This commit removes Kerberos bootstrap checks as they were more
validation checks and better done in Kerberos realm constructor
than as bootstrap checks. This also moves the check
for one Kerberos realm per node to where we initialize realms.
This commit adds few validations which were missing earlier
like missing read permissions on keytab file or if it is directory
to throw exception with error message.
The default behaviour for "GetPrivileges" is to get all application
privileges. This should only be allowed if the user has access to
the "*" application.
In #29623 we added `Request` object flavored requests to the low level
REST client and in #30315 we deprecated the old `performRequest`s. This
changes all calls in the `x-pack/plugin/security` project to use the new
versions.
The main highlight is the removal of the reclaim_deletes_weight in the TieredMergePolicy.
The es setting index.merge.policy.reclaim_deletes_weight is deprecated in this commit and the value is ignored. The new merge policy setting setDeletesPctAllowed should be added in a follow up.
In #29623 we added `Request` object flavored requests to the low level
REST client and in #30315 we deprecated the old `performRequest`s. This
changes all calls in the `x-pack/plugin` project to use the new versions.
Previously we had two patterns for naming of strict
and lenient parsers.
Some classes had CONFIG_PARSER and METADATA_PARSER,
and used an enum to pass the parser type to nested
parsers.
Other classes had STRICT_PARSER and LENIENT_PARSER
and used ternary operators to pass the parser type
to nested parsers.
This change makes all ML classes use the second of
the patterns described above.
Removing some dead code or supressing warnings where apropriate. Most of the
time the variable tested for null is dereferenced earlier or never used before.
Today we allow plugins to add index store implementations yet we are not
doing this in our new way of managing plugins as pull versus push. That
is, today we still allow plugins to push index store providers via an on
index module call where they can turn around and add an index
store. Aside from being inconsistent with how we manage plugins today
where we would look to pull such implementations from plugins at node
creation time, it also means that we do not know at a top-level (for
example, in the indices service) which index stores are available. This
commit addresses this by adding a dedicated plugin type for index store
plugins, removing the index module hook for adding index stores, and by
aggregating these into the top-level of the indices service.
Java 11 uses more verbose exceptions messages, causing this assertion
to fail. Changed the test to be less restrictive and only look
for the classes we care about.
This commit reverts to the pre-6.3 way of merging automata as the
change in 6.3 significantly impacts the performance for roles with a
large number of concrete indices. In addition, the maximum number of
states for security automata has been increased to 100,000 in order
to allow users to use roles that caused problems pre-6.3 and 6.3 fixed.
As an escape hatch, the maximum number of states is configurable with
a setting so that users with complex patterns in roles can increase
the states with the knowledge that there is more memory usage.
* Introduce fips_mode setting and associated checks
Introduce xpack.security.fips_mode.enabled setting ( default false)
When it is set to true, a number of Bootstrap checks are performed:
- Check that Secure Settings are of the latest version (3)
- Check that no JKS keystores are configured
- Check that compliant algorithms ( PBKDF2 family ) are used for
password hashing
In the HL REST client we replace the License object with a string, because of
complexity of this class. It is also not really needed on the client side since
end-users are not interacting with the license besides passing it as a string
to the server.
Relates #29827
This commit introduces "Application Privileges" to the X-Pack security
model.
Application Privileges are managed within Elasticsearch, and can be
tested with the _has_privileges API, but do not grant access to any
actions or resources within Elasticsearch. Their purpose is to allow
applications outside of Elasticsearch to represent and store their own
privileges model within Elasticsearch roles.
Access to manage application privileges is handled in a new way that
grants permission to specific application names only. This lays the
foundation for more OLS on cluster privileges, which is implemented by
allowing a cluster permission to inspect not just the action being
executed, but also the request to which the action is applied.
To support this, a "conditional cluster privilege" is introduced, which
is like the existing cluster privilege, except that it has a Predicate
over the request as well as over the action name.
Specifically, this adds
- GET/PUT/DELETE actions for defining application level privileges
- application privileges in role definitions
- application privileges in the has_privileges API
- changes to the cluster permission class to support checking of request
objects
- a new "global" element on role definition to provide cluster object
level security (only for manage application privileges)
- changes to `kibana_user`, `kibana_dashboard_only_user` and
`kibana_system` roles to use and manage application privileges
Closes#29820Closes#31559
This bundles the x-pack:protocol project into the x-pack:plugin:core
project because we'd like folks to consider it an implementation detail
of our build rather than a separate artifact to be managed and depended
on. It is now bundled into both x-pack:plugin:core and
client:rest-high-level. To make this work I had to fix a few things.
Firstly, I had to make PluginBuildPlugin work with the shadow plugin.
In that case we have to bundle only the `shadow` dependencies and the
shadow jar.
Secondly, every reference to x-pack:plugin:core has to use the `shadow`
configuration. Without that the reference is missing all of the
un-shadowed dependencies. I tried to make it so that applying the shadow
plugin automatically redefines the `default` configuration to mirror the
`shadow` configuration which would allow us to use bare project references
to the x-pack:plugin:core project but I couldn't make it work. It'd *look*
like it works but then fail for transitive dependencies anyway. I think
it is still a good thing to do but I don't have the willpower to do it
now.
Finally, I had to fix an issue where Eclipse and IntelliJ didn't properly
reference shadowed transitive dependencies. Neither IDE supports shadowing
natively so they have to reference the shadowed projects. We fix this by
detecting `shadow` dependencies when in "Intellij mode" or "Eclipse mode"
and adding `runtime` dependencies to the same target. This convinces
IntelliJ and Eclipse to play nice.
This commit adds support for Kerberos authentication with a platinum
license. Kerberos authentication support relies on SPNEGO, which is
triggered by challenging clients with a 401 response with the
`WWW-Authenticate: Negotiate` header. A SPNEGO client will then provide
a Kerberos ticket in the `Authorization` header. The tickets are
validated using Java's built-in GSS support. The JVM uses a vm wide
configuration for Kerberos, so there can be only one Kerberos realm.
This is enforced by a bootstrap check that also enforces the existence
of the keytab file.
In many cases a fallback authentication mechanism is needed when SPNEGO
authentication is not available. In order to support this, the
DefaultAuthenticationFailureHandler now takes a list of failure response
headers. For example, one realm can provide a
`WWW-Authenticate: Negotiate` header as its default and another could
provide `WWW-Authenticate: Basic` to indicate to the client that basic
authentication can be used in place of SPNEGO.
In order to test Kerberos, unit tests are run against an in-memory KDC
that is backed by an in-memory ldap server. A QA project has also been
added to test against an actual KDC, which is provided by the krb5kdc
fixture.
Closes#30243
* Complete changes for running IT in a fips JVM
- Mute :x-pack:qa:sql:security:ssl:integTest as it
cannot run in FIPS 140 JVM until the SQL CLI supports key/cert.
- Set default JVM keystore/truststore password in top level build
script for all integTest tasks in a FIPS 140 JVM
- Changed top level x-pack build script to use keys and certificates
for trust/key material when spinning up clusters for IT
Adds a new single-value metrics aggregation that computes the weighted
average of numeric values that are extracted from the aggregated
documents. These values can be extracted from specific numeric
fields in the documents.
When calculating a regular average, each datapoint has an equal "weight"; it
contributes equally to the final value. In contrast, weighted averages
scale each datapoint differently. The amount that each datapoint contributes
to the final value is extracted from the document, or provided by a script.
As a formula, a weighted average is the `∑(value * weight) / ∑(weight)`
A regular average can be thought of as a weighted average where every value has
an implicit weight of `1`.
Closes#15731
In #29623 we added `Request` object flavored requests to the low level
REST client and in #30315 we deprecated the old `performRequest`s. This
changes all calls in the `x-pack/plugin/monitoring` project to use the new
versions.
In #29623 we added `Request` object flavored requests to the low level
REST client and in #30315 we deprecated the old `performRequest`s. This
changes all calls in the `x-pack:core` project to use the new versions.
New data is reported from Beats to the monitoring endpoint. This PR adds the template change necessary for it. See https://github.com/elastic/beats/issues/7521 for more details.
Queue data is skipped for now as implementation is not finished yet.
* Remove BouncyCastle dependency from runtime
This commit introduces a new gradle project that contains
the classes that have a dependency on BouncyCastle. For
the default distribution, It builds a jar from those and
in puts it in a subdirectory of lib
(/tools/security-cli) along with the BouncyCastle jars.
This directory is then passed in the
ES_ADDITIONAL_CLASSPATH_DIRECTORIES of the CLI tools
that use these classes.
BouncyCastle is removed as a runtime dependency (remains
as a compileOnly one) from x-pack core and x-pack security.
The initial decision to use async durability was made a long time ago
for performance reasons. That argument no longer applies and we
prefer the safety of request durability.
Resolving wildcards in aliases expression is challenging as we may end
up with no aliases to replace the original expression with, but if we
replace with an empty array that means _all which is quite the opposite.
Now that we support and serialize the original requested aliases,
whenever aliases are replaced we will be able to know what was
initially requested. `MetaData#findAliases` can then be updated to not
return anything in case it gets empty aliases, but the original aliases
were not empty. That means that empty aliases are interpreted as _all
only if they were originally requested that way.
Relates to #31516
Prior to 6.3 a trial license default to security enabled. Since 6.3
they default to security disabled. If a cluster is upgraded from <6.3
to >6.3, then we detect this and mimic the old behaviour with respect
to security.
Now write operations like Index, Delete, Update rely on the write-index associated with
an alias to operate against. This means writes will be accepted even when an alias points to multiple indices, so long as one is the write index. Routing values will be used from the AliasMetaData for the alias in the write-index. All read operations are left untouched.
Relates #29827
This implementation behaves like the current transport client, that you basically cannot configure a Watch POJO representation as an argument to the put watch API, but only a bytes reference. You can use the the `WatchSourceBuilder` from the `org.elasticsearch.plugin:x-pack-core` dependency to build watches.
This commit also changes the license type to trial, so that watcher is available in high level rest client tests.
/cc @hub-cap
The rollup indexer uses a range query to select the next page
of results based on the last time bucket of the previous round
and the `delay` configured on the rollup job. This query uses
the `epoch_millis` format implicitly but doesn't set the `format`.
This result in errors during the rollup job if the field
definition doesn't allow this format. It can also miss documents
if the format is not accepted but another format in the field
definition is able to parse the query (e.g.: `epoch_second`).
This change ensures that we use `epoch_millis` as the only format
to parse the rollup range query.
This is related to #32122. A number of things changed related to adding
TLS 1.3 support in JDK11. Some exception messages and other SSLEngine
behavior changed. This commit fixes assertions on exception messages.
Additionally it identifies two bugs related to how the SSLDriver behaves
in regards to JDK11 changes. Finally, it mutes a tests until correct
behavior can be identified. There is another open issue for that muted
test (#32144).
* Add basic support for field aliases in index mappings. (#31287)
* Allow for aliases when fetching stored fields. (#31411)
* Add tests around accessing field aliases in scripts. (#31417)
* Add documentation around field aliases. (#31538)
* Add validation for field alias mappings. (#31518)
* Return both concrete fields and aliases in DocumentFieldMappers#getMapper. (#31671)
* Make sure that field-level security is enforced when using field aliases. (#31807)
* Add more comprehensive tests for field aliases in queries + aggregations. (#31565)
* Remove the deprecated method DocumentFieldMappers#getFieldMapper. (#32148)
Moves the customizations to the build to produce nice shadow jars and
javadocs into common build code, mostly BuildPlugin with a little into
the root build.gradle file. This means that any project that applies the
shadow plugin will automatically be set up just like the high level rest
client:
* The non-shadow jar will not be built
* The shadow jar will not have a "classifier"
* Tests will run against the shadow jar
* Javadoc will include all of the shadowed classes
* Service files in `META-INF/services` will be merged
The ML config classes will shortly be moved to the X-Pack protocol
library to allow the ML APIs to be moved to the high level REST
client. Dependencies on server functionality should be removed
from the config classes before this is done.
This change is entirely about moving code between packages. It
does not add or remove any functionality or tests.
Ensure our tests can run in a FIPS JVM
JKS keystores cannot be used in a FIPS JVM as attempting to use one
in order to init a KeyManagerFactory or a TrustManagerFactory is not
allowed.( JKS keystore algorithms for private key encryption are not
FIPS 140 approved)
This commit replaces JKS keystores in our tests with the
corresponding PEM encoded key and certificates both for key and trust
configurations.
Whenever it's not possible to refactor the test, i.e. when we are
testing that we can load a JKS keystore, etc. we attempt to
mute the test when we are running in FIPS 140 JVM. Testing for the
JVM is naive and is based on the name of the security provider as
we would control the testing infrastrtucture and so this would be
reliable enough.
Other cases of tests being muted are the ones that involve custom
TrustStoreManagers or KeyStoreManagers, null TLS Ciphers and the
SAMLAuthneticator class as we cannot sign XML documents in the
way we were doing. SAMLAuthenticator tests in a FIPS JVM can be
reenabled with precomputed and signed SAML messages at a later stage.
IT will be covered in a subsequent PR
This introduces a new GetRollupIndexCaps API which allows the user to retrieve rollup capabilities of a specific rollup index (or index pattern). This is distinct from the existing RollupCaps endpoint.
- Multiple jobs can be stored in multiple indices and point to a single target data index pattern (logstash-*). The existing API finds capabilities/config of all jobs matching that data index pattern.
- One rollup index can hold data from multiple jobs, targeting multiple data index patterns. This new API finds the capabilities based on the concrete rollup indices.
Due to the way ANTLR works, any declared tokens need to be accounted for
manually inside function names (otherwise a different rule gets applied).
Fix#32046
There is currently no way to see what user executed a watch. This commit
adds the decrypted username to each execution in the watch history, in a
new field "user".
Closes#31772
The old RollupIT was a node IT, an flaky for a number of reasons.
This new version is an ESRestTestCase and should be a little more robust.
This was added to the multi-node QA tests as that seemed like the most
appropriate location. It didn't seem necessary to create a whole new
QA module.
Note: The only test that was ported was the "Big" test for validating
a larger dataset. The rest of the tests are represented in existing
yaml tests.
Closes#31258Closes#30232
Related to #30290
Ensure that the same algorithm is used for settings and
change password requests for consistency, even if we
do not expext to reach the code where the algorithm is
checked for now.
Completes a7eaa409e8
The build was broken due to some issues with the merging of #32018. A
method that was public went private before the PR was merged. That did
not cause a merge conflict (so the PR was merged successfully). But it
did cause the build to fail.
This is related to #27260. It adds the SecurityNioHttpServerTransport
to the security plugin. It randomly uses the nio http transport in
security integration tests.
Metric config already whitelist scaled_floats, but it wasn't added to
the histo group config. This centralizes the mapping types map
so that both metrics and histo (and any future configs) use the same
map.
Fixes#32035
This commit allows for rebuilding watcher secure secrets via the
reload_secure_settings API call. The commit also renames a method in the
Notification Service to make it a bit more readable.
This commit adds the _xpack/usage api to the high level rest client.
Currently in the transport api, the usage data is exposed in a limited
fashion, at most giving one level of helper methods for the inner keys
of data, but then exposing thos subobjects as maps of objects. Rather
than making parsers for every set of usage data from each feature, this
PR exposes the entire set of usage data as a map of maps.
Other watcher actions already account for secure settings in their
sensitive settings, whereas the email sending action did not. This adds
the ability to optionally set a secure_password for email accounts.
Previously, the ensureWatchExists was overridable. This commit makes
it final so that it cannot be overridden, and cleans up some redundant
code in the process.
There was still a case with a null text that allowed for 0 attachments
to be created. This commit ensures that greater than zero are created
if the text is null. Otherwise, it uses the same logic to create 0 to 3
random attachments.
Closes#31948
We can leverage the composite agg's new `missing_bucket` feature on
terms groupings. This means the aggregation criteria used in the indexer
will now return null buckets for missing keys.
Because all buckets are now returned (even if a key is null),
we can guarantee correct doc counts with
"combined" jobs (where a job rolls up multiple schemas). This was
previously impossible since composite would ignore documents that
didn't have _all_ the keys, meaning non-overlapping schemas would
cause composite to return no buckets.
Note: date_histo does not use `missing_bucket`, since a timestamp is
always required.
The docs have been adjusted to recommend a single, combined job. It
also makes reference to the previous issue to help users that are upgrading
(rather than just deleting the sections).
The previous errors in compileJava were not cause by the brackets but my the
content of the @link section. Corrected this so its a working javadoc link again.
When an ML job cannot be allocated to a node the exception
contained an explanation of why the job couldn't be
allocated to each node in the cluster. For large clusters
this was not particularly easy to read and made the error
displayed in the UI look very scary.
This commit changes the structure of the error to an outer
ElasticsearchException with a high level message and an
inner IllegalStateException containing the detailed
explanation. Because the definition of root cause is the
innermost ElasticsearchException the detailed explanation
will not be the root cause (which is what Kibana displays).
Fixes#29950
Historically we have loaded SSL objects (such as SSLContext,
SSLIOSessionStrategy) by passing in the SSL settings, constructing a
new SSL configuration from those settings and then looking for a
cached object that matches those settings.
The primary issue with this approach is that it requires a fully
configured Settings object to be available any time the SSL context
needs to be loaded. If the Settings include SecureSettings (such as
passwords for keys or keystores) then this is not true, and the cached
SSL object cannot be loaded at runtime.
This commit introduces an alternative approach of naming every cached
ssl configuration, so that it is possible to load the SSL context for
a named configuration (such as "xpack.http.ssl"). This means that the
calling code does not need to have ongoing access to the secure
settings that were used to load the configuration.
This change also allows monitoring exporters to use SSL passwords
from secure settings, however an exporter that uses a secure SSL setting
(e.g. truststore.secure_password) may not have its SSL settings updated
dynamically (this is prevented by a settings validator).
Exporters without secure settings can continue to be defined and updated
dynamically.
This is related to #27260. It adds the SecurityNioTransport to the
security plugin. Additionally, it adds support for ip filtering. And it
randomly uses the nio transport in security integration tests.
Added support for ASCII, BIT_LENGTH, CHAR, CHAR_LENGTH, LCASE, LENGTH, LTRIM, RTRIM, SPACE, UCASE functions.
Wherever Painless scripting is necessary (WHERE conditions, ORDER BY etc), those scripts are being used.
The test failure in #31916 revealed that updating
rules on a job was modifying the detectors list
in-place. That meant the old cluster state and the
updated cluster state had no difference and thus the
change was not propagated to non-master nodes.
This commit fixes that and also reviews all of ML
metadata in order to ensure immutability.
Closes#31916
A new commit was merged that does not allow a null attachement &&
text. This is valid for the slack API, as it does not allow this, but
our unit tests did. This commit fixes the broken unit test.
Closes#31948
Currently Role.Builder keeps a reference to the FieldPermissionsCache that is
passed into its constructors. This seems to be unused except for passing it on
to convertFromIndicesPrivileges() in the second ctor itself, but we don't need
to keep the internal reference in that case, so it can be removed.
Relates to #31876
* Adding Beats x-pack plugin + index templates
* Adding built-in roles for Beats central management
* Fixing typo
* Refactoring: extract common code into method
* More refactoring for more code reuse
* Use a single index for Beats management
* Rename "fragment" to "block"
* Adding configuration block type
* Expand kibana_system role to include Beats management index privileges
* Fixing syntax
* Adding test
* Adding asserting for reserved role
* Fixing privileges
* Updating template
* Removing beats plugin
* Fixing tests
* Fixing role variable name
* Fixing assertions
* Switching to preferred syntax for boolean false checks
* Making class final
* Making variables final
* Updating Basic license message to be more accurate
Originally I put the X-Pack info object into the top level rest client
object. I did that because we thought we'd like to squash `xpack` from
the name of the X-Pack APIs now that it is part of the default
distribution. We still kind of want to do that, but at least for now we
feel like it is better to keep the high level rest client aligned with
the other language clients like C# and Python. This shifts the X-Pack
info API to align with its json spec file.
Relates to #31870
Slack accepts an empty text or attachments, but not both. This commit
ensures that both are not empty when creating a watch.
Closes#30071
Replacing old pull request: #31288
The steps to read the settings and build URLs happen in a non-obvious
order, which meant that we would build the default URL (from the
domain name, and port) before we'd actually read the port settings.
This would cause the URL to always have a port of `0`.
Relates: bccf988
With https://github.com/elastic/beats/pull/7075 Beats introduces state reporting for X-Pack Monitoring. The data sent up to Elasticsearch ends up stored in the following format.
```
"beats_state": {
"timestamp": "2018-07-05T07:21:03.581Z",
"state": {
"module": {
"count": 1,
"names": [
"http"
]
}
},
"beat": {
"uuid": "594039b5-6353-4d78-9bad-778ecc0fe83f",
"type": "metricbeat",
"version": "7.0.0-alpha1",
"name": "ruflin",
"host": "ruflin"
}
}
```
This PR adds the new fields to the template.
This is the first x-pack API we're adding to the high level REST client
so there is a lot to talk about here!
= Open source
The *client* for these APIs is open source. We're taking the previously
Elastic licensed files used for the `Request` and `Response` objects and
relicensing them under the Apache 2 license.
The implementation of these features is staying under the Elastic
license. This lines up with how the rest of the Elasticsearch language
clients work.
= Location of the new files
We're moving all of the `Request` and `Response` objects that we're
relicensing to the `x-pack/protocol` directory. We're adding a copy of
the Apache 2 license to the root fo the `x-pack/protocol` directory to
line up with the language in the root `LICENSE.txt` file. All files in
this directory will have the Apache 2 license header as well. We don't
want there to be any confusion. Even though the files are under the
`x-pack` directory, they are Apache 2 licensed.
We chose this particular directory layout because it keeps the X-Pack
stuff together and easier to think about.
= Location of the API in the REST client
We've been following the layout of the rest-api-spec files for other
APIs and we plan to do this for the X-Pack APIs with one exception:
we're dropping the `xpack` from the name of most of the APIs. So
`xpack.graph.explore` will become `graph().explore()` and
`xpack.license.get` will become `license().get()`.
`xpack.info` and `xpack.usage` are special here though because they
don't belong to any proper category. For now I'm just calling
`xpack.info` `xPackInfo()` and intend to call usage `xPackUsage` though
I'm not convinced that this is the final name for them. But it does get
us started.
= Jars, jars everywhere!
This change makes the `xpack:protocol` project a `compile` scoped
dependency of the `x-pack:plugin:core` and `client:rest-high-level`
projects. I intend to keep it a compile scoped dependency of
`x-pack:plugin:core` but I intend to bundle the contents of the protocol
jar into the `client:rest-high-level` jar in a follow up. This change
has grown large enough at this point.
In that followup I'll address javadoc issues as well.
= Breaking-Java
This breaks that transport client by a few classes around. We've
traditionally been ok with doing this to the transport client.
For historical reasons SQL restricts GROUP BY to only one field.
This commit removes the restriction and improves the test suite with
multi group by tests.
Close#31793
If a get alias api call requests a specific alias pattern then
indices not having any matching aliases should not be included in the response.
This is a second attempt to fix this (first attempt was #28294).
The reason that the first attempt was reverted is because when xpack
security is enabled then index expression (like * or _all) are resolved
prior to when a request is processed in the get aliases transport action,
then `MetaData#findAliases` can't know whether requested all where
requested since it was already expanded in concrete alias names. This
change replaces aliases(...) replaceAliases(...) method on AliasesRequests
class and leave the aliases(...) method on subclasses. So there is a distinction
between when xpack security replaces aliases and a user setting aliases via
the transport or high level http client.
Closes#27763
Removes support for storing scripts without the usual json around the
script. So You can no longer do:
```
POST _scripts/<templatename>
{
"query": {
"match": {
"title": "{{query_string}}"
}
}
}
```
and must instead do:
```
POST _scripts/<templatename>
{
"script": {
"lang": "mustache",
"source": {
"query": {
"match": {
"title": "{{query_string}}"
}
}
}
}
}
```
This improves error reporting when you attempt to store a script but don't
quite get the syntax right. Before, there was a good chance that we'd
think of it as a "raw" template and just store it. Now we won't do that.
Nice.
It seems that java 11 tightened some validations with regard to
time formats. The random instance creator was setting an odd
time format to the data description which is invalid when run
with java 11. This commit changes it to a valid format.
Job persistent tasks with stale allocation IDs used to always be
considered as OPENING jobs in the ML job node allocation decision.
However, FAILED jobs are not relocated to other nodes, which leads
to them blocking up the nodes they failed on after node restarts.
FAILED jobs should not restrict how many other jobs can open on a
node, regardless of whether they are stale or not.
Closes#31794
Job updates or changes to calendars or filters may
result into updating the job process if it has been
running. To preserve the order of updates, process
updates are queued through the UpdateJobProcessNotifier
which is only running on the master node. All actions
performing such updates must run on the master node.
However, the CRUD actions for calendars and filters
are not master node actions. They have been submitting
the updates to the UpdateJobProcessNotifier even though
it might have not been running (given the action was
run on a non-master node). When that happens, the update
never reaches the process.
This commit fixes this problem by ensuring the notifier
runs on all nodes and by ensuring the process update action
gets the resources again before updating the process
(instead of having those resources passed in the request).
This ensures that even if the order of the updates
gets messed up, the latest update will read the latest
state of those resource and the process will get back
in sync.
This leaves us with 2 types of updates:
1. updates to the job config should happen on the master
node. This is because we cannot refetch the entire job
and update it. We need to know the parts that have been changed.
2. updates to resources the job uses. Those can be handled
on non-master nodes but they should be re-fetched by the
update process action.
Closes#31803
Yannick Welsch
Yogesh Gaikwad
Tim Vernum
Nik Everett
David Roberts
Jim Ferenczi
Jack Conradson
Christoph Büscher
Jason Tedor
Zachary Tong
Igor Motov
Jay Modi
Lee Hinman
Ioannis Kakavas
David Kyle
Andy Bristol
Nicolas Ruflin
Alexander Reelsen
Luca Cavanna
Tal Levy
Ryan Ernst
Tim Brooks
Julie Tibshirani
Ioannis Kakavas
Armin Braun
Costin Leau
Michael Basnight
Tim Vernum
Andrei Stefan
Dimitris Athanasiou
Shaunak Kashyap
albendz
hanbj
Martijn van Groningen
Sohaib Iftikhar
Vladimir Dolzhenko