Hide the `.async-search-*` in Security by making it a restricted index namespace.
The namespace is hard-coded.
To grant privileges on restricted indices, one must explicitly toggle the
`allow_restricted_indices` flag in the indices permission in the role definition.
As is the case with any other index, if a certain user lacks all permissions for an
index, that index is effectively nonexistent for that user.
The OpenIdConnectRealm had a bug which would cause it not to populate
User metadata for collections contained in the user JWT claims.
This commit fixes that bug.
Backport of: #50521
* [ML][Inference] PUT API (#50852)
This adds the `PUT` API for creating trained models that support our format.
This includes
* HLRC change for the API
* API creation
* Validations of model format and call
* fixing backport
* Fix SLM check for restore in progress (#50868)
* Fix SLM check for restore in progress
This commit fixes the check in SLM where the `RestoreInProgress`
metadata was checked for existence. Rather than check existence we
should instead check the `isEmpty` method. Prior to this, a successful
restore for a repository that used SLM retention would prevent SLM
retention from running in subsequent invocations, due to SLM thinking
that a restore was still running.
* Fix 7.x-isms
This commit removes validation logic of source and dest indices
for data frame analytics and replaces it with using the common
`SourceDestValidator` class which is already used by transforms.
This way the validations and their messages become consistent
while we reduce code.
This means that where these validations fail the error messages
will be slightly different for data frame analytics.
Backport of #50841
Replaces the "funny"
`Function<String, ConstructingObjectParser<T, Void>>` with a much
simpler `ConstructingObjectParser<T, String>`. This makes pretty much
all of our object parsers static.
A very large number of recursive calls can cause a stack overflow
exception. This commit forks the recursive calls for non-async
processors. Once forked, each thread will handle at most 10
recursive calls to help keep the stack size and thread count
down to a reasonable size.
This adds the necessary named XContent classes to the HLRC for the lang ident model. This is so the HLRC can call `GET _ml/inference/lang_ident_model_1?include_definition=true` without XContent parsing errors.
The constructors are package private as since this classes are used exclusively within the pre-packaged model (and require the specific weights, etc. to be of any use).
If a pipeline referenced by a transform does not exist, we should not allow the transform to be created.
We do allow the pipeline existence check to be skipped with defer_validations, but if the pipeline still does not exist on `_start`, the pipeline will fail to start.
relates: #50135
This test replaces the watch index after watcher got started.
This triggers watches being reloaded and while this happens the
trigger engine is paused, which disallows watches from being
triggered. At this time there are no watches in the .watches
index and I think this is just unlucky timing.
Reloading of watches happens in the background and
the watch state can be started when that happens.
For normal schedule trigger engines this is not an issue,
because watches that are meant to be triggered are triggered
when the engine triggers the next time. However for the
mock scheduled trigger engine this is different,
because watches are triggered programatically and
there is no retry in this test.
I think just adding `timeWarp().trigger("mywatch");` inside
a `assertBusy(...)`` is the right fix here. If it fails
because the mock schedule trigger engine is paused then
the test will try again. In the mean time the the watches
can be reloaded, which then resumes the mock scheduled trigger engine.
Closes#50658
Currently, if an updateable synonym filter is included in a multiplexer filter,
it is not reloaded via the _reload_search_analyzers because the multiplexer
itself doesn't pass on the analysis mode of the filters it contains, so its not
recognized as "updateable" in itself. Instead we can check and merge the
AnalysisMode settings of all filters in the multiplexer and use the resulting
mode (e.g. search-time only) for the multiplexer itself, thus making any synonym
filters contained in it reloadable. This, of course, will also make the
analyzers using the multiplexer be usable at search-time only.
Closes#50554
This commits makes the "init" ILM step retryable. It also adds a test
where an index is created with a non-parsable index name and then fails.
Related to #48183
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
This makes the "update-rollover-lifecycle-date" step, which is part of the
rollover action, retryable. It also adds an integration test to check the
step is retried and it eventually succeeds.
(cherry picked from commit 5bf068522deb2b6cd2563bcf80f34fdbf459c9f2)
Signed-off-by: Andrei Dan <andrei.dan@elastic.co>
In security we currently monitor a set of files for changes:
- config/role_mapping.yml (or alternative configured path)
- config/roles.yml
- config/users
- config/users_roles
This commit prevents unnecessary reloading when the file change actually doesn't change the internal structure.
Backport of: #50207
Co-authored-by: Anton Shuvaev <anton.shuvaev91@gmail.com>
vector REST tests occasionally fail on 7.x because
we don't receive the expected response headers with deprecation warnings.
This happens as searchers were executed against all indices including
internal indices, whose shards did not produce expected warnings.
This PR ensures that searchers are executed only against expected
indices.
Closes#50716
In 7.x an internal API used for validating remote cluster does not throw, see #50420 for the
details. This change implements a workaround for remote cluster validation, only for 7.x branches.
fixes#50420
Switch from a 32 bit Java hash to a 128 bit Murmur hash for
creating document IDs from by/over/partition field values.
The 32 bit Java hash was not sufficiently unique, and could
produce identical numbers for relatively common combinations
of by/partition field values such as L018/128 and L017/228.
Fixes#50613
The end offset of a tokenizer is supposed to point one past the
end of the input, not to the end character of the input. The
ml_classic tokenizer was erroneously doing the latter.
* Add aditional logging for ILM history store tests (#50624)
These tests use the same index name, making it hard to read logs when
diagnosing the failures. Additionally more information about the current
state of the index could be retrieved when failing.
This changes these two things in the hope of capturing more data about
why this fails on some CI nodes but not others.
Relates to #50353
This drops all remaining references to `BaseRestHandler.logger` which
has been deprecated for something like a year now. I replaced all of the
references with locally declared loggers which is so much less spooky
action at a distance to me.
Sharing a random generator may cause test failures as non-threadsafe random generators are periodically utilized in tests (see: https://github.com/elastic/elasticsearch/issues/50651)
This change constructs a calls `Randomness.get()` within the `bulkIndexWithRetry` method so that the returned `Random` object is only used in a single thread. Before, the member variable could have been used between threads, which caused test failures.
* [ML][Inference] lang_ident model (#50292)
This PR contains a java port of Google's CLD3 compact NN model https://github.com/google/cld3
The ported model is formatted to fit within our inference model formatting and stored as a resource in the `:xpack:ml:` plugin and is under basic license.
The model is broken up into two major parts:
- Preprocessing through the custom embedding (based on CLD3's embedding layer)
- Pushing the embedded text through the two layers of fully connected shallow NN.
Main differences between this port and CLD3:
- We take advantage of Java's internal Unicode handling where possible (i.e. codepoints, characters, decoders, etc.)
- We do not trim down input text by removing duplicated tokens
- We do not encode doubles/floats as longs/integers.
This adds a new cluster privilege `monitor_snapshot` which is a restricted
version of `create_snapshot`, granting the same privileges to view
snapshot and repository info and status but not granting the actual
privilege to create a snapshot.
Co-authored-by: j-bean <anton.shuvaev91@gmail.com>
Some changes had to be made in order to make the test pass due to the removal or types.
Added some more assertions. The failure description in this comment [0] indicates that the rest handler couldn't be found. The test passes now.
I plan to merge this into master and see how CI reacts, if it handles this change well then I will also unmute this test in 7 dot x branch.
Also check watch count after stopping watcher in test teardown and
disabled slm in smoke test watcher qa test.
Relates to #41172
0: https://github.com/elastic/elasticsearch/issues/41172#issuecomment-496993976
* 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.
*Most* of our parsing can be done without passing any extra context into
the parser that isn't already part of the xcontent stream. While I was
looking around at the places that *do* need a context I found a few
places that were declared to need a context but don't actually need it.
This adds support for retrying AsyncActionSteps by triggering the async
step after ILM was moved back on the failed step (the async step we'll
be attempting to run after the cluster state reflects ILM being moved
back on the failed step).
This also marks the RolloverStep as retryable and adds an integration
test where the RolloverStep is failing to execute as the rolled over
index already exists to test that the async action RolloverStep is
retried until the rolled over index is deleted.
(cherry picked from commit 8bee5f4cb58a1242cc2ef4bc0317dae6c8be49d3)
Signed-off-by: Andrei Dan <andrei.dan@elastic.co>
Adds a `force` parameter to the delete data frame analytics
request. When `force` is `true`, the action force-stops the
jobs and then proceeds to the deletion. This can be used in
order to delete a non-stopped job with a single request.
Closes#48124
Backport of #50553
We have about 800 `ObjectParsers` in Elasticsearch, about 700 of which
are final. This is *probably* the right way to declare them because in
practice we never mutate them after they are built. And we certainly
don't change the static reference. Anyway, this adds `final` to a bunch
of these parsers, mostly the ones in xpack and their "paired" parsers in
the high level rest client. I picked these just to have somewhere to
break the up the change so it wouldn't be huge.
I found the non-final parsers with this:
```
diff \
<(find . -type f -name '*.java' -exec grep -iHe 'static.*PARSER\s*=' {} \+ | sort) \
<(find . -type f -name '*.java' -exec grep -iHe 'static.*final.*PARSER\s*=' {} \+ | sort) \
2>&1 | grep '^<'
```
Eclipse 4.13 shows a type mismatch error in the affected line because it cannot
correctly infer the boolean return type for the method call. Assigning return
value to a local variable resolves this problem.
XPackPlugin created an SSLService within the plugin contructor.
This has 2 negative consequences:
1. The service may be constructed based on a partial view of settings.
Other plugins are free to add setting values via the
additionalSettings() method, but this (necessarily) happens after
plugins have been constructed.
2. Any exceptions thrown during the plugin construction are handled
differently than exceptions thrown during "createComponents".
Since SSL configurations exceptions are relatively common, it is
far preferable for them to be thrown and handled as part of the
createComponents flow.
This commit moves the creation of the SSLService to
XPackPlugin.createComponents, and alters the sequence of some other
steps to accommodate this change.
Backport of: #49667
PR #44238 changed several links related to the Elasticsearch search request body API. This updates several places still using outdated links or anchors.
This will ultimately let us remove some redirects related to those link changes.
The docs/reference/redirects.asciidoc file stores a list of relocated or
deleted pages for the Elasticsearch Reference documentation.
This prunes several older redirects that are no longer needed and
don't require work to fix broken links in other repositories.
We are matching on the exact number of shards in this test, but may run into
snapshotting more than the single index created in it due to auto-created indices like
`.watcher`.
Fixed by making the test only take a snapshot of the single index used by this test.
Closes#50450
* 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
refactors source and dest validation, adds support for CCS, makes resolve work like reindex/search, allow aliased dest index with a single write index.
fixes#49988fixes#49851
relates #43201
Previously, during expression optimisation, CAST would be considered
nullable if the casted expression resulted to a NULL literal, and would
be always non-nullable otherwise. As a result if CASE was wrapped by a
null check function like IS NULL or IS NOT NULL it was simplified to
TRUE/FALSE, eliminating the actual casting operation. So in case of an
expression with an erroneous casting like CAST('foo' AS DATETIME) IS NULL
it would be simplified to FALSE instead of throwing an Exception signifying
the attempt to cast 'foo' to a DATETIME type.
CAST now always returns Nullability.UKNOWN except from the case that
its result evaluated to a constant NULL, where it returns Nullability.TRUE.
This way the IS NULL/IS NOT NULL don't get simplified to FALSE/TRUE
and the CAST actually gets evaluated resulting to a thrown Exception.
Fixes: #50191
(cherry picked from commit 671e07a931cd828661e226cba22a5d38804a17a5)
* Update remote cluster stats to support simple mode (#49961)
Remote cluster stats API currently only returns useful information if
the strategy in use is the SNIFF mode. This PR modifies the API to
provide relevant information if the user is in the SIMPLE mode. This
information is the configured addresses, max socket connections, and
open socket connections.
* Send hostname in SNI header in simple remote mode (#50247)
Currently an intermediate proxy must route conncctions to the
appropriate remote cluster when using simple mode. This commit offers
a additional mechanism for the proxy to route the connections by
including the hostname in the TLS SNI header.
* Rename the remote connection mode simple to proxy (#50291)
This commit renames the simple connection mode to the proxy connection
mode for remote cluster connections. In order to do this, the mode specific
settings which we namespaced by their mode (ex: sniff.seed and
proxy.addresses) have been reverted.
* Modify proxy mode to support a single address (#50391)
Currently, the remote proxy connection mode uses a list setting for the
proxy address. This commit modifies this so that the setting is
proxy_address and only supports a single remote proxy address.
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
In order to ensure any persisted model state is searchable by the moment
the job reports itself as `stopped`, we need to refresh the state index
before completing.
This should fix the occasional failures we see in #50168 and #50313 where
the model state appears missing.
Closes#50168Closes#50313
Backport of #50322
This fixes support for nested fields
We now support fully nested, fully collapsed, or a mix of both on inference docs.
ES mappings allow the `_source` to be any combination of nested objects + dot delimited fields.
So, we should do our best to find the best path down the Map for the desired field.
Our REST infrastructure will reject requests that have a body where the
body of the request is never consumed. This ensures that we reject
requests on endpoints that do not support having a body. This requires
cooperation from the REST handlers though, to actually consume the body,
otherwise the REST infrastructure will proceed with rejecting the
request. This commit addresses an issue in the has privileges API where
we would prematurely try to reject a request for not having a username,
before consuming the body. Since the body was not consumed, the REST
infrastructure would instead reject the request as a bad request.
This commit adds removal of unused data frame analytics state
from the _delete_expired_data API (and in extend th ML daily
maintenance task). At the moment the potential state docs
include the progress document and state for regression and
classification analyses.
Backport of #50243
Backport #50244 to 7.x branch.
If a processor executes asynchronously and the ingest simulate api simulates with
multiple documents then the order of the documents in the response may not match
the order of the documents in the request.
Alexander Reelsen discovered this issue with the enrich processor with the following reproduction:
```
PUT cities/_doc/munich
{"zip":"80331","city":"Munich"}
PUT cities/_doc/berlin
{"zip":"10965","city":"Berlin"}
PUT /_enrich/policy/zip-policy
{
"match": {
"indices": "cities",
"match_field": "zip",
"enrich_fields": [ "city" ]
}
}
POST /_enrich/policy/zip-policy/_execute
GET _cat/indices/.enrich-*
POST /_ingest/pipeline/_simulate
{
"pipeline": {
"processors" : [
{
"enrich" : {
"policy_name": "zip-policy",
"field" : "zip",
"target_field": "city",
"max_matches": "1"
}
}
]
},
"docs": [
{ "_id": "first", "_source" : { "zip" : "80331" } } ,
{ "_id": "second", "_source" : { "zip" : "50667" } }
]
}
```
* fixed test compile error
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.
This adds a new "xpack.license.upload.types" setting that restricts
which license types may be uploaded to a cluster.
By default all types are allowed (excluding basic, which can only be
generated and never uploaded).
This setting does not restrict APIs that generate licenses such as the
start trial API.
This setting is not documented as it is intended to be set by
orchestrators and not end users.
Backport of: #49418
In the yaml cluster upgrade tests, we start a scroll in a mixed-version cluster,
then attempt to continue the scroll after the upgrade is complete. This test
occasionally fails because the scroll can expire before the cluster is done
upgrading.
The current scroll keep-alive time 5m. This PR bumps it to 10m, which gives a
good buffer since in failing tests the time was only exceeded by ~30 seconds.
Addresses #46529.
Backport of #49612.
The current Docker entrypoint script picks up environment variables and
translates them into -E command line arguments. However, since any tool
executes via `docker exec` doesn't run the entrypoint, it results in
a poorer user experience.
Therefore, refactor the env var handling so that the -E options are
generated in `elasticsearch-env`. These have to be appended to any
existing command arguments, since some CLI tools have subcommands and
-E arguments must come after the subcommand.
Also extract the support for `_FILE` env vars into a separate script, so
that it can be called from more than once place (the behaviour is
idempotent).
Finally, add noop -E handling to CronEvalTool for parity, and support
`-E` in MultiCommand before subcommands.
The testclusters shutdown code was killing child processes
of the ES JVM before the ES JVM. This causes any running
ML jobs to be recorded as failed, as the ES JVM notices that
they have disconnected from it without being told to stop,
as they would if they crashed. In many test suites this
doesn't matter because the test cluster will never be
restarted, but in the case of upgrade tests it makes it
impossible to test what happens when an ML job is running
at the time of the upgrade.
This change reverses the order of killing the ES process
tree such that the parent processes are killed before their
children. A list of children is stored before killing the
parent so that they can subsequently be killed (if they
don't exit by themselves as a side effect of the parent
dying).
Backport of #50175
Executing the data frame analytics _explain API with a config that contains
a field that is not in the includes list but at the same time is the excludes
list results to trying to remove the field twice from the iterator. That causes
an `IllegalStateException`. This commit fixes this issue and adds a test that
captures the scenario.
Backport of #50192
* 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
Lucene 8.4 added support for "CONTAINS", therefore in this commit those
changes are integrated in Elasticsearch. This commit contains as well a
bug fix when querying with a geometry collection with "DISJOINT" relation.
The release builds use a production license key, and our rest test load
licenses that are signed by the dev license key.
This change adds the new enterprise license Rest tests to the
blacklist on release builds.
Backport of: #50163
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>
This adds a new field for the inference processor.
`warning_field` is a place for us to write warnings provided from the inference call. When there are warnings we are not going to write an inference result. The goal of this is to indicate that the data provided was too poor or too different for the model to make an accurate prediction.
The user could optionally include the `warning_field`. When it is not provided, it is assumed no warnings were desired to be written.
The first of these warnings is when ALL of the input fields are missing. If none of the trained fields are present, we don't bother inferencing against the model and instead provide a warning stating that the fields were missing.
Also, this adds checks to not allow duplicated fields during processor creation.
This exchanges the direct use of the `Client` for `ResultsPersisterService`. State doc persistence will now retry. Failures to persist state will still not throw, but will be audited and logged.
This commit fixes a bug that caused the data frame analytics
_explain API to time out in a multi-node setup when the source
index was missing. When we try to create the extracted fields detector,
we check the index settings. If the index is missing that responds
with a failure that could be wrapped as a remote exception.
While we unwrapped correctly to check if the cause was an
`IndexNotFoundException`, we then proceeded to cast the original
exception instead of the cause.
Backport of #50176
This test was fixed as part of #49736 so that it used a
TokenService mock instance that was enabled, so that token
verification fails because the token is invalid and not because
the token service is not enabled.
When the randomly generated token we send, decodes to being of
version > 7.2 , we need to have mocked a GetResponse for the call
that TokenService#getUserTokenFromId will make, otherwise this
hangs and times out.
Today the HTTP exporter settings without the exporter type having been
configured to HTTP. When it is time to initialize the exporter, we can
blow up. Since this initialization happens on the cluster state applier
thread, it is quite problematic that we do not reject settings updates
where the type is not configured to HTTP, but there are HTTP exporter
settings configured. This commit addresses this by validating that the
exporter type is not only set, but is set to HTTP.
The "code_user" and "code_admin" reserved roles existed to support
code search which is no longer included in Kibana.
The "kibana_system" role included privileges to read/write from the
code search indices, but no longer needs that access.
Backport of: #50068
* [ML] Add graceful retry for anomaly detector result indexing failures (#49508)
All results indexing now retry the amount of times configured in `xpack.ml.persist_results_max_retries`. The retries are done in a semi-random, exponential backoff.
* fixing test
This adds "enterprise" as an acceptable type for a license loaded
through the PUT _license API.
Internally an enterprise license is treated as having a "platinum"
operating mode.
The handling of License types was refactored to have a new explicit
"LicenseType" enum in addition to the existing "OperatingMode" enum.
By default (in 7.x) the GET license API will return "platinum" when an
enterprise license is active in order to be compatible with existing
consumers of that API.
A new "accept_enterprise" flag has been introduced to allow clients to
opt-in to receive the correct "enterprise" type.
Backport of: #49223
The `sparse_vector` REST tests occasionally fail on 7.x because we don't receive the expected response headers with deprecation warnings.
One theory as to what is happening is that there is an extra empty index present in addition to the test index. Since the search doesn't specify an index name, it hits both the test index and this extra empty index and shard responses from the extra index don't produce deprecation warnings. If not all shard responses contain the warning headers, then certain deprecation warnings can be lost (due to the bug described in #33936).
This PR tries to harden the `sparse_vector` tests by always specifying the index name during a search. This doesn't fix the root causes of the issue, but is good practice and can help avoid intermittent failures.
Addresses #49383.
The `ClassificationIT.testTwoJobsWithSameRandomizeSeedUseSameTrainingSet`
test was previously set up to just have 10 rows. With `training_percent`
of 50%, only 5 rows will be used for training. There is a good chance that
all 5 rows will be of one class which results to failure.
This commit increases the rows to 100. Now 50 rows should be used for training
and the chance of failure should be very small.
Backport of #50072
Watcher logs when actions fail in ActionWrapper, but failures to
generate an email attachment are not logged and we thus only know the
type of the exception and not where/how it occurred.
Adjusts the subclasses of `TransportMasterNodeAction` to use their own loggers
instead of the one for the base class.
Relates #50056.
Partial backport of #46431 to 7.x.
Return a 401 in all cases when a request is submitted with an
access token that we can't consume. Before this change, we would
throw a 500 when a request came in with an access token that we
had generated but was then invalidated/expired and deleted from
the tokens index.
Resolves: #38866
Backport of #49736
This adds a new `randomize_seed` for regression and classification.
When not explicitly set, the seed is randomly generated. One can
reuse the seed in a similar job in order to ensure the same docs
are picked for training.
Backport of #49990
The elasticsearch-node tools allow manipulating the on-disk cluster state. The tool is currently
unaware of plugins and will therefore drop custom metadata from the cluster state once the
state is written out again (as it skips over the custom metadata that it can't read). This commit
preserves unknown customs when editing on-disk metadata through the elasticsearch-node
command-line tools.
Today settings can declare dependencies on another setting. This
declaration is implemented so that if the declared setting is not set
when the declaring setting is, settings validation fails. Yet, in some
cases we want not only that the setting is set, but that it also has a
specific value. For example, with the monitoring exporter settings, if
xpack.monitoring.exporters.my_exporter.host is set, we not only want
that xpack.monitoring.exporters.my_exporter.type is set, but that it is
also set to local. This commit extends the settings infrastructure so
that this declaration is possible. The use of this in the monitoring
exporter settings will be implemented in a follow-up.
The `testReplaceChildren()` has been fixed for Pivot as part
of #49693.
Reverting: #49045
(cherry picked from commit 4b9b9edbcf2041a8619b65580bbe192bf424cebc)
When checking the cardinality of a field, the query should be take into account. The user might know about some bad data in their index and want to filter down to the target_field values they care about.
Work in progress in the c++ side is increasing memory estimates
a bit and this test fails. At the time of this commit the mem
estimate when there is no source query is a about 2Mb. So I
am relaxing the test to assert memory estimate is less than 1Mb
instead of 500Kb.
Backport of #49924
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.
Makes sure that CCR also properly works with _source disabled.
Changes one exception in LuceneChangesSnapshot as the case of missing _recovery_source
because of a missing lease was not properly properly bubbled up to CCR (testIndexFallBehind
was failing).
To recap, Attributes form the properties of a derived table.
Each LogicalPlan has Attributes as output since each one can be part of
a query and as such its result are sent to its consumer.
This change essentially removes the name id comparison so any changes
applied to existing expressions should work as long as the said
expressions are semantically equivalent.
This change enforces the hashCode and equals which has the side-effect
of using hashCode as identifiers for each expression.
By removing any property from an Attribute, the various components need
to look the original source for comparison which, while annoying, should
prevent a reference from getting out of sync with its source due to
optimizations.
Essentially going forward there are only 3 types of NamedExpressions:
Alias - user define (implicit or explicit) name
FieldAttribute - field backed by Elasticsearch
ReferenceAttribute - a reference to another source acting as an
Attribute. Typically the Attribute of an Alias.
* Remove the usage of NamedExpression as basis for all Expressions.
Instead, restrict their use only for named context, such as projections
by using Aliasing instead.
* Remove different types of Attributes and allow only FieldAttribute,
UnresolvedAttribute and ReferenceAttribute. To avoid issues with
rewrites, resolve the references inside the QueryContainer so the
information always stays on the source.
* Side-effect, simplify the rules as the state for InnerAggs doesn't
have to be contained anymore.
* Improve ResolveMissingRef rule to handle references to named
non-singular expression tree against the same expression used up the
tree.
#49693 backport to 7.x
(cherry picked from commit 5d095e2173bcbf120f534a6f2a584185a7879b57)
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**
This commit refactors the `IndexLifecycleRunner` to split out and
consolidate the number of methods that change state from within ILM. It
adds a new class `IndexLifecycleTransition` that contains a number of
static methods used to modify ILM's state. These methods all return new
cluster states rather than making changes themselves (they can be
thought of as helpers for modifying ILM state).
Rather than having multiple ways to move an index to a particular step
(like `moveClusterStateToStep`, `moveClusterStateToNextStep`,
`moveClusterStateToPreviouslyFailedStep`, etc (there are others)) this
now consolidates those into three with (hopefully) useful names:
- `moveClusterStateToStep`
- `moveClusterStateToErrorStep`
- `moveClusterStateToPreviouslyFailedStep`
In the move, I was also able to consolidate duplicate or redundant
arguments to these functions. Prior to this commit there were many calls
that provided duplicate information (both `IndexMetaData` and
`LifecycleExecutionState` for example) where the duplicate argument
could be derived from a previous argument with no problems.
With this split, `IndexLifecycleRunner` now contains the methods used to
actually run steps as well as the methods that kick off cluster state
updates for state transitions. `IndexLifecycleTransition` contains only
the helpers for constructing new states from given scenarios.
This also adds Javadocs to all methods in both `IndexLifecycleRunner`
and `IndexLifecycleTransition` (this accounts for almost all of the
increase in code lines for this commit). It also makes all methods be as
restrictive in visibility, to limit the scope of where they are used.
This refactoring is part of work towards capturing actions and
transitions that ILM makes, by consolidating and simplifying the places
we make state changes, it will make adding operation auditing easier.
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
* 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.
Some extended testing with MS-SQL server and H2 (which agree on
results) revealed bugs in the implementation of WEEK related extraction
and diff functions.
Non-iso WEEK seems to be broken since #48209 because
of the replacement of Calendar and the change in the ISO rules.
ISO_WEEK failed for some edge cases around the January 1st.
DATE_DIFF was previously based on non-iso WEEK extraction which seems
not to be the case.
Fixes: #49376
(cherry picked from commit 54fe7f57289c46bb0905b1418f51a00e8c581560)
This stems from a time where index requests were directly forwarded to
TransportReplicationAction. Nowadays they are wrapped in a BulkShardRequest, and this logic is
obsolete.
In contrast to prior PR (#49647), this PR also fixes (see b3697cc) a situation where the previous
index expression logic had an interesting side effect. For bulk requests (which had resolveIndex
= false), the reroute phase was waiting for the index to appear in case where it was not present,
and for all other replication requests (resolveIndex = true) it would right away throw an
IndexNotFoundException while resolving the name and exit. With #49647, every replication
request was now waiting for the index to appear, which was problematic when the given index
had just been deleted (e.g. deleting a follower index while it's still receiving requests from the
leader, where these requests would now wait up to a minute for the index to appear). This PR
now adds b3697cc on top of that prior PR to make sure to reestablish some of the prior behavior
where the reroute phase waits for the bulk request for the index to appear. That logic was in
place to ensure that when an index was created and not all nodes had learned about it yet, that
the bulk would not fail somewhere in the reroute phase. This is now only restricted to the
situation where the current node has an older cluster state than the one that coordinated the
bulk request (which checks that the index is present). This also means that when an index is
deleted, we will no longer unnecessarily wait up to the timeout for the index o appear, and
instead fail the request.
Closes#20279
This adds a `_source` setting under the `source` setting of a data
frame analytics config. The new `_source` is reusing the structure
of a `FetchSourceContext` like `analyzed_fields` does. Specifying
includes and excludes for source allows selecting which fields
will get reindexed and will be available in the destination index.
Closes#49531
Backport of #49690
* 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
- Improves HTTP client hostname verification failure messages
- Adds "DiagnosticTrustManager" which logs certificate information
when trust cannot be established (hostname failure, CA path failure,
etc)
These diagnostic messages are designed so that many common TLS
problems can be diagnosed based solely (or primarily) on the
elasticsearch logs.
These diagnostics can be disabled by setting
xpack.security.ssl.diagnose.trust: false
Backport of: #48911
Add a mirror of the maven repository of the shibboleth project
and upgrade opensaml and related dependencies to the latest
version available version
Resolves: #44947
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
This stems from a time where index requests were directly forwarded to
TransportReplicationAction. Nowadays they are wrapped in a BulkShardRequest, and this logic is
obsolete.
Closes#20279
This change adds a dynamic cluster setting named `indices.id_field_data.enabled`.
When set to `false` any attempt to load the fielddata for the `_id` field will fail
with an exception. The default value in this change is set to `false` in order to prevent
fielddata usage on this field for future versions but it will be set to `true` when backporting
to 7x. When the setting is set to true (manually or by default in 7x) the loading will also issue
a deprecation warning since we want to disallow fielddata entirely when https://github.com/elastic/elasticsearch/issues/26472
is implemented.
Closes#43599
Authentication has grown more complex with the addition of new realm
types and authentication methods. When user authentication does not
behave as expected it can be difficult to determine where and why it
failed.
This commit adds DEBUG and TRACE logging at key points in the
authentication flow so that it is possible to gain addition insight
into the operation of the system.
Backport of: #49575
The AuthenticationService has a feature to "smart order" the realm
chain so that whicherver realm was the last one to successfully
authenticate a given user will be tried first when that user tries to
authenticate again.
There was a bug where the building of this realm order would
incorrectly drop the first realm from the default chain unless that
realm was the "last successful" realm.
In most cases this didn't cause problems because the first realm is
the reserved realm and so it is unusual for a user that authenticated
against a different realm to later need to authenticate against the
resevered realm.
This commit fixes that bug and adds relevant asserts and tests.
Backport of: #49473
All the implementations of `EsBlobStoreTestCase` use the exact same
bootstrap code that is also used by their implementation of
`EsBlobStoreContainerTestCase`.
This means all tests might as well live under `EsBlobStoreContainerTestCase`
saving a lot of code duplication. Also, there was no HDFS implementation for
`EsBlobStoreTestCase` which is now automatically resolved by moving the tests over
since there is a HDFS implementation for the container tests.
Grouping By YEAR() is translated to a histogram aggregation, but
previously if there was a scalar function invloved (e.g.:
`YEAR(date + INTERVAL 2 YEARS)`), there was no proper script created
and the histogram was applied on a field with name: `date + INTERVAL 2 YEARS`
which doesn't make sense, and resulted in null result.
Check the underlying field of YEAR() and if it's a function call
`asScript()` to properly get the painless script on which the histogram
is applied.
Fixes: #49386
(cherry picked from commit 93c37abc943d00d3a14ba08435d118a6d48874c7)
Add TRUNC as alias to already implemented TRUNCATE
numeric function which is the flavour supported by
Oracle and PostgreSQL.
Relates to: #41195
(cherry picked from commit f2aa7f0779bc5cce40cc0c1f5e5cf1a5bb7d84f0)
We depend on the number of data frame rows in order to report progress
for the writing of results, the last phase of a job run. However, results
include other objects than just the data frame rows (e.g, progress, inference model, etc.).
The problem this commit fixes is that if we receive the last data frame row results
we'll report that progress is complete even though we still have more results to process
potentially. If the job gets stopped for any reason at this point, we will not be able
to restart the job properly as we'll think that the job was completed.
This commit addresses this by limiting the max progress we can report for the
writing_results phase before the results processor completes to 98.
At the end, when the process is done we set the progress to 100.
The commit also improves failure capturing and reporting in the results processor.
Backport of #49551
Previously, CaseProcessor was pre-calculating (called `process()`)
on all the building elements of a CASE/IIF expression, not only the
conditions involved but also the results, as well as the final else result.
In case one of those results had an erroneous calculation
(e.g.: division by zero) this was executed and resulted in
an Exception to be thrown, even if this result was not used because of
the condition guarding it. e.g.:
```
SELECT CASE myField1 = 0 THEN NULL ELSE myField2 / myField1 END
FROM test;
```
Fixes: #49388
(cherry picked from commit dbd169afc98686cae1bc72024fad0ca32b272efd)
This commit back ports three commits related to enabling the simple
connection strategy.
Allow simple connection strategy to be configured (#49066)
Currently the simple connection strategy only exists in the code. It
cannot be configured. This commit moves in the direction of allowing it
to be configured. It introduces settings for the addresses and socket
count. Additionally it introduces new settings for the sniff strategy
so that the more generic number of connections and seed node settings
can be deprecated.
The simple settings are not yet registered as the registration is
dependent on follow-up work to validate the settings.
Ensure at least 1 seed configured in remote test (#49389)
This fixes#49384. Currently when we select a random subset of seed
nodes from a list, it is possible for 0 seeds to be selected. This test
depends on at least 1 seed being selected.
Add the simple strategy to cluster settings (#49414)
This is related to #49067. This commit adds the simple connection
strategy settings and strategy mode setting to the cluster settings
registry. With these changes, the simple connection mode can be used.
Additionally, it adds validation to ensure that settings cannot be
misconfigured.
The categorization job wizard in the ML UI will use this
information when showing the effect of the chosen categorization
analyzer on a sample of input.
Before this change excluding an unsupported field resulted in
an error message that explained the excluded field could not be
detected as if it doesn't exist. This error message is confusing.
This commit commit changes this so that there is no error in this
scenario. When excluding a field that does exist but has been
automatically been excluded from the analysis there is no harm
(unlike excluding a missing field which could be a typo).
Backport of #49535
This test must check for state `SUCCESS` as well. `SUCESS` in
`SnapshotsInProgress` means "all data nodes finished snapshotting sucessfully but master must still finalize the snapshot in the repo".
`SUCESS` does not mean that the snapshot is actually fully finished in this object.
You can easily reporduce the scenario in #49303 that has an in-progress snapshot in `SUCCESS` state
by waiting 20s before running the busy assert loop on the snapshot status so that all steps but the blocked
finalization can finish.
Closes#49303
This commit replaces the _estimate_memory_usage API with
a new API, the _explain API.
The API consolidates information that is useful before
creating a data frame analytics job.
It includes:
- memory estimation
- field selection explanation
Memory estimation is moved here from what was previously
calculated in the _estimate_memory_usage API.
Field selection is a new feature that explains to the user
whether each available field was selected to be included or
not in the analysis. In the case it was not included, it also
explains the reason why.
Backport of #49455
This commit enhances the required pipeline functionality by changing it
so that default/request pipelines can also be executed, but the required
pipeline is always executed last. This gives users the flexibility to
execute their own indexing pipelines, but also ensure that any required
pipelines are also executed. Since such pipelines are executed last, we
change the name of required pipelines to final pipelines.
Add extra checks to prevent ConstantFolding rule to try to fold
the CASE/IIF functions early before the SimplifyCase rule gets applied.
Fixes: #49387
(cherry picked from commit f35c9725350e35985d8dd3001870084e1784a5ca)
This change automatically pre-sort search shards on search requests that use a primary sort based on the value of a field. When possible, the can_match phase will extract the min/max (depending on the provided sort order) values of each shard and use it to pre-sort the shards prior to running the subsequent phases. This feature can be useful to ensure that shards that contain recent data are executed first so that intermediate merge have more chance to contain contiguous data (think of date_histogram for instance) but it could also be used in a follow up to early terminate sorted top-hits queries that don't require the total hit count. The latter could significantly speed up the retrieval of the most/least
recent documents from time-based indices.
Relates #49091
This commit adds a deprecation warning when starting
a node where either of the server contexts
(xpack.security.transport.ssl and xpack.security.http.ssl)
meet either of these conditions:
1. The server lacks a certificate/key pair (i.e. neither
ssl.keystore.path not ssl.certificate are configured)
2. The server has some ssl configuration, but ssl.enabled is not
specified. This new validation does not care whether ssl.enabled is
true or false (though other validation might), it simply makes it
an error to configure server SSL without being explicit about
whether to enable that configuration.
Backport of: #45892
Backport of #48277
Otherwise integration tests may fail if the monitoring interval is low:
```
[2019-10-21T09:57:25,527][ERROR][o.e.b.ElasticsearchUncaughtExceptionHandler] [integTest-0] fatal error in thread [elasticsearch[integTest-0][generic][T#4]], exiting
java.lang.AssertionError: initial cluster state not set yet
at org.elasticsearch.cluster.service.ClusterApplierService.state(ClusterApplierService.java:208) ~[elasticsearch-7.6.0-SNAPSHOT.jar:7.6.0-SNAPSHOT]
at org.elasticsearch.cluster.service.ClusterService.state(ClusterService.java:125) ~[elasticsearch-7.6.0-SNAPSHOT.jar:7.6.0-SNAPSHOT]
at org.elasticsearch.xpack.monitoring.MonitoringService$MonitoringExecution$1.doRun(MonitoringService.java:231) ~[?:?]
at org.elasticsearch.common.util.concurrent.AbstractRunnable.run(AbstractRunnable.java:37) ~[elasticsearch-7.6.0-SNAPSHOT.jar:7.6.0-SNAPSHOT]
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515) ~[?:?]
at java.util.concurrent.FutureTask.run(FutureTask.java:264) ~[?:?]
at org.elasticsearch.common.util.concurrent.ThreadContext$ContextPreservingRunnable.run(ThreadContext.java:703) ~[elasticsearch-7.6.0-SNAPSHOT.jar:7.6.0-SNAPSHOT]
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128) ~[?:?]
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628) ~[?:?]
at java.lang.Thread.run(Thread.java:835) [?:?]
```
I ran into this when lowering the monitoring interval when investigating
enrich monitoring test: #48258
fix force stopping transform if indexer state hasn't been written and/or is set to STOPPED. In certain situations the transform could not be stopped, which means the task could not be removed. Introduces improved abstraction in order to better test state handling in future.
* SLM set the operation mode to RUNNING on first run
Set the SLM operation mode to RUNNING when setting the first SLM lifecycle
policy. Historically, SLM was not decoupled from ILM but now they are
independent components. Setting the SLM operation mode to what the ILM running
mode was when we set the first SLM lifecycle policy was a remain from those
times.
* SLM update package info
* SLM suppress unusued warning
* SLM use logger for the correct class
* SLM Add integration test for operation mode
* Use ESSingleNodeTestCase instead of ESIntegTestCase
(cherry picked from commit 4ad3d93f89d03bf9a25685a990d1a439f33ce0e6)
Signed-off-by: Andrei Dan <andrei.dan@elastic.co>
If a datafeed is stopped normally and force stopped at the same
time then it is possible that the force stop removes the
persistent task while the normal stop is performing actions.
Currently this causes the normal stop to error, but since
stopping a stopped datafeed is not an error this doesn't make
sense. Instead the force stop should just take precedence.
This is a followup to #49191 and should really have been
included in the changes in that PR.
Albert Zaharovits
Tim Vernum
Benjamin Trent
Lee Hinman
Julie Tibshirani
Dimitris Athanasiou
Nik Everett
Jake Landis
Sean Story
Martijn van Groningen
Ioannis Kakavas
Christoph Büscher
Nhat Nguyen
Adrien Grand
Armin Braun
Andrei Dan
Mayya Sharipova
Hendrik Muhs
Przemysław Witek
David Roberts
Lisa Cawley
James Rodewig
Orhan Toy
Igor Motov
Marios Trivyzas
Tim Brooks
Stuart Tettemer
Jason Tedor
Yannick Welsch
Florian Kelbert
David Kyle
Andrei Stefan
Rory Hunter
Ignacio Vera
Henning Andersen
David Turner
Stuart Cam
cachedout
Costin Leau
Zachary Tong
Jim Ferenczi
Przemyslaw Gomulka
j-bean
debadair
István Zoltán Szabó
Bogdan Pintea