDK warnings.
With node ordinals gone, there's no longer a need for such a complicated full cluster restart
procedure (as we can now uniquely associate nodes to data folders).
Follow-up to #41652
Follow up to #49729
This change removes falling back to listing out the repository contents to find the latest `index-N` in write-mounted blob store repositories.
This saves 2-3 list operations on each snapshot create and delete operation. Also it makes all the snapshot status APIs cheaper (and faster) by saving one list operation there as well in many cases.
This removes the resiliency to concurrent modifications of the repository as a result and puts a repository in a `corrupted` state in case loading `RepositoryData` failed from the assumed generation.
* Remove BlobContainer Tests against Mocks
Removing all these weird mocks as asked for by #30424.
All these tests are now part of real repository ITs and otherwise left unchanged if they had
independent tests that didn't call the `createBlobStore` method previously.
The HDFS tests also get added coverage as a side-effect because they did not have an implementation
of the abstract repository ITs.
Closes#30424
Since 7.4, we switch from translog to Lucene as the source of history
for peer recoveries. However, we reduce the likelihood of
operation-based recoveries when performing a full cluster restart from
pre-7.4 because existing copies do not have PPRL.
To remedy this issue, we fallback using translog in peer recoveries if
the recovering replica does not have a peer recovery retention lease,
and the replication group hasn't fully migrated to PRRL.
Relates #45136
Today we do not use retention leases in peer recovery for closed indices
because we can't sync retention leases on closed indices. This change
allows that ability and adjusts peer recovery to use retention leases
for all indices with soft-deletes enabled.
Relates #45136
Co-authored-by: David Turner <david.turner@elastic.co>
See discussion in #50047 (comment).
There are reproducible issues with Thread#suspend in Jdk11 and Jdk12 for me locally
and we have one failure for each on CI.
Jdk8 and Jdk13 are stable though on CI and in my testing so I'd selectively disable
this test here to keep the coverage. We aren't using suspend in production code so the
JDK bug behind this does not affect us.
Closes#50047
* Remove Unused Single Delete in BlobStoreRepository
There are no more production uses of the non-bulk delete or the delete that throws
on missing so this commit removes both these methods.
Only the bulk delete logic remains. Where the bulk delete was derived from single deletes,
the single delete code was inlined into the bulk delete method.
Where single delete was used in tests it was replaced by bulk deleting.
* Better Logging GCS Blobstore Mock
Two things:
1. We should just throw a descriptive assertion error and figure out why we're not reading a multi-part instead of
returning a `400` and failing the tests that way here since we can't reproduce these 400s locally.
2. We were missing logging the exception on a cleanup delete failure that coincides with the `400` issue in tests.
Relates #49429
Batch deletes get a response for every delete request, not just those that actually hit an existing blob.
The fact that we only responded for existing blobs leads to a degenerate response that throws a parse exception if a batch delete only contains non-existant blobs.
Multiple version ranges are allowed to be used in section skip in yml
tests. This is useful when a bugfix was backported to latest versions
and all previous releases contain a wire breaking bug.
examples:
6.1.0 - 6.3.0, 6.6.0 - 6.7.9, 7.0 -
- 7.2, 8.0.0 -
backport #50014
Step on the road to #49060.
This commit adds the logic to keep track of a repository's generation
across repository operations. See changes to package level Javadoc for the concrete changes in the distributed state machine.
It updates the write side of new repository generations to be fully consistent via the cluster state. With this change, no `index-N` will be overwritten for the same repository ever. So eventual consistency issues around conflicting updates to the same `index-N` are not a possibility any longer.
With this change the read side will still use listing of repository contents instead of relying solely on the cluster state contents.
The logic for that will be introduced in #49060. This retains the ability to externally delete the contents of a repository and continue using it afterwards for the time being. In #49060 the use of listing to determine the repository generation will be removed in all cases (except for full-cluster restart) as the last step in this effort.
In order to cache script results in the query shard cache, we need to
check if scripts are deterministic. This change adds a default method
to the script factories, `isResultDeterministic() -> false` which is
used by the `QueryShardContext`.
Script results were never cached and that does not change here. Future
changes will implement this method based on whether the results of the
scripts are deterministic or not and therefore cacheable.
Refs: #49466
**Backport**
Historically only two things happened in the final reduction:
empty buckets were filled, and pipeline aggs were reduced (since it
was the final reduction, this was safe). Usage of the final reduction
is growing however. Auto-date-histo might need to perform
many reductions on final-reduce to merge down buckets, CCS
may need to side-step the final reduction if sending to a
different cluster, etc
Having pipelines generate their output in the final reduce was
convenient, but is becoming increasingly difficult to manage
as the rest of the agg framework advances.
This commit decouples pipeline aggs from the final reduction by
introducing a new "top level" reduce, which should be called
at the beginning of the reduce cycle (e.g. from the SearchPhaseController).
This will only reduce pipeline aggs on the final reduce after
the non-pipeline agg tree has been fully reduced.
By separating pipeline reduction into their own set of methods,
aggregations are free to use the final reduction for whatever
purpose without worrying about generating pipeline results
which are non-reducible
Adds `GET /_script_language` to support Kibana dynamic scripting
language selection.
Response contains whether `inline` and/or `stored` scripts are
enabled as determined by the `script.allowed_types` settings.
For each scripting language registered, such as `painless`,
`expression`, `mustache` or custom, available contexts for the language
are included as determined by the `script.allowed_contexts` setting.
Response format:
```
{
"types_allowed": [
"inline",
"stored"
],
"language_contexts": [
{
"language": "expression",
"contexts": [
"aggregation_selector",
"aggs"
...
]
},
{
"language": "painless",
"contexts": [
"aggregation_selector",
"aggs",
"aggs_combine",
...
]
}
...
]
}
```
Fixes: #49463
**Backport**
* Copying the request is not necessary here. We can simply release it once the response has been generated and a lot of `Unpooled` allocations that way
* Relates #32228
* I think the issue that preventet that PR that PR from being merged was solved by #39634 that moved the bulk index marker search to ByteBuf bulk access so the composite buffer shouldn't require many additional bounds checks (I'd argue the bounds checks we add, we save when copying the composite buffer)
* I couldn't neccessarily reproduce much of a speedup from this change, but I could reproduce a very measureable reduction in GC time with e.g. Rally's PMC (4g heap node and bulk requests of size 5k saw a reduction in young GC time by ~10% for me)
This commit fixes a number of issues with data replication:
- Local and global checkpoints are not updated after the new operations have been fsynced, but
might capture a state before the fsync. The reason why this probably went undetected for so
long is that AsyncIOProcessor is synchronous if you index one item at a time, and hence working
as intended unless you have a high enough level of concurrent indexing. As we rely in other
places on the assumption that we have an up-to-date local checkpoint in case of synchronous
translog durability, there's a risk for the local and global checkpoints not to be up-to-date after
replication completes, and that this won't be corrected by the periodic global checkpoint sync.
- AsyncIOProcessor also has another "bad" side effect here: if you index one bulk at a time, the
bulk is always first fsynced on the primary before being sent to the replica. Further, if one thread
is tasked by AsyncIOProcessor to drain the processing queue and fsync, other threads can
easily pile more bulk requests on top of that thread. Things are not very fair here, and the thread
might continue doing a lot more fsyncs before returning (as the other threads pile more and
more on top), which blocks it from returning as a replication request (e.g. if this thread is on the
primary, it blocks the replication requests to the replicas from going out, and delaying
checkpoint advancement).
This commit fixes all these issues, and also simplifies the code that coordinates all the after
write actions.
This rewrites long sort as a `DistanceFeatureQuery`, which can
efficiently skip non-competitive blocks and segments of documents.
Depending on the dataset, the speedups can be 2 - 10 times.
The optimization can be disabled with setting the system property
`es.search.rewrite_sort` to `false`.
Optimization is skipped when an index has 50% or more data with
the same value.
Optimization is done through:
1. Rewriting sort as `DistanceFeatureQuery` which can
efficiently skip non-competitive blocks and segments of documents.
2. Sorting segments according to the primary numeric sort field(#44021)
This allows to skip non-competitive segments.
3. Using collector manager.
When we optimize sort, we sort segments by their min/max value.
As a collector expects to have segments in order,
we can not use a single collector for sorted segments.
We use collectorManager, where for every segment a dedicated collector
will be created.
4. Using Lucene's shared TopFieldCollector manager
This collector manager is able to exchange minimum competitive
score between collectors, which allows us to efficiently skip
the whole segments that don't contain competitive scores.
5. When index is force merged to a single segment, #48533 interleaving
old and new segments allows for this optimization as well,
as blocks with non-competitive docs can be skipped.
Backport for #48804
Co-authored-by: Jim Ferenczi <jim.ferenczi@elastic.co>
* Make BlobStoreRepository Aware of ClusterState (#49639)
This is a preliminary to #49060.
It does not introduce any substantial behavior change to how the blob store repository
operates. What it does is to add all the infrastructure changes around passing the cluster service to the blob store, associated test changes and a best effort approach to tracking the latest repository generation on all nodes from cluster state updates. This brings a slight improvement to the consistency
by which non-master nodes (or master directly after a failover) will be able to determine the latest repository generation. It does not however do any tricky checks for the situation after a repository operation
(create, delete or cleanup) that could theoretically be used to get even greater accuracy to keep this change simple.
This change does not in any way alter the behavior of the blobstore repository other than adding a better "guess" for the value of the latest repo generation and is mainly intended to isolate the actual logical change to how the
repository operates in #49060
Removing a lot of needless buffering and array creation
to reduce the significant memory usage of tests using this.
The incoming stream from the `exchange` is already buffered
so there is no point in adding a ton of additional buffers
everywhere.
This commit adds a function in NodeClient that allows to track the progress
of a search request locally. Progress is tracked through a SearchProgressListener
that exposes query and fetch responses as well as partial and final reduces.
This new method can be used by modules/plugins inside a node in order to track the
progress of a local search request.
Relates #49091
Yannick Welsch
Nik Everett
Armin Braun
Tim Brooks
Nhat Nguyen
Adrien Grand
Rory Hunter
David Turner
Henning Andersen
Martijn van Groningen
Lee Hinman
Jim Ferenczi
Stuart Tettemer
Przemyslaw Gomulka
Zachary Tong
Mayya Sharipova