Adds tests similar to `:qa:full-cluster-restart` for x-pack. You
run them with `gradle :x-pack:qa:full-cluster-restart:check`.
The actual tests are as basic as it gets: create a doc and load it,
shut down, upgrade to master, startup, and load it. Create a user
and load it, shut down, upgrade to master, startup, and load it.
Relates to elastic/x-pack-elasticsearch#1629
Original commit: elastic/x-pack-elasticsearch@8994bec8e7
This commit removes unnecessary initialization of the system key in tests that no longer make use
of the system key. It also removes the feature usage for the system key in the SecurityFeatureSet.
Original commit: elastic/x-pack-elasticsearch@b9fffe0bd3
Logstash now has ephemeral id at the instance level and also at the
pipeline level, we need to add them to the logstash monitoring template.
Original commit: elastic/x-pack-elasticsearch@dfac702d59
* Use bulk request to persist model plots and model size stats
* Revert persisting model size stats in the bulk request
* Refactor results persister
Original commit: elastic/x-pack-elasticsearch@f51297bfc2
We recently added logging for critical authentication failures as they had previously been silent (with respect to logs) but would cause authentication processing to stop.
However, the reserved realm intentionally uses exceptions to stop any other realm authenticating a reserved user if the password is entered incorrectly.
Since this is the most common use of exceptions in the authc chain, we reduce the logging verbosity in normal cases (drop the stack trace, remove "unexpected") and only log the full details in debug.
Original commit: elastic/x-pack-elasticsearch@686a98010b
This is related to elastic/x-pack-elasticsearch#1217. This change introduces a tool
bin/x-pack/setup-passwords that will streamline the setting of
internal user passwords. There are two modes of operation. One mode
called auto, automatically generates passwords and prints them to
the console. The second mode called interactive allows the user to
enter passwords.
All passwords are changed using the elastic superuser. The elastic
password is the first password to be set.
Original commit: elastic/x-pack-elasticsearch@00974234a2
After improving the authorization of scroll requests and backporting to 5.x, we no longer need to
have any signing code in master. This commit removes it.
Original commit: elastic/x-pack-elasticsearch@8b65fd9338
If an exception occurs while sending the initial setup messages to the autodetect
such that it fails rather than reaching the open state then the autodetect process
needs to be killed to prevent it hogging resources.
Relates elastic/x-pack-elasticsearch#1684
Original commit: elastic/x-pack-elasticsearch@1ee80ed9b0
In does not make sense for the time_field in the data_description to
be used as a by/over/partition field name, nor the summary_count_field,
categorization_field or as an influencer. Therefore, configurations
where the time_field in the data_description is used in the
analysis_config are now rejected.
Additionally, it causes a problem communicating with the C++ code if
the control field name (which is '.') is used in the analysis_config,
so this is also rejected at the validation stage.
Relates elastic/x-pack-elasticsearch#1684
Original commit: elastic/x-pack-elasticsearch@e6750a2cda
- Don't attempt to upgrade from 2.x
- Attempt up to 10 retries if the migration fails (with increasing back-off between attempts)
- If a cached user is disabled, recheck with the underlying store
The last change is required if the migration takes a long time.
While users are being migrated, they might be marked as disabled, but when the migration is complete they need to be usable immediately.
Original commit: elastic/x-pack-elasticsearch@2621867014
Removes the `assemble` task from projects that aren't published
to speed up `gradle assemble` so the unified release can call it.
Original commit: elastic/x-pack-elasticsearch@43dfcc15f3
These tests are starting their own nodes and do not use the testing
trigger schedule class.
There are occasional test failure due to a race condition where watcher
is in the process of being started, but cannot be shut down properly,
because starting up was not finished when the shut down was called for.
These filter tests do not rely on watcher, so we can disable them for
now, but we still need to fix a race condition in starting/stopping
watcher.
relates elastic/x-pack-elasticsearch#1422
Original commit: elastic/x-pack-elasticsearch@f13bb7a6fb
There is no need to handle any _status field in
the 6.0 release from now on, as everything has been
taken care in the upgrade API.
Original commit: elastic/x-pack-elasticsearch@606581f4d7
This changes part of the logic that was added in elastic/x-pack-elasticsearch#644 and extended
in elastic/x-pack-elasticsearch#1495 so that when ML is disabled we never try to communicate with
the native controller during node shutdown.
The original reason for needing to communicate with the native controller
when ML is disabled was the problem of elastic/prelert-legacy#803.
However, this was fixed in a better way in elastic/elasticsearch#24579.
Now there is considerable benefit in never talking to the native
controller from the plugin code when ML is disabled, because it means
anyone suffering some obscure problem with ML can disable it without
running the risk of uncovering some other obscure problem with shutdown.
Original commit: elastic/x-pack-elasticsearch@9d329483a7
This adds a check in the REST tests to ensure that
watcher is started, and if not, tries to start watcher.
This eliminates test failures where watcher was not in
the correct state due to other tests stopping watcher.
Original commit: elastic/x-pack-elasticsearch@fc547d49b4
This commit adds a new Logstash component to x-pack to support the config management work. Currently, the functionality in this component is really simple; all it does is upload a new index template for `.logstash` index. This index stores the actual LS configuration.
On this template is bootstrapped in ES, Kibana can write user-created LS configs which adhere to the mapping defined here. In the future, we're looking into adding more functionality on the ES side to handle config documents, but for now, this is simple.
relates elastic/x-pack-elasticsearch#1499, relates elastic/x-pack-elasticsearch#1471
Original commit: elastic/x-pack-elasticsearch@d7cc8675f7
In the case where a field is a text multi-field, it has
no doc values and it is not in source. Thus, the datafeed
will not be able to extract it.
However, it is possible to extract it by getting its parent
field instead. This commit implements the logic to look
in parent fields when the field in question is a text field.
Original commit: elastic/x-pack-elasticsearch@f116e89921
In 5.4.x, the datafeed attempts to get all fields from
doc_values by default. It has a `_source` parameter which
when enabled changes the strategy to instead try to get
all fields from the source.
This has been the most common issue users have been
reporting as it means the datafeed will fail to fetch
any text fields by default.
This change uses the field capabilities API in order
to automatically detect whether a field is aggregatable.
It then extracts such fields from doc_values while the
rest are taken from source. The change also adds
validation to the start datafeed action so that if
fields are missing mappings or the time field is not
aggregatable we respond with an appropriate error.
relates elastic/x-pack-elasticsearch#1649
Original commit: elastic/x-pack-elasticsearch@76e2cc6cb2
Specifying types for a datafeed should be optional
as no types is equal to searching through all types.
Original commit: elastic/x-pack-elasticsearch@f61ac01b45
Only log an entry when an actual stop is executed instead of
always logging.
Also added a reason to stop watcher to the methods, so that
debug logs will yield that information.
Original commit: elastic/x-pack-elasticsearch@8efaed0e9a
The constructor for CreateIndexResponse changed to include the index
name. This commit adapts x-pack-elasticsearch to this change.
Original commit: elastic/x-pack-elasticsearch@b078d80cd9
Commit elastic/x-pack-elasticsearch@b07aa78a7b was a forward port of logic needed in 5.x to get
the correct bwc branch. However, other changes on master meant that this forward port was not
needed and actually broke the bwc tests. This change removes the incorrect if statement and project name.
Original commit: elastic/x-pack-elasticsearch@9a77269fa6
This is mainly a commit to reduce noise in test logfiles when going
through them. When watcher shuts down and another node takes over, it
might try to start watcher again and tries to load triggered watches.
However the triggered watches index could be gone in the meantime due to
further shutdown. This results in logging a stack trace that the index
does not exist.
This commit checks the cluster state before trying to load triggered
watches to prevent an IndexNotFoundException in the logs.
Original commit: elastic/x-pack-elasticsearch@9f26d557d0
This introduces a check to only pause the execution of watcher
when there is no metadata but there was a shard on this node before
that inside of the ClusterStateListener.
This prevents repeated logging that watcher was paused even
though it was not necessary to call anything.
Original commit: elastic/x-pack-elasticsearch@8d3a829ffb
Watcher had an undocumented feature to configure the settings of any
index template via updating the cluster settings. Instead of changing
the template one could add a setting during runtime that overwrote
what is written in the index template.
As index template are created once and not overwritten, users should
just change the index template - or the concrete index settings like
number of replicas.
This feature was not exposed in our documentation at all.
Original commit: elastic/x-pack-elasticsearch@32e1769925
This removes the Cluster State collector and resolver and moves the collection of the cluster state (and cluster health, which is already included in cluster stats).
This makes the tests a little more stable and removes an extra network hop during monitoring data collection.
Original commit: elastic/x-pack-elasticsearch@44851d2dd6
When testing against the previous 5.x release, the bwc project incorrectly would checkout the 5.x
branch instead of the 5.5 branch as it still had the logic that applies for major versions bwc. This change adds
a check to compare the major version when making the decision on the branch to use.
Original commit: elastic/x-pack-elasticsearch@b07aa78a7b
The aim of this change is to prevent many identical requests to create
watcher index templates being submitted when a cluster first starts up
and many cluster state updates are happening. Prior to this change, if
watcher's original index template creation requests queued up behind other
cluster state change requests then for each other request watcher would
re-request creation of all its index templates. After this change it
uses a strategy similar to that used by ML to only have one creation
request per index template in the cluster state change queue at any time.
Relates elastic/x-pack-elasticsearch#1368
Relates elastic/x-pack-elasticsearch#1631
Relates elastic/x-pack-elasticsearch#1650
Original commit: elastic/x-pack-elasticsearch@ad87bf3f78
Although the job stats for jobs with missing results indices are clearly
ruined, it's better to provide zeroes for the missing values and show the
stats for other jobs than to fail the whole request. This means the UI
can continue to function.
relates elastic/x-pack-elasticsearch#1656
Original commit: elastic/x-pack-elasticsearch@a06fa994a5
* Wait for job deletion if it is in the deleting state
* Tolerate errors if multiple force delete requests
Original commit: elastic/x-pack-elasticsearch@1f0c9fbb86
When a user executes a watch and specifies it as part of the
execute watch API, no triggered watch is created, as the watch
cannot be picked up anymore (it only leaves for the duration of
the request).
However until now the TriggeredWatchStore was invoked and tried
to delete this non-existing triggered watch, resulting in some
log cluttering.
This commit removes this try to delete a non-existing triggered
watch.
Original commit: elastic/x-pack-elasticsearch@3db125cea2
This is just the culmination of all of the minor PRs associated with 1068. It will:
- Drop the `.monitoring-data-N` index
- Drop use of `_type` in all cases (replaced by `doc` and a new `type` field)
- Drop the API version from the template name (e.g., instead of `.monitoring-es-6` we now use `.monitoring-es`).
- Change API version to `-6-` from `-2-`.
- Both exporters handle versioned resources (templates, pipelines, and watches)
- HTTP exporters will optionally (true by default) publish placeholders for the old, `-2` templates.
When this is backported, it will need to:
- Change `index_patterns` to `template` within the templates.
- Downgrade the version requirements for the templates, pipeline, and watches _and_ the HTTP exporter itself (all require 6.0)
This is a companion to the feature branch in X-Pack Kibana elastic/x-pack-kibana/pull/1318 and they need to be merged at the same time.
Original commit: elastic/x-pack-elasticsearch@6031cfffa4
This commit adds better security for scroll requests in that they are now tied to a single user as
we only authorize the request that creates the scroll. This is accomplished by adding a
SearchOperationListener that listens for new scroll contexts and stores the authentication on the
ScrollContext. Then upon
retrieval of the search context for a query or fetch, the current authentication is compared to the
authentication that was present when the scroll context was created. If the current authentication
belongs to a different user, then a SearchContextMissingException will be thrown to prevent leaking
a valid vs invalid scroll id.
Additionally, signing of a scroll id is only performed when there is a older node in the cluster
that would expect the scroll id to be signed. Once this is backported to 5.x, we can remove this
bwc layer for 6.0/master.
Original commit: elastic/x-pack-elasticsearch@0e5dcafd32
When a user or client intend to delete a datafeed
and its job, there is benefit into ensuring the
datafeed has gracefully stopped (ie no data loss).
In constrast, the desired behaviour is to stop and
delete the datafeed as quickly as possible.
This change adds a force option to the delete
datafeed action. When the delete is forced,
the datafeed is isolated, its task removed and,
finally, the datafeed itself is removed from the
metadata.
relates elastic/x-pack-elasticsearch#1533
Original commit: elastic/x-pack-elasticsearch@5ae0168bf2
We try to install empty ML metadata as soon as possible after startup
if none exists. However, this still leaves a short gap when the cluster
is active with no ML metadata. To avoid problems, functions that use
the ML metadata should treat this situation as equivalent to having
empty ML metadata.
relates elastic/x-pack-elasticsearch#1643
Original commit: elastic/x-pack-elasticsearch@8f0e00cda8
ignoreAliases allows to resolve index expressions against concrete indices only, rather than against indices and aliases. It is used for now only in IndicesAliasesRequest and the indices resolution code in the security plugin needs to be adapted accordingly.
Original commit: elastic/x-pack-elasticsearch@ae964eade9
This commit switches over to two index aliases per job: one for reading
and one for writing. In the future this will allow the addition of a
rollover endpoint for ML results indices. (Rollover is still not possible
following this change, but the change to make it possible in the future
should not be a breaking change now.)
Relates elastic/x-pack-elasticsearch#1599
relates elastic/x-pack-elasticsearch#827
Original commit: elastic/x-pack-elasticsearch@d648f4631f
* Add force delete job option
* Can’t kill a process on a 5.4 node
* Address review comments
* Rename KillAutodetectAction -> KillProcessAction
* Review comments
* Cancelling task is superfluous after it has been killed
* Update docs
* Revert "Cancelling task is superfluous after it has been killed"
This reverts commit 576950e2e1ee095b38174d8b71de353c082ae953.
* Remove unnecessary TODOs and logic that doesn't alwasys force close
Original commit: elastic/x-pack-elasticsearch@f8c8b38217
This test would sometime leak threads.
The "Timer thread for LDAPConnection" is created by the unboundid SDK - closing the connection should force the thread to terminate
Original commit: elastic/x-pack-elasticsearch@bd58a17a59
The cluster state listener to decide if watcher should be reloaded was
assuming that no aliases could be used and thus wrongly could trigger
a reload, which could have lead to wrong test results.
During debugging I also added a reason for reloading and fixed another
wrong test assumption.
Also the listener does not rely on previous cluster state, but stores this
in instance variable, as we need to compare with local state and not the
previous cluster state.
Original commit: elastic/x-pack-elasticsearch@582783a66d
Includes:
- Extensive changes to "mapping roles" section
- New section for role mapping API
- Updates to LDAP/AD/PKI realms to refer to API based role mapping
- Updates to LDAP/AD realms: `unmapped_groups_as_roles` only looks at file-based mappings
- Updates to LDAP/AD realms: new setting for "metadata"
Original commit: elastic/x-pack-elasticsearch@6349f665f5
This avoids log spam about being unable to create new mappings in indices
that are set to only allow one type. (It doesn't actually have any effect
on the deletion, which was working before despite the failure to create new
mappings for the legacy types referenced by the delete request.)
relates elastic/x-pack-elasticsearch#1634
Original commit: elastic/x-pack-elasticsearch@061ce7acf1
Because:
1. It's pointless, as new detector_index values are assigned when an
analysis_config is parsed
2. It creates a backwards compatibility issue when upgrading from v5.4
Original commit: elastic/x-pack-elasticsearch@2f61aa457e
This is the x-pack side of the removal of `accumulateExceptions()` for both `TransportNodesAction` and `TransportTasksAction`.
There are occasional, random failures that occur during API calls that are silently ignored from the caller's perspective, which also leads to weird API responses that have no response and also no errors, which is obviously untrue.
Original commit: elastic/x-pack-elasticsearch@9b57321549
Detectors now have a field called detector_index. This is also now the
field that needs to be supplied when updating a detector. (Previously
it was simply index, which was confusing.)
When detectors are added to an analysis_config it will reassign
ascending detector_index values starting from 0. The intention is
never to allow deletion of detectors from an analysis_config, but
possibly to allow disabling them in the future. This ensures that
detector_index values in results will always tie up with detector_ids
in the detectors that created them.
relates elastic/x-pack-elasticsearch#1275
Original commit: elastic/x-pack-elasticsearch@20a660b07b
At the end of the test, LocalExporterTests checks if no more monitoring
data are exporter by checking multiple times the last time nodes_stats
documents were exported, stopping after 10 seconds. It does this in a
@After annotated method but it would be better to do this in a finally
block. Also, it should search for node_stats documents only if the
monitoring indices exist and are searchable to avoid some "all shards
failed" failures.
Original commit: elastic/x-pack-elasticsearch@90ffb4affd
Now we've set the option for one type per index it causes a stack trace
in to be logged if we issue a request to delete two documents with
different types. We only do this to cover the case of documents left
over from v5.4. We can avoid it by deleting by query using just the
document IDs.
Original commit: elastic/x-pack-elasticsearch@2abffc7d95
This commit removes ClientProxy and WatcherClientProxy classes. They
were added in times, where there were issues with guice and circular
dependencies. However there is no guice anymore and on top of that
the classes do not add any value.
We can switch to use a regular client, but have to make sure that
the InternalClient is injected in all the transport actions as those
is able to query data, when security is enabled.
Original commit: elastic/x-pack-elasticsearch@763a79b2f7
The goal of this change is to allow datafeeds to start
when the job is in the opening state. This makes the API
more async and it allows clients like the ML UI to open a
job and start its datafeed without having to manage the
complexity of dealing with timeouts due to the job taking
time to open due to restoring a large state.
In order to achieve this, this commit does a number of things:
- accepts a start datafeed request when the job is opening
- adds logic to the DatafeedManager to wait before running the
datafeed task until the job is opened
- refactord the datafeed node selection logic into its own class
- splitd selection issues in critical and non-critical with regard
to creating the datafeed task
- refactord the unit tests to make simpler to write & understand
- adds unit tests for added and modified functionality
- changes the response when the datafeed cannot be started to
be a conflict exception
relates elastic/x-pack-elasticsearch#1535
Original commit: elastic/x-pack-elasticsearch@c83196155d
In preparation for the removal of types, new security types like invalidated-tokens are stored in the .security
index under the generic "doc" type, with a query filter on `doc_type`.
In order to avoid id clashes, we also need to use that doc_type as part of the document id.
relates elastic/x-pack-elasticsearch#1300
Original commit: elastic/x-pack-elasticsearch@469724a228
When an error response contains multiple layers of errors, Kibana displays
the one labelled root_cause. The definition of root_cause is the most
deeply nested ElasticsearchException. Therefore, it is of great benefit to
the UI if our config validation returns the actual problem in an
ElasticsearchException rather than an IllegalArgumentException.
This commit also adds an extra validation check to catch the case of a
single job config containing fields x.y as well as x earlier. Previously
this was caught when we tried to create results mappings, and was
accompanied by an error suggesting that using a dedicated results index
would help, when clearly it won't for a clash in a single job config.
Fixeselastic/x-pack-kibana#1387Fixeselastic/prelert-legacy#349
Original commit: elastic/x-pack-elasticsearch@7d1b7def6c
Otherwise it's possible that the get_filter endpoint can return a filter that's been
deleted. Although this is the behaviour of the search API, specific metadata
management APIs should provide better guarantees.
Original commit: elastic/x-pack-elasticsearch@818495f176
This allows us to build both 5.5.0-SNAPSHOT and 5.4.1-SNAPSHOT
artifacts for backwards compatibility testing. It is a port of
elastic/elasticsearch:24870 to x-pack and will be super useful
when elastic/elasticsearch:24846 is ported to x-pack.
Original commit: elastic/x-pack-elasticsearch@0ea443f488
Previously there were two @After methods in the XPackRestIT class, and
there is no guarantee about the order in which these run. This commit
replaces these with a single @After method that calls the cleanup methods
in a well-defined order.
Original commit: elastic/x-pack-elasticsearch@d3ab366591
The has_privileges API now supports wildcards.
The semantics are that the user must have a superset of the wildcard being checked.
---------------------
Role | Check | Result
---------------------
* | foo* | true
f* | foo* | true
foo* | foo* | true
foo* | foo? | true
foo? | foo? | true
foo? | foo* | false
foo | foo* | false
Original commit: elastic/x-pack-elasticsearch@817550db17
We don't hyphenate metadata anywhere else.
Also added tests for the LdapMetaDataResolver as they were completely absent.
Original commit: elastic/x-pack-elasticsearch@eec647ba93
This deletes tests getting results from MlJobIT since
such tests already exist in a form that is simpler to
understand and maintain in the YAML suite.
Original commit: elastic/x-pack-elasticsearch@b708e24877
This commit means that newly created ML state indices will have a single
type named "doc", and newly persisted state documents will have type
"doc" too.
Retrieving state is only supported for type "doc".
When deleting state, documents with the old types are deleted in addition
to those with type "doc". This means jobs created by the beta can be fully
deleted.
Relates elastic/x-pack-elasticsearch#668
Original commit: elastic/x-pack-elasticsearch@29c07d40f1
This commit adds an internal project call ml-cpp-snapshot which when
built will pull the ml cpp zip file from the prelert bucket. The GET
request has retries added to handle the dynamic aws creds eventual
consistency.
Original commit: elastic/x-pack-elasticsearch@1bba7d0f08
This commit cleans up the check for SSL with client authentication when a PKI realm is enabled by
moving it from the realm to a actual bootstrap check.
A bug was found during this cleanup in the check for transport profiles and that is also fixed in
this commit.
relates elastic/x-pack-elasticsearch#420
Original commit: elastic/x-pack-elasticsearch@3aa6a3edc0
The commit that converted the results index into single type
broke the search for fetching results for renormalization.
This commit fixes that.
Original commit: elastic/x-pack-elasticsearch@1ca7517adc
REST endpoints that support GET and POST need
to also support source parsing. As these
endpoints can accept a body but some clients
do not allow doing a GET with a request body,
elasticsearch has support for parsing via a
source URI parameter. This commit adds source
handling to all such endpoints.
relates elastic/x-pack-elasticsearch#1204
Original commit: elastic/x-pack-elasticsearch@3949ea31fe
Previously we used to normalize records with their buckets. This required
nested scrolls: an outer scroll over buckets, then a nested scroll for
records in each bucket. This was fragile.
The new approach is to simply scroll first through buckets, then through
records. This is made possible because we no longer store max_record_score
on buckets nor bucket anomaly_score on records.
While making these changes I noticed that the PerPartitionMaxProbabilities
class was redundant (because it was storing max_record_score in the case of
per-partition normalization), so I removed it. I also removed a redundant
Map from the Bucket class and fixed its equals() and hashCode() methods.
relates elastic/x-pack-elasticsearch#1115
Original commit: elastic/x-pack-elasticsearch@efbee63573
Deleting a job issues 2 cluster state updates.
The first marks the job as deleted.
The second actually removes the job.
Both check that there is no datafeed referring to the job.
If a put datafeed request arrives between those 2 cluster
state updates, the datafeed gets created and the final
job cluster state update fails. This means we end up with
both the job and the datafeed, but the job's results and
state have been deleted.
This commit changes the behaviour so that the put
datafeed request fails for a job that is marked as deleted
as this scenario is avoiding partially executing an action.
relates elastic/x-pack-elasticsearch#1510
Original commit: elastic/x-pack-elasticsearch@76fa0f0b1a
This commit changes all results to use the single doc type.
All searches were adjusted to work without the need to specify
type which ensures BWC with 5.4 results.
Additional work is needed to put the new doc mapping in indices
created with 5.4 but it will be done in separate PR.
Relates elastic/x-pack-elasticsearch#668
Original commit: elastic/x-pack-elasticsearch@041c88ac2d
* Reduced a longish timeout to a shorter one, as a watch should be
executed in a HTTP test.
* Ensured that the TimeThrottleIntegration tests only query for own
watches in the watch history, also use random names for watch ids
* HipChatServiceTests configured deprecated logging package, so it was
not possible to follow the HTTP calls to the hipchat service endpoint.
relates elastic/x-pack-elasticsearch#1514
Relates elastic/x-pack-elasticsearch#1515
Original commit: elastic/x-pack-elasticsearch@adb492e4e9
* Remove sequenceNum from anomaly records and influencers
* Generate unqiue IDs without sequence numbers
* Remove more instances of sequence_num
* Handle parsing sequnce_num from v5.4
Original commit: elastic/x-pack-elasticsearch@e60b206daf
This commit simplifies the WatchStatsTests to use a MockScriptPlugin.
The latch script engine previously depended on a static instance of the
engine to contain the latches. These are now moved to statics of the
test class itself.
Original commit: elastic/x-pack-elasticsearch@4170cd1bd3
This commit enables security to work with an index named .security (as
it could before) OR an alias named .security that points to a concrete
index by a different name that has the security index. This prepares
the ability to migrate from a 5.x to 6.x security index that allows
changing and re-indexing the underlying security index while maintaining
a .security alias that points to the underlying updated index.
relates elastic/x-pack-elasticsearch#1216
Original commit: elastic/x-pack-elasticsearch@9fee12e5a0
Suppress many job/datafeed errors if a node is known to be shutting down. Also, ensure started datafeeds and open jobs don't end up stopped/failed due to errors as the shutdown progresses, as this would prevent them automatically relocating to a different node.
relates elastic/x-pack-elasticsearch#1114
Original commit: elastic/x-pack-elasticsearch@e56a7dbea1
Set job to failed if autodetect manager fails closing, fix force closing of jobs that hang in closing
state, set timeout when waiting for clusterstate update, disallow closing of failed jobs with normal
close
relates elastic/x-pack-elasticsearch#1453
Original commit: elastic/x-pack-elasticsearch@493cf85e22
Aliases might be contained in requests that check index permissions
to disable caches etc. This commit preserves permissions for
aliases as well.
Original commit: elastic/x-pack-elasticsearch@233195aeba
The authentication object was changed in 5.4.0 in that it was conditionally signed depending on
the version and other factors. A bug was introduced however that causes the authentication to
actually get written with the version of the node it is being sent to even if that version is
greater than the version of the current node, which causes rolling upgrades to fail.
Original commit: elastic/x-pack-elasticsearch@a718ff8a52
* Add sort parameter for get buckets
* Add secondary sort by time
* Use default values from actions in rest requests
Original commit: elastic/x-pack-elasticsearch@a530c0bed6
This commit ensures the old 5.x index templates are removed
using the existing plugin hook, instead of the self written part.
Original commit: elastic/x-pack-elasticsearch@6faf08d98d
This commit adds compatibility checks while opening a job:
- Checks that jobs without versions (< 5.5) are not opened
- Checks that jobs with incompatible types are not opened
Original commit: elastic/x-pack-elasticsearch@a3adab733e
* Hide ML actions for tribe node client
* Remove unused parameters
* Enable ML actions and rest endpoints for the transport client
* Create the ML components for the transport client
* Add ml transport client tests
Original commit: elastic/x-pack-elasticsearch@509007ca29
* Remove repeated calls to validateAndReturnJobTask
* Wait for closing job
* Refactor resolving job ids
* More close job unit tests
* Don’t finalise closing jobs twice
Original commit: elastic/x-pack-elasticsearch@20616d6f0a
The LocalExporterTests.testLocalExporterFlush() test was removed in elastic/x-pack-elasticsearch#835 when
LocalExporterTests was changed. This test verified that export exceptions are
thrown when monitoring documents are exported using the Monitoring Bulk API but
the underlying monitoring indices are closed.
This commit reintroduces this test, but as a REST test this time.
relates elastic/x-pack-elasticsearch#416
Original commit: elastic/x-pack-elasticsearch@0a42f9a1be
This is the xpack side of elastic/elasticsearch#24447. The one caveat is
there are a number of places within the xpack api that use
BytesReference to pass down to templates. These are convert to encode
into a BytesArray when necessary, so as not to require changing all of
those apis here at once, but they should all be convert to String as
well.
Original commit: elastic/x-pack-elasticsearch@8399b9d8c3
For user tokens, we were storing the expiration time as a date with the date time format in the
mappings. Occasionally, we would get CI failures due to parsing the date and having an invalid
format. This change simplifies the user tokens to simply use an Instant and convert it to the epoch
millis when we index the document. This eliminates the need for a date formatter and should ensure
we no longer have issues with parsing the dates.
Additionally, the TokenAuthIntegTests#testExpireMultipleTimes could fail if the token was expired
but the approximation for the expiration time was after the current time. In order to resolve this
we get the exact expiration time from the token and use that in the test.
relates elastic/x-pack-elasticsearch#1255
Original commit: elastic/x-pack-elasticsearch@d4bf53e7bc
When we update a model snapshot we need to write it back to
the exact index where we read it from. This is necessary
for rollover as otherwise we would end up with two different
versions of the same snapshot.
Relates elastic/x-pack-elasticsearch#827
Original commit: elastic/x-pack-elasticsearch@b5d1ab38a7
Support the resolution of remote index names, including those that contain wildcards in the cluster name or index part)
Specifically these work:
- `GET /remote*:foo/_search`
- `GET /*:foo/_search`
- `GET /*:foo,*/_search`
- `GET /remote:*/_search`
- `GET /*:*/_search`
This change assumes that every user is allowed to attempt a cross-cluster search against any remote index, and the actual authorisation of indices happens on the remote nodes. Thus ` GET /*:foo/_search` will expand to search the `foo` index on every registered remote without consideration of the roles and privileges that the user has on the source cluster.
Original commit: elastic/x-pack-elasticsearch@b45041aaa3
As this does not require any reindexing this is easy to fix by just
changing the watch history template.
In addition the old templates are deleted on start up and the new ones
are instantiated.
Original commit: elastic/x-pack-elasticsearch@7e1ad495ad
I broke this by adding more randomization to core's randomTimeValue
method. Now it produces valid time values that don't round trip
properly over the XContent API. This change causes the test to skip
trying to round trip time these sorts of time values.
Original commit: elastic/x-pack-elasticsearch@dcdd588bdb
This is in preparation of introducing a write alias.
It adjusts all requests to persist results to do so
using a method that returns the write alias (even though
it currently returns the same as the read alias).
Relates elastic/x-pack-elasticsearch#827
Original commit: elastic/x-pack-elasticsearch@1358dd8dcf
This commit updates the RunAsIntegTests to randomly assign the superuser role to the user that
is authenticating with the cluster but the request is being run as a different user. This provides
additional validation that the authorization errors are actually coming from the user the request
is running as and not due to the authenticating user's privileges.
Original commit: elastic/x-pack-elasticsearch@c6360d13e6
Each job introduces new fields to the results index matching
the analysis terms. When the job is created, mappings for those
are added explicitly. However, when rollover is introduced, that
will not be the case. This commit prepares for that by adding
dynamic mapping of new fields as keyword.
Relates elastic/x-pack-elasticsearch#827
Original commit: elastic/x-pack-elasticsearch@8f6cd09a71
This was the behaviour in Shield 2.x, but it was accidentally changed during migration to X-Pack 5.x
Original commit: elastic/x-pack-elasticsearch@de0bf5e688
It just wastes 20 seconds while we timeout trying to open named pipes that cannot
possibly have been created.
Original commit: elastic/x-pack-elasticsearch@4e447874f6
fixes build errors. Still ensures that the timestamp is set to 'now' if the parsed logfile misses it.
Original commit: elastic/x-pack-elasticsearch@cf60e8d76b
If invalid job configs are transported to the master node then the root
cause of the validation exception gets reported as a remote_transport_exception,
which is extremely confusing.
This commit moves the validation of job configurations to the first node that
handles the action.
Fixeselastic/x-pack-kibana#1172
Original commit: elastic/x-pack-elasticsearch@5ed59d2a6f
Before searching for documents in monitoring indices, we need to ensure
that they exist and are available.
Original commit: elastic/x-pack-elasticsearch@29db55a1fe
This commit adds a base rest handler for security that handles the license checking in the security
apis. This was done previously in some rest handlers but not all and actually had issues where a
value would be returned but we may not have consumed all of the request parameters, which could
lead to a different response being returned than what we would have expected.
relates elastic/x-pack-elasticsearch#1236
Original commit: elastic/x-pack-elasticsearch@2f1100b64a
Adds defenseagainst broken scrolls to the fetching roles and users.
While Elasticsearch *shouldn't* send broken scrolls it has done so
in the past and when it does this causes security to consume the
entire heap and crash. This changes it so we instead fail the request
with a message about the scroll being broken.
Relates to elastic/x-pack-elasticsearch#1299
Original commit: elastic/x-pack-elasticsearch@dfef87e757
Some JDKs do not support the ECDSA cipher suites that we use in the EllipticCurveSSLTests, which
is the underlying cause of some CI failures. This change ensures there is at least one enabled
ECDSA cipher before testing that a connection can be made.
relates elastic/x-pack-elasticsearch#1278
Original commit: elastic/x-pack-elasticsearch@f6c93d776c
* [ML] Not an error to close a job twice
* Error if job is opening
* Address review comments
* Test closed job isn’t resolved
Original commit: elastic/x-pack-elasticsearch@7da7b24c08
Originally we used to try to get the native code version even when ML was disabled.
However, this has proved to be an annoyance in cases where people running on
unsupported platforms have disabled ML. This commit means we'll only try to get
the native code version if ML is enabled on a node.
Original commit: elastic/x-pack-elasticsearch@778d6708d2
This removes the "node" type from `.monitoring-data-2`. This data is sent to _both_ the time-based and non-time-based indexes for Elasticsearch, but the UI only used the time-based variant already.
This is another step in the process of removing the `.monitoring-data-2` index. There is now only one `_type` left in that index: `cluster_info`, which is used by the UI and phone home stats because it contains the license details _and_ the `stack_stats` (e.g., `xpack_usage`).
Original commit: elastic/x-pack-elasticsearch@2cadb5939d
This is a task towards allowing rollover.
Multiple model_size_stats are stored in order to allow
analytics of the memory usage over time. The job _stats
need to display the latest model_size_stats. Before this
commit, the latest model_size_stats was being stored with
a special ID and it was retrieved using that ID. This
does not lend itself well for rollover as we would end up
with multiple of those special IDs in the rolled indices.
This commit removes the need to store a special model_size_stats
version. Job _stats now retrieve the model_size_stats by searching
for the latest one. It also uses a manual ID for all model_size_stats
in order to maintain a single document per log_time.
Relates elastic/x-pack-elasticsearch#827
Original commit: elastic/x-pack-elasticsearch@b2796e9b08
Now that the Monitoring UI no longer checks the `.monitoring-data-2` index
for Kibana, Logstash, or Beats data, we can stop accepting the duplicated
data in that index (the _exact_ same documents are also indexed into the
time-based index for each product).
This ignores rather than rejects requests that contain such documents to
allow older clients to communicate with a 5.5+ monitoring cluster.
Original commit: elastic/x-pack-elasticsearch@def472cf2e
- role_mapping.native is always present, but contains no entries if the security index is unavailable
- file based role mapping does not allow duplicate keys
Original commit: elastic/x-pack-elasticsearch@734cf7e2c0
As fields with underscores will be disallowed in master, and we have to
prepare the upgrade, this commit renames the _status field to status.
When the 5.x upgrade logic is in place in the 5.x we can remove all the
old style _status handling from the master branch.
Note: All the BWC compatibility tests, that load 5.x indices are now
faking a finished upgrade by adding the `status` field to the mapping
of the watches index.
Original commit: elastic/x-pack-elasticsearch@9d5cc9aaec
In some cases (based on the randomisation) the primary-realm would have no mappings, but the secondary-realm would.
If this occurred we wouldn't write out any mappings and the secondary-realm would not authorize the access being tested
Original commit: elastic/x-pack-elasticsearch@8e81ec9dd7
This introduces a role-mapping API to X-Pack security.
Features:
- A `GET`/`PUT`/`DELETE` API at `/_xpack/security/role_mapping/`
- Role-mappings are stored in the `.security` index
- A custom expression language (in JSON) for expressing the mapping rules
- Supported in LDAP/AD and PKI realms
- LDAP realm also supports loading arbitrary meta-data (which can be used in the mapping rules)
- A CompositeRoleMapper unifies roles from the existing file based mapper, and the new API based mapper.
- Usage stats for native role mappings
Original commit: elastic/x-pack-elasticsearch@d9972ed1da
This is related to elastic/elasticsearch#24412. That commit changed how
ListenableActionFuture implementations are created. This commit
updates x-pack to be compatible with those changes. In particular, all
the usages of ListenableActionFuture in x-pack could be replaced with
PlainActionFuture as the "listening" functionality was not being used.
Original commit: elastic/x-pack-elasticsearch@7c8d8e3df9
JobDataDeleter handles the deletion logic for 3 cases:
1. deleting a model snapshot and its state docs
2. deleting all results after a timestamp
3. deleting all interim results
The last 2 are currently implemented by manually performing
a search and scroll and then adding matching hits in a bulk
delete action. This operation is exactly what delete-by-query
does.
This commit changes JobDataDeleter to use delete-by-query. This
makes the code simpler and less error-prone. The downside is
losing some logging which seems non-critical. Unit tests for
JobDataDeleter are also removed as they are heavily mocked tests,
adding little value and high maintenance cost. This functionality
is tested by integration tests already.
relates elastic/x-pack-elasticsearch#821
Original commit: elastic/x-pack-elasticsearch@7da91332bd
The distribution of watches now happens on the node which holds the
watches index, instead of on the master node. This requires several
changes to the current implementation.
1. Running on shards and replicas
In order to run watches on the nodes with the watches index on its
primaries and replicas. To ensure that watches do not run twice, there is
a logic which checks the local shards, runs a murmurhash on the id and
runs modulo against the number of shards and replicas, this is the way to
find out, if a watch should run local. Reloading happens
2. Several master node actions moved to a HandledTransportAction, as they
are basically just aliases for indexing actions, among them the
put/delete/get watch actions, the acknowledgement action, the de/activate
actions
3. Stats action moved to a broadcast node action, because we potentially
have to query every node to get watcher statistics
4. Starting/Stopping watcher now is a master node action, which updates
the cluster state and then listeners acts on those. Because of this watches
can be running on two systems, if you those have different cluster state
versions, until the new watcher state is propagated
5. Watcher is started on all nodes now. With the exception of the ticker
schedule engine most classes do not need a lot of resources while running.
However they have to run, because of the execute watch API, which can hit
any node - it does not make sense to find the right shard for this watch
and only then execute (as this also has to work with a watch, that has not
been stored before)
6. By using a indexing operation listener, each storing of a watch now
parses the watch first and only stores on successful parsing
7. Execute watch API now uses the watcher threadpool for execution
8. Getting the number of watches for the stats now simply queries the
different execution engines, how many watches are scheduled, so this is
not doing a search anymore
There will be follow up commits on this one, mainly to ensure BWC compatibility.
Original commit: elastic/x-pack-elasticsearch@0adb46e658
Within the same JVM, setting the number of processors available to Netty
can only be done once. However, tests randomize the number of processors
and so without intervention would attempt to set this value multiple
times. Therefore, we need to use a flag that prevents setting this value
in tests.
Relates elastic/x-pack-elasticsearch#1266
Original commit: elastic/x-pack-elasticsearch@d127149725
Many of the tests assume that the trial license has already been generated before the test gets to run. As this is asynchronously triggered upon node
startup, however, there is no guarantee that trial license generation has completed before the tests get to execute, leading to null values when
checking clusterService.state().metaData().custom(LicensesMetaData.TYPE).
Original commit: elastic/x-pack-elasticsearch@d909c9ba95
LicenseManagerServiceTests sometimes fails in jenkins, but fairly
rarely. We don't get useful logs when it does. This cranks up
the log level and adds some more assertions so we can better track
down where the failure comes from.
Relates to elastic/x-pack-elasticsearch#222
Original commit: elastic/x-pack-elasticsearch@3e08725fc7
When a user creates a datafeed, as well as checking they have permission
to create a datafeed we also check that they have permission to search the
indices they've configured the datafeed to search.
Previously this second check was erroneously done for the user who issued
the put_datafeed request, whereas it should be done as the runas user for
that request.
Original commit: elastic/x-pack-elasticsearch@4c35204c66
When we revert to snapshot, if we delete intervening results
we should delete with an open end on the time range for the
case when future data has been posted to the job.
Original commit: elastic/x-pack-elasticsearch@c3f5e8f19e
The one argument ctor for `Script` creates a script with the
default language but most usages of are for testing and either
don't care about the language or are for use with
`MockScriptEngine`. This replaces most usages of the one argument
ctor on `Script` with calls to `ESTestCase#mockScript` to make
it clear that the tests don't need the default scripting language.
Original commit: elastic/x-pack-elasticsearch@c1d05b7357
When the active directory realm was refactored to add support for authenticating against multiple
domains, only the default authenticator respected the user_search.filter setting. This commit moves
this down to the base authenticator and also changes the UPN filter to not include sAMAccountName
in the filter.
Original commit: elastic/x-pack-elasticsearch@d2c19c9bee
- Removes need to handle exception from action methods
- Clearly renames DatafeedJobIT to DatafeedJobsRestIT to distinguish
from DatafeedJobsIT
- Refactors DatafeedJobsIT to reuse MlNativeAutodetectIntegTestCase
Original commit: elastic/x-pack-elasticsearch@5bd0c01391
Cross cluster search uses ClusterSearchShardsAction under the covers.
Without this change, you would need both "read_cross_cluster" and "view_index_metadata" privilegs in order to have permission to execute searches from a remote cluster.
Original commit: elastic/x-pack-elasticsearch@65a6aff329
If a single permission set does not have a query defined then this should be considered as the user
not having document level security for the indices matching that pattern. However, the lack of
document level security was not being taken into account and document level security was being
applied when it should not have been.
Original commit: elastic/x-pack-elasticsearch@f5777c2f37
Many tests in monitoring use the pattern of calling first awaitMonitoringDocsCount, and then doing a search that checks certain properties, assuming
that the doc count is correct at that point. In the presence of replicas, awaitMonitoringDocsCount might not wait for all shard copies to have the
desired property. A subsequent search might then hit a shard where the property does not hold. As these tests randomize the number of replicas
(through the random_index_template), it easier to constrain awaitMonitoringDocsCount just to the primary and then do subsequent checks just by
querying the primary.
Original commit: elastic/x-pack-elasticsearch@4165beb903
This commit creates a single server socket that will be connected to by local sockets. The local
sockets will use the port of the previously stopped ldap server as the local port. This will
prevent the ldap library from establishing a connection. The previous use of server sockets for
this did not work on all operating systems as the backlog parameter has platform specific meaning.
Original commit: elastic/x-pack-elasticsearch@03b6bf39d4
This commit adds a timeout for the expiration of invalidated tokens so that we can expect that the
requests will have been finished before we do the assertions on the internal test cluster.
Original commit: elastic/x-pack-elasticsearch@2928706224
This commit adds a token based access mechanism that is a subset of the OAuth 2.0 protocol. The
token mechanism takes the same values as a OAuth 2 standard (defined in RFC 6749 and RFC 6750),
but differs in that we use XContent for the body instead of form encoded values. Additionally, this
PR provides a mechanism for expiration of a token; this can be used to implement logout
functionality that prevents the token from being used again.
The actual tokens are encrypted using AES-GCM, which also provides authentication. The key for
encryption is derived from a salt value and a passphrase that is stored on each node in the
secure settings store. By default, the tokens have an expiration time of 20 minutes and is
configurable up to a maximum of one hour.
Relates elastic/x-pack-elasticsearch#8
Original commit: elastic/x-pack-elasticsearch@3d201ac2bf
This is an issue where a bucket can have both interim results and
non-interim results, a bucket should never have both at the same time.
The steps to cause this situation are:
1. Flush a running job and create interim results
2. Close that job (this does not delete interim results)
3. Re-open the job and POST data
4. The job will eventually emit a bucket result which mingles with the
existing interim results
Normally interim results are deleted by AutoDetectResultProcessor when a
bucket is parsed following a flush command. Because of the close and
re-opening of the job AutoDetectResultProcessor no longer knows that a
previous flush command creating interim results.
The fix is to always delete interim results the first time
AutoDetectResultProcessor sees a bucket.
relates elastic/x-pack-elasticsearch#1188
Original commit: elastic/x-pack-elasticsearch@5326455f54
In the SessionFactoryLoadBalancingTests, we sometime want a connection to a certain IP and Port to
fail as a way to mock an unresponsive/disconnected LDAP server. The test does this by starting up
multiple LDAP servers and then shutting some down. When the server is shut down the port that it
was bound to is open for another process or test to bind to, which can lead to sporadic failures in
CI. This change is a best effort attempt to prevent this by binding a server socket to the port and
filling its backlog so other connections should fail.
Relates elastic/x-pack-elasticsearch#1195
Original commit: elastic/x-pack-elasticsearch@b31a560c93
The DatafeedJobsIT.testRealtime test fails from time to time.
The test seems to take a long time to execute the flush action
after the lookback. This could make sense as the test produces
a few records over the span of a week with 5 minutes bucket_span.
Thus, flush will end up doing a lot of word to create results
for so many buckets.
This change increases the bucket_span to 1 hour. Hopefully, this
will stop the failures.
Relates elastic/x-pack-elasticsearch#1162
Original commit: elastic/x-pack-elasticsearch@4366907371
This commit fixes the support for elliptic curve certificates that are specified as a PEM file.
These certificates and private keys can now be read properly and a integration test was added to
ensure that TLS also functions correctly with these certificates.
Original commit: elastic/x-pack-elasticsearch@6d6d579c88
This commit reduces spamming of the logs when a common SSL exception is encountered such as a
client not trusting the server's certificate or a plaintext request sent to a channel that expects
TLS traffic.
relates elastic/x-pack-elasticsearch#1062
Original commit: elastic/x-pack-elasticsearch@94959e79f6
This change prevents the situation where cleanup of ML indices immediately
after deleting a job leaves the audit notification in limbo because the index
it was due to be indexed into has been deleted.
Relates elastic/x-pack-elasticsearch#1142
Original commit: elastic/x-pack-elasticsearch@300e9c36ce
Ordinary Kibana users should not have access to the cluster state of ES,
and therefore they should not be able to access ML jobs without explicit
permission.
Original commit: elastic/x-pack-elasticsearch@77273d561a
When a condition is unmet, the ack status of the actions needs to be
resetted again, so that new alerts can be triggered.
Due to a bugfix this functionality was removed from ES 5.0.0-alpha5
onwards.
relates elastic/x-pack-elasticsearch#1123
Original commit: elastic/x-pack-elasticsearch@83db2cecf9
Persistent tasks should verify that completion notification is done for correct version of the task, otherwise a delayed notification from an old node can accidentally close a newly reassigned task.
Original commit: elastic/x-pack-elasticsearch@478bb6e730
* Adds a check to wait for active tasks for XPackRestIT
* uses test logger
* Change to use assertBusy instead of awaitBusy
* fixes failures with active tasks remaining
* Moves wait for pending tasks into MlRestTestStateCleaner
* remove unecessary log line
Original commit: elastic/x-pack-elasticsearch@1f098dbb64
By creating the watches via the exporter, we get to afford ourselves
with a much more automatic and simpler set of security permissions.
This does limit us in a few ways (e.g., every exporter has to deal with
cluster alerts itself, which means that newer releases of Kibana cannot
help by adding newer cluster alerts for older, still-monitored
clusters).
Original commit: elastic/x-pack-elasticsearch@448ef313c3
When Logstash 5.2 - 5.3 submit documents via the `_xpack/monitoring/_bulk`
endpoint, it sends its time-based documents with an explicit `_id` of
`""`.
This used to be automatically ignored by Monitoring, but we now accept the
_id that we are given (including `null`). ES, prior to 5.3.1, accepted
`""` as a valid `_id` through the `_bulk` endpoint, which means that it
blindly accepted and overwrote documents given that ID, meaning that all
Logstash instances "shared" the exact same document and therefore the UI
becomes useless.
This change allows `""` to be used and it simply replaces that value, and
only that value, with `null`. This enables backwards compatibility with LS
5.2 - 5.3.0.
Original commit: elastic/x-pack-elasticsearch@889578e61e
PersistentTasksCustomMetadata was using a generic param named `Params`. This conflicted with the imported interface `ToXContent.Params`. The java compiler was preferring the generic param over the interface so everything was fine but Eclipse apparently prefers the interface int his case which was screwing up the Hierarchy and causing compile errors in Eclipse. This changes fixes it by renaming the Generic param to `P`
Original commit: elastic/x-pack-elasticsearch@8528870684
- Mark all security indices (that is all indices managed by SecurityLifecycleService) as "superuser only" (only superuser role can have direct permissions)
- Add unit tests for IndexLifecycleManager
Original commit: elastic/x-pack-elasticsearch@e4478825e0
This commit removes the SecuredString class that was previously used throughout the security code
and replaces it with the SecureString class from core that was added as part of the new secure
settings infrastructure.
relates elastic/x-pack-elasticsearch#421
Original commit: elastic/x-pack-elasticsearch@e9cd117ca1
Luca Cavanna
Nik Everett
Jay Modi
Pier-Hugues Pellerin
David Kyle
Tim Vernum
Martijn van Groningen
Tim Brooks
Jim Ferenczi
David Roberts
Boaz Leskes
Igor Motov
Simon Willnauer
Dimitrios Athanasiou
Alexander Reelsen
Chris Earle
Ryan Ernst
Suyog Rao
Jason Tedor
jaymode
Chris Earle
Jack Conradson
Colin Goodheart-Smithe
Tanguy Leroux
markharwood
Andrew Cholakian
Ali Beyad
Hendrik Muhs
Tim Vernum
javanna
Adrien Grand
Daniel Mitterdorfer
Yannick Welsch
Zachary Tong