The existing privilege model for API keys with privileges like
`manage_api_key`, `manage_security` etc. are too permissive and
we would want finer-grained control over the cluster privileges
for API keys. Previously APIs created would also need these
privileges to get its own information.
This commit adds support for `manage_own_api_key` cluster privilege
which only allows api key cluster actions on API keys owned by the
currently authenticated user. Also adds support for retrieval of
the API key self-information when authenticating via API key
without the need for the additional API key privileges.
To support this privilege, we are introducing additional
authentication context along with the request context such that
it can be used to authorize cluster actions based on the current
user authentication.
The API key get and invalidate APIs introduce an `owner` flag
that can be set to true if the API key request (Get or Invalidate)
is for the API keys owned by the currently authenticated user only.
In that case, `realm` and `username` cannot be set as they are
assumed to be the currently authenticated ones.
The changes cover HLRC changes, documentation for the API changes.
Closes#40031
This commit introduces PKI realm delegation. This feature
supports the PKI authentication feature in Kibana.
In essence, this creates a new API endpoint which Kibana must
call to authenticate clients that use certificates in their TLS
connection to Kibana. The API call passes to Elasticsearch the client's
certificate chain. The response contains an access token to be further
used to authenticate as the client. The client's certificates are validated
by the PKI realms that have been explicitly configured to permit
certificates from the proxy (Kibana). The user calling the delegation
API must have the delegate_pki privilege.
Closes#34396
This commit allows the Transport Actions for the SSO realms to
indicate the realm that should be used to authenticate the
constructed AuthenticationToken. This is useful in the case that
many authentication realms of the same type have been configured
and where the caller of the API(Kibana or a custom web app) already
know which realm should be used so there is no need to iterate all
the realms of the same type.
The realm parameter is added in the relevant REST APIs as optional
so as not to introduce any breaking change.
Most of our CLI tools use the Terminal class, which previously did not provide methods for writing to standard output. When all output goes to standard out, there are two basic problems. First, errors and warnings are "swallowed" in pipelines, making it hard for a user to know when something's gone wrong. Second, errors and warnings are intermingled with legitimate output, making it difficult to pass the results of interactive scripts to other tools.
This commit adds a second set of print commands to Terminal for printing to standard error, with errorPrint corresponding to print and errorPrintln corresponding to println. This leaves it to developers to decide which output should go where. It also adjusts existing commands to send errors and warnings to stderr.
Usage is printed to standard output when it's correctly requested (e.g., bin/elasticsearch-keystore --help) but goes to standard error when a command is invoked incorrectly (e.g. bin/elasticsearch-keystore list-with-a-typo | sort).
The current implementations make it difficult for
adding new privileges (example: a cluster privilege which is
more than cluster action-based and not exposed to the security
administrator). On the high level, we would like our cluster privilege
either:
- a named cluster privilege
This corresponds to `cluster` field from the role descriptor
- or a configurable cluster privilege
This corresponds to the `global` field from the role-descriptor and
allows a security administrator to configure them.
Some of the responsibilities like the merging of action based cluster privileges
are now pushed at cluster permission level. How to implement the predicate
(using Automaton) is being now enforced by cluster permission.
`ClusterPermission` helps in enforcing the cluster level access either by
performing checks against cluster action and optionally against a request.
It is a collection of one or more permission checks where if any of the checks
allow access then the permission allows access to a cluster action.
Implementations of cluster privilege must be able to provide information
regarding the predicates to the cluster permission so that can be enforced.
This is enforced by making implementations of cluster privilege aware of
cluster permission builder and provide a way to specify how the permission is
to be built for a given privilege.
This commit renames `ConditionalClusterPrivilege` to `ConfigurableClusterPrivilege`.
`ConfigurableClusterPrivilege` is a renderable cluster privilege exposed
as a `global` field in role descriptor.
Other than this there is a requirement where we would want to know if a cluster
permission is implied by another cluster-permission (`has-privileges`).
This is helpful in addressing queries related to privileges for a user.
This is not just simply checking of cluster permissions since we do not
have access to runtime information (like request object).
This refactoring does not try to address those scenarios.
Relates #44048
* Restrict which tasks can use testclusters
This PR fixes a problem between the interaction of test-clusters and
build cache.
Before this any task could have used a cluster without tracking it as
input.
With this change a new interface is introduced to track the tasks that
can use clusters and we do consider the cluster as input for all of
them.
When using the implicit flow in OpenID Connect, the
op.token_endpoint_url should not be mandatory as there is no need
to contact the token endpoint of the OP.
Uses JDK 11's per-socket configuration of TCP keepalive (supported on Linux and Mac), see
https://bugs.openjdk.java.net/browse/JDK-8194298, and exposes these as transport settings.
By default, these options are disabled for now (i.e. fall-back to OS behavior), but we would like
to explore whether we can enable them by default, in particular to force keepalive configurations
that are better tuned for running ES.
Currently in the transport-nio work we connect and bind channels on the
a thread before the channel is registered with a selector. Additionally,
it is at this point that we set all the socket options. This commit
moves these operations onto the event-loop after the channel has been
registered with a selector. It attempts to set the socket options for a
non-server channel at registration time. If that fails, it will attempt
to set the options after the channel is connected. This should fix
#41071.
This change improves the exception messages that are thrown when the
system cannot read TLS resources such as keystores, truststores,
certificates, keys or certificate-chains (CAs).
This change specifically handles:
- Files that do not exist
- Files that cannot be read due to file-system permissions
- Files that cannot be read due to the ES security-manager
Backport of: #44787
There are no realms that can be configured exclusively with secure
settings. Every realm that supports secure settings also requires one
or more non-secure settings.
However, sometimes a node will be configured with entries in the
keystore for which there is nothing in elasticsearch.yml - this may be
because the realm we removed from the yml, but not deleted from the
keystore, or it could be because there was a typo in the realm name
which has accidentially orphaned the keystore entry.
In these cases the realm building would fail, but the error would not
always be clear or point to the root cause (orphaned keystore
entries). RealmSettings would act as though the realm existed, but
then fail because an incorrect combination of settings was provided.
This change causes realm building to fail early, with an explicit
message about incorrect keystore entries.
Backport of: #44471
When we create API key we check if the API key with the name
already exists. It searches with scroll enabled and this causes
the request to fail when creating large number of API keys in
parallel as it hits the number of open scroll limit (default 500).
We do not need the search context to be created so this commit
removes the scroll parameter from the search request for duplicate
API key.
* We shouldn't be recreating wrapped REST handlers over and over for every request. We only use this hook in x-pack and the wrapper there does not have any per request state.
This is inefficient and could lead to some very unexpected memory behavior
=> I made the logic create the wrapper on handler registration and adjusted the x-pack wrapper implementation to correctly forward the circuit breaker and content stream flags
A mismatched configuration between the IdP and SP will often result in
SAML authentication attempts failing because the audience condition is
not met (because the IdP and SP disagree about the correct form of the
SP's Entity ID).
Previously the error message in this case did not provide sufficient
information to resolve the issue because the IdP's expected audience
would be truncated if it exceeeded 32 characters. Since the error did
not provide both IDs in full, it was not possible to determine the
correct fix (in detail) based on the error alone.
This change expands the message that is included in the thrown
exception, and also adds additional logging of every failed audience
condition, with diagnostics of the match failure.
Backport of: #44334
The existing equals check was broken, and would always be false.
The correct behaviour is to return "Collections.emptyList()" whenever
the the active(licensed)-realms equals the configured-realms.
Backport of: #44399
We often start testing with early access versions of new Java
versions and this have caused minor issues in our tests
(i.e. #43141) because the version string that the JVM reports
cannot be parsed as it ends with the string -ea.
This commit changes how we parse and compare Java versions to
allow correct parsing and comparison of the output of java.version
system property that might include an additional alphanumeric
part after the version numbers
(see [JEP 223[(https://openjdk.java.net/jeps/223)). In short it
handles a version number part, like before, but additionally a
PRE part that matches ([a-zA-Z0-9]+).
It also changes a number of tests that would attempt to parse
java.specification.version in order to get the full version
of Java. java.specification.version only contains the major
version and is thus inappropriate when trying to compare against
a version that might contain a minor, patch or an early access
part. We know parse java.version that can be consistently
parsed.
Resolves#43141
Registering a channel with a selector is a required operation for the
channel to be handled properly. Currently, we mix the registeration with
other setup operations (ip filtering, SSL initiation, etc). However, a
fail to register is fatal. This PR modifies how registeration occurs to
immediately close the channel if it fails.
There are still two clear loopholes for how a user can interact with a
channel even if registration fails. 1. through the exception handler.
2. through the channel accepted callback. These can perhaps be improved
in the future. For now, this PR prevents writes from proceeding if the
channel is not registered.
This commit converts all the StreamableResponseActionType security
classes in xpack core to ActionType, implementing Writeable for their
response classes.
relates #34389
When getting authentication info from the thread context, it might be
that we encounter an I/O exception. Today we swallow this exception and
return a null authentication info to the caller. Yet, this could be
hiding bugs or errors. This commits adjusts this behavior so that we no
longer swallow the exception.
Test clusters currently has its own set of logic for dealing with
finding different versions of Elasticsearch, downloading them, and
extracting them. This commit converts testclusters to use the
DistributionDownloadPlugin.
This commit moves the Supplier variant of HandledTransportAction to have
a different ordering than the Writeable.Reader variant. The Supplier
version is used for the legacy Streamable, and currently having the
location of the Writeable.Reader vs Supplier in the same place forces
using casts of Writeable.Reader to select the correct super constructor.
This change in ordering allows easier migration to Writeable.Reader.
relates #34389
Fixes a bug in the PKI authentication. This manifests when there
are multiple PKI realms configured in the chain, with different
principal parse patterns. There are a few configuration scenarios
where one PKI realm might parse the principal from the Subject
DN (according to the `username_pattern` realm setting) but
another one might do the truststore validation (according to
the truststore.* realm settings).
This is caused by the two passes through the realm chain, first to
build the authentication token and secondly to authenticate it, and
that the X509AuthenticationToken sets the principal during
construction.
Simplifies AbstractSimpleTransportTestCase to use JVM-local ports and also adds an assertion so
that cases like #44134 can be more easily debugged. The likely reason for that one is that a test,
which was repeated again and again while always spawning a fresh Gradle worker (due to Gradle
daemon) kept increasing Gradle worker IDs, causing an overflow at some point.
The base classes for transport requests and responses currently
implement Streamable and Writeable. The writeTo method on these base
classes is implemented with an empty implementation. Not only does this
complicate subclasses to think they need to call super.writeTo, but it
also can lead to not implementing writeTo when it should have been
implemented, or extendiong one of these classes when not necessary,
since there is nothing to actually implement.
This commit removes the empty writeTo from these base classes, and fixes
subclasses to not call super and in some cases implement an empty
writeTo themselves.
relates #34389
This commit converts the ConnectionManager's openConnection and connectToNode methods to
async-style. This will allow us to not block threads anymore when opening connections. This PR also
adapts the cluster coordination subsystem to make use of the new async APIs, allowing to remove
some hacks in the test infrastructure that had to account for the previous synchronous nature of the
connection APIs.
This commit deprecates the `transport.profiles.*.xpack.security.type`
setting. This setting is used to configure a profile that would only
allow client actions. With the upcoming removal of the transport client
the setting should also be deprecated so that it may be removed in
a future version.
All valid licenses permit security, and the only license state where
we don't support security is when there is a missing license.
However, for safety we should attach the system (or xpack/security)
user to internally originated actions even if the license is missing
(or, more strictly, doesn't support security).
This allows all nodes to communicate and send internal actions (shard
state, handshake/pings, etc) even if a license is transitioning
between a broken state and a valid state.
Relates: #42215
Backport of: #43468
Document level security was depending on the shared
"BitsetFilterCache" which (by design) never expires its entries.
However, when using DLS queries - particularly templated ones - the
number (and memory usage) of generated bitsets can be significant.
This change introduces a new cache specifically for BitSets used in
DLS queries, that has memory usage constraints and access time expiry.
The whole cache is automatically cleared if the role cache is cleared.
Individual bitsets are cleared when the corresponding lucene index
reader is closed.
The cache defaults to 50MB, and entries expire if unused for 7 days.
Backport of: #43669
If an item in the bulk request fails, that could be for a variety of
reasons - it may be that the underlying behaviour of security has
changed, or it may just be a transient failure during testing.
Simply asserting a `true`/`false` value produces failure messages that
are difficult to diagnose and debug. Using hamcert (`assertThat`) will
make it easier to understand the causes of failures in this test.
Backport of: #43725
This adds a new cluster privilege for manage_api_key. Users with this
privilege are able to create new API keys (as a child of their own
user identity) and may also get and invalidate any/all API keys
(including those owned by other users).
Backport of: #43728
As defined in https://tools.ietf.org/html/rfc6749#section-2.3.1
both client id and client secret need to be encoded with the
application/x-www-form-urlencoded encoding algorithm when used as
credentials for HTTP Basic Authentication in requests to the OP.
Resolves#43709
This is an odd backport of #41774
UserRoleMapper.UserData is constructed by each realm and it is used to
"match" role mapping expressions that eventually supply the role names
of the principal.
This PR filters out `null` collection values (lists and maps), for the groups
and metadata, which get to take part in the role mapping, in preparation
for using Java 9 collection APIs. It filters them as soon as possible, during
the construction.
Action is a class that encapsulates meta information about an action
that allows it to be called remotely, specifically the action name and
response type. With recent refactoring, the action class can now be
constructed as a static constant, instead of needing to create a
subclass. This makes the old pattern of creating a singleton INSTANCE
both misnamed and lacking a common placement.
This commit renames Action to ActionType, thus allowing the old INSTANCE
naming pattern to be TYPE on the transport action itself. ActionType
also conveys that this class is also not the action itself, although
this change does not rename any concrete classes as those will be
removed organically as they are converted to TYPE constants.
relates #34389
This change removes the ability to wrap an IndexSearcher in plugins. The IndexSearcherWrapper is replaced by an IndexReaderWrapper and allows to wrap the DirectoryReader only. This simplifies the creation of the context IndexSearcher that is used on a per request basis. This change also moves the optimization that was implemented in the security index searcher wrapper to the ContextIndexSearcher that now checks the live docs to determine how the search should be executed. If the underlying live docs is a sparse bit set the searcher will compute the intersection
betweeen the query and the live docs instead of checking the live docs on every document that match the query.
TransportNodesAction provides a mechanism to easily broadcast a request
to many nodes, and collect the respones into a high level response. Each
node has its own request type, with a base class of BaseNodeRequest.
This base request requires passing the nodeId to which the request will
be sent. However, that nodeId is not used anywhere. It is private to the
base class, yet serialized to each node, where the node could just as
easily find the nodeId of the node it is on locally.
This commit removes passing the nodeId through to the node request
creation, and guards its serialization so that we can remove the base
request class altogether in the future.
* Use atomic boolean to guard wakeups
* Don't trigger wakeups from the select loops thread itself for registering and closing channels
* Don't needlessly queue writes
Co-authored-by: Tim Brooks <tim@uncontended.net>
Currently nio implements ip filtering at the channel context level. This
is kind of a hack as the application logic should be implemented at the
handler level. This commit moves the ip filtering into a channel
handler. This requires adding an indicator to the channel handler to
show when a channel should be closed.
This replaces the use of char[] in the password length validation
code, with the use of SecureString
Although the use of char[] is not in itself problematic, using a
SecureString encourages callers to think about the lifetime of the
password object and to clear it after use.
Backport of: #42884
This commit removes some very old test logging annotations that appeared
to be added to investigate test failures that are long since closed. If
these are needed, they can be added back on a case-by-case basis with a
comment associating them to a test failure.
Kibana wants to create access_token/refresh_token pair using Token
management APIs in exchange for kerberos tickets. `client_credentials`
grant_type requires every user to have `cluster:admin/xpack/security/token/create`
cluster privilege.
This commit introduces `_kerberos` grant_type for generating `access_token`
and `refresh_token` in exchange for a valid base64 encoded kerberos ticket.
In addition, `kibana_user` role now has cluster privilege to create tokens.
This allows Kibana to create access_token/refresh_token pair in exchange for
kerberos tickets.
Note:
The lifetime from the kerberos ticket is not used in ES and so even after it expires
the access_token/refresh_token pair will be valid. Care must be taken to invalidate
such tokens using token management APIs if required.
Closes#41943
* TestClusters: Convert the security plugin
This PR moves security tests to use TestClusters.
The TLS test required support in testclusters itself, so the correct
wait condition is configgured based on the cluster settings.
* PR review
The description field of xpack featuresets is optionally part of the
xpack info api, when using the verbose flag. However, this information
is unnecessary, as it is better left for documentation (and the existing
descriptions describe anything meaningful). This commit removes the
description field from feature sets.
It turns out that key rotation on the OP, can manifest as both
a BadJWSException and a BadJOSEException in nimbus-jose-jwt. As
such we cannot depend on matching only BadJWSExceptions to
determine if we should poll the remote JWKs for an update.
This has the side-effect that a remote JWKs source will be polled
exactly one additional time too for errors that have to do with
configuration, or for errors that might be caused by not synched
clocks, forged JWTs, etc. ( These will throw a BadJWTException
which extends BadJOSEException also )
WriteActionsTests#testBulk and WriteActionsTests#testIndex sometimes
fail with a pending retention lock. We might leak retention locks when
switching to async recovery. However, it's more likely that ongoing
recoveries prevent the retention lock from releasing.
This change increases the waiting time when we check for no pending
retention lock and also ensures no ongoing recovery in
WriteActionsTests.
Closes#41054
Kibana alerting is going to be built using API Keys, and should be
permitted on a basic license.
This commit moves API Keys (but not Tokens) to the Basic license
Relates: elastic/kibana#36836
Backport of: #42787
Currently, when the SSLEngine needs to produce handshake or close data,
we must manually call the nonApplicationWrite method. However, this data
is only required when something triggers the need (starting handshake,
reading from the wire, initiating close, etc). As we have a dedicated
outbound buffer, this data can be produced automatically. Additionally,
with this refactoring, we combine handshake and application mode into a
single mode. This is necessary as there are non-application messages that
are sent post handshake in TLS 1.3. Finally, this commit modifies the
SSLDriver tests to test against TLS 1.3.
Enable audit logs in docker by creating console appenders for audit loggers.
also rename field @timestamp to timestamp and add field type with value audit
The docker build contains now two log4j configuration for oss or default versions. The build now allows override the default configuration.
Also changed the format of a timestamp from ISO8601 to include time zone as per this discussion #36833 (comment)
closes#42666
backport#42671
This commit fixes the version parsing in various tests. The issue here is that
the parsing was relying on java.version. However, java.version can contain
additional characters such as -ea for early access builds. See JEP 233:
Name Syntax
------------------------------ --------------
java.version $VNUM(\-$PRE)?
java.runtime.version $VSTR
java.vm.version $VSTR
java.specification.version $VNUM
java.vm.specification.version $VNUM
Instead, we want java.specification.version.
This commit removes the TLS cluster join validator.
This validator existed to prevent v6.x nodes (which mandated
TLS) from joining an existing cluster of v5.x nodes (which did
not mandate TLS) unless the 6.x node (and by implication the
5.x nodes) was configured to use TLS.
Since 7.x nodes cannot talk to 5.x nodes, this validator is no longer
needed.
Removing the validator solves a problem where single node clusters
that were bound to local interfaces were incorrectly requiring TLS
when they recovered cluster state and joined their own cluster.
Backport of: #42826
Whether security is enabled/disabled is dependent on the combination
of the node settings and the cluster license.
This commit adds a license state listener that logs when the license
change causes security to switch state (or to be initialised).
This is primarily useful for diagnosing cluster formation issues.
Backport of: #42488
If the security index is closed, it should be treated as unavailable
for security purposes.
Prior to 8.0 (or in a mixed cluster) a closed security index has
no routing data, which would cause a NPE in the cluster change
handler, and the index state would not be updated correctly.
This commit fixes that problem
Backport of: #42191
This commit clones the existing AnalyzeRequest/AnalyzeResponse classes
to the high-level rest client, and adjusts request converters to use these new
classes.
This is a prerequisite to removing the Streamable interface from the internal
server version of these classes.
This commit changes the way token ids are hashed so that the output is
url safe without requiring encoding. This follows the pattern that we
use for document ids that are autogenerated, see UUIDs and the
associated classes for additional details.
This change ensures that:
- We only attempt to refresh the remote JWKS when there is a
signature related error only ( BadJWSException instead of the
geric BadJOSEException )
- We do call OpenIDConnectAuthenticator#getUserClaims upon
successful refresh.
- We test this in OpenIdConnectAuthenticatorTests.
Without this fix, when using the OpenID Connect realm with a remote
JWKSet configured in `op.jwks_path`, the refresh would be triggered
for most configuration errors ( i.e. wrong value for `op.issuer` )
and the kibana wouldn't get a response and timeout since
`getUserClaims` wouldn't be called because
`ReloadableJWKSource#reloadAsync` wouldn't call `onResponse` on the
future.
Test was using ClockMock#rewind passing the amount of nanoseconds
in order to "strip" nanos from the time value. This was intentional
as the expiration time of the UserToken doesn't have nanosecond
precision.
However, ClockMock#rewind doesn't support nanos either, so when it's
called with a TimeValue, it rewinds the clock by the TimeValue's
millis instead. This was causing the clock to go enough millis
before token expiration time and the test was passing. Once every
few hundred times though, the TimeValue by which we attempted to
rewind the clock only had nanos and no millis, so rewind moved the
clock back just a few millis, but still after expiration time.
This change moves the clock explicitly to the same instant as expiration,
using clock.setTime and disregarding nanos.
* Safer Wait for Snapshot Success in ClusterPrivilegeTests
* The snapshot state returned by the API might become SUCCESS before it's fully removed from the cluster state.
* We should fix this race in the transport API but it's not trivial and will be part of the incoming big round of refactoring the repository interaction, this added check fixes the test for now
* closes#38030
This commit makes creators of GetField split the fields into document fields and metadata fields. It is part of larger refactoring that aims to remove the calls to static methods of MapperService related to metadata fields, as discussed in #24422.
rp.client_secret is a required secure setting. Make sure we fail with
a SettingsException and a clear, actionable message when building
the realm, if the setting is missing.
Enhance the handling of merging the claims sets of the
ID Token and the UserInfo response. JsonObject#merge would throw a
runtime exception when attempting to merge two objects with the
same key and different values. This could happen for an OP that
returns different vales for the same claim in the ID Token and the
UserInfo response ( Google does that for profile claim ).
If a claim is contained in both sets, we attempt to merge the
values if they are objects or arrays, otherwise the ID Token claim
value takes presedence and overwrites the userinfo response.
SHA256 was recently added to the Hasher class in order to be used
in the TokenService. A few tests were still using values() to get
the available algorithms from the Enum and it could happen that
SHA256 would be picked up by these.
This change adds an extra convenience method
(Hasher#getAvailableAlgoCacheHash) and enures that only this and
Hasher#getAvailableAlgoStoredHash are used for getting the list of
available password hashing algorithms in our tests.
This performs a simple restart test to move a basic licensed
cluster from no security (the default) to security & transport TLS
enabled.
Backport of: #41933
If there are no realms that depend on the native role mapping store,
then changes should it should not perform any cache refresh.
A refresh with an empty realm array will refresh all realms.
This also fixes a spurious log warning that could occur if the
role mapping store was notified that the security index was recovered
before any realm were attached.
Backport of: #42169
This commit changes how access tokens and refresh tokens are stored
in the tokens index.
Access token values are now hashed before being stored in the id
field of the `user_token` and before becoming part of the token
document id. Refresh token values are hashed before being stored
in the token field of the `refresh_token`. The tokens are hashed
without a salt value since these are v4 UUID values that have
enough entropy themselves. Both rainbow table attacks and offline
brute force attacks are impractical.
As a side effect of this change and in order to support multiple
concurrent refreshes as introduced in #39631, upon refreshing an
<access token, refresh token> pair, the superseding access token
and refresh tokens values are stored in the superseded token doc,
encrypted with a key that is derived from the superseded refresh
token. As such, subsequent requests to refresh the same token in
the predefined time window will return the same superseding access
token and refresh token values, without hitting the tokens index
(as this only stores hashes of the token values). AES in GCM
mode is used for encrypting the token values and the key
derivation from the superseded refresh token uses a small number
of iterations as it needs to be quick.
For backwards compatibility reasons, the new behavior is only
enabled when all nodes in a cluster are in the required version
so that old nodes can cope with the token values in a mixed
cluster during a rolling upgrade.
This commit updates the default ciphers and TLS protocols that are used
when the runtime JDK supports them. New cipher support has been
introduced in JDK 11 and 12 along with performance fixes for AES GCM.
The ciphers are ordered with PFS ciphers being most preferred, then
AEAD ciphers, and finally those with mainstream hardware support. When
available stronger encryption is preferred for a given cipher.
This is a backport of #41385 and #41808. There are known JDK bugs with
TLSv1.3 that have been fixed in various versions. These are:
1. The JDK's bundled HttpsServer will endless loop under JDK11 and JDK
12.0 (Fixed in 12.0.1) based on the way the Apache HttpClient performs
a close (half close).
2. In all versions of JDK 11 and 12, the HttpsServer will endless loop
when certificates are not trusted or another handshake error occurs. An
email has been sent to the openjdk security-dev list and #38646 is open
to track this.
3. In JDK 11.0.2 and prior there is a race condition with session
resumption that leads to handshake errors when multiple concurrent
handshakes are going on between the same client and server. This bug
does not appear when client authentication is in use. This is
JDK-8213202, which was fixed in 11.0.3 and 12.0.
4. In JDK 11.0.2 and prior there is a bug where resumed TLS sessions do
not retain peer certificate information. This is JDK-8212885.
The way these issues are addressed is that the current java version is
checked and used to determine the supported protocols for tests that
provoke these issues.
The migrate tool was added when the native realm was created, to aid
users in converting from file realms that were per node, into the
cluster managed native realm. While this tool was useful at the time,
users should now be using the native realm directly. This commit
deprecates the tool, to be removed in a followup for 8.0.
If a basic license enables security, then we should also enforce TLS
on the transport interface.
This was already the case for Standard/Gold/Platinum licenses.
For Basic, security defaults to disabled, so some of the process
around checking whether security is actuallY enabled is more complex
now that we need to account for basic licenses.
Because realms are configured at node startup, but license levels can
change dynamically, it is possible to have a running node that has a
particular realm type configured, but that realm is not permitted under
the current license.
In this case the realm is silently ignored during authentication.
This commit adds a warning in the elasticsearch logs if authentication
fails, and there are realms that have been skipped due to licensing.
This message is not intended to imply that the realms could (or would)
have successfully authenticated the user, but they may help reduce
confusion about why authentication failed if the caller was expecting
the authentication to be handled by a particular realm that is in fact
unlicensed.
Backport of: #41778
The run task is supposed to run elasticsearch with the given plugin or
module. However, for modules, this is most realistic if using the full
distribution. This commit changes the run setup to use the default or
oss as appropriate.
This is related to #27260. Currently we have a single read buffer that
is no larger than a single TLS packet. This prevents us from reading
multiple TLS packets in a single socket read call. This commit modifies
our TLS work to support reading similar to the plaintext case. The data
will be copied to a (potentially) recycled TLS packet-sized buffer for
interaction with the SSLEngine.
This commit is a refactoring of how we filter addresses on
interfaces. In particular, we refactor all of these methods into a
common private method. We also change the order of logic to first check
if an address matches our filter and then check if the interface is
up. This is to possibly avoid problems we are seeing where devices are
flapping up and down while we are checking for loopback addresses. We do
not expect the loopback device to flap up and down so by reversing the
logic here we avoid that problem on CI machines. Finally, we expand the
error message when this does occur so that we know which device is
flapping.
This is related to #27260. Currently there is a setting
http.read_timeout that allows users to define a read timeout for the
http transport. This commit implements support for this functionality
with the transport-nio plugin. The behavior here is that a repeating
task will be scheduled for the interval defined. If there have been
no requests received since the last run and there are no inflight
requests, the channel will be closed.
This is modelled on the qa test for TLS on basic.
It starts a cluster on basic with security & performs a number of
security related checks.
It also performs those same checks on a trial license.
This adds support for using security on a basic license.
It includes:
- AllowedRealmType.NATIVE realms (reserved, native, file)
- Roles / RBAC
- TLS (already supported)
It does not support:
- Audit
- IP filters
- Token Service & API Keys
- Advanced realms (AD, LDAP, SAML, etc)
- Advanced roles (DLS, FLS)
- Pluggable security
As with trial licences, security is disabled by default.
This commit does not include any new automated tests, but existing tests have been updated.
This commit introduces the `.security-tokens` and `.security-tokens-7`
alias-index pair. Because index snapshotting is at the index level granularity
(ie you cannot snapshot a subset of an index) snapshoting .`security` had
the undesirable effect of storing ephemeral security tokens. The changes
herein address this issue by moving tokens "seamlessly" (without user
intervention) to another index, so that a "Security Backup" (ie snapshot of
`.security`) would not be bloated by ephemeral data.
Today we allow adding entries from a file or from a string, yet we
internally maintain this distinction such that if you try to add a value
from a file for a setting that expects a string or add a value from a
string for a setting that expects a file, you will have a bad time. This
causes a pain for operators such that for each setting they need to know
this difference. Yet, we do not need to maintain this distinction
internally as they are bytes after all. This commit removes that
distinction and includes logic to upgrade legacy keystores.
This is related to #27260. Currently for the SSLDriver we allocate a
dedicated network write buffer and encrypt the data into that buffer one
buffer at a time. This requires constantly switching between encrypting
and flushing. This commit adds a dedicated outbound buffer for SSL
operations that will internally allocate new packet sized buffers as
they are need (for writing encrypted data). This allows us to totally
encrypt an operation before writing it to the network. Eventually it can
be hooked up to buffer recycling.
This commit also backports the following commit:
Handle WRAP ops during SSL read
It is possible that a WRAP operation can occur while decrypting
handshake data in TLS 1.3. The SSLDriver does not currently handle this
well as it does not have access to the outbound buffer during read call.
This commit moves the buffer into the Driver to fix this issue. Data
wrapped during a read call will be queued for writing after the read
call is complete.
This is related to #27260. Currently for the SSLDriver we allocate a
dedicated network write buffer and encrypt the data into that buffer one
buffer at a time. This requires constantly switching between encrypting
and flushing. This commit adds a dedicated outbound buffer for SSL
operations that will internally allocate new packet sized buffers as
they are need (for writing encrypted data). This allows us to totally
encrypt an operation before writing it to the network. Eventually it can
be hooked up to buffer recycling.
TLS 1.3 changes to the SSLEngine introduced a scenario where a UNWRAP
call during a handshake can consume a close notify alerty without
throwing an exception. This means that we continue down a codepath where
we assert that we are still in handshaking mode. Transitioning to closed
from handshaking is a valid scenario. This commit removes this
assertion.
Today the `_field_caps` API returns the list of indices where a field
is present only if this field has different types within the requested indices.
However if the request is an index pattern (or an alias, or both...) there
is no way to infer the indices if the response contains only fields that have
the same type in all indices. This commit changes the response to always return
the list of indices in the response. It also adds a way to retrieve unmapped field
in a specific section per field called `unmapped`. This section is created for each field
that is present in some indices but not all if the parameter `include_unmapped` is set to
true in the request (defaults to false).
The Has Privileges API allows to tap into the authorization process, to validate
privileges without actually running the operations to be authorized. This commit
fixes a bug, in which the Has Privilege API returned spurious results when checking
for index privileges over restricted indices (currently .security, .security-6,
.security-7). The actual authorization process is not affected by the bug.
hamcrest has some improvements in newer versions, like FileMatchers
that make assertions regarding file exists cleaner. This commit upgrades
to the latest version of hamcrest so we can start using new and improved
matchers.
The `DistinguishedNamePredicate`, used for matching users to role mapping
expressions, should handle users with null DNs. But it fails to do so (and this is
a NPE bug), if the role mapping expression contains a lucene regexp or a wildcard.
The fix simplifies `DistinguishedNamePredicate` to not handle null DNs at all, and
instead use the `ExpressionModel#NULL_PREDICATE` for the DN field, just like
any other missing user field.
When the same alias points to multiple indices we can write to only one index
with `is_write_index` value `true`. The special handling in case of the put
mapping request(to resolve authorized indices) has a check on indices size
for a concrete index. If multiple indices existed then it marked the request
as unauthorized.
The check has been modified to consider write index flag and only when the
requested index matches with the one with write index alias, the alias is considered
for authorization.
Closes#40831
This commit adds an OpenID Connect authentication realm to
elasticsearch. Elasticsearch (with the assistance of kibana or
another web component) acts as an OpenID Connect Relying
Party and supports the Authorization Code Grant and Implicit
flows as described in http://ela.st/oidc-spec. It adds support
for consuming and verifying signed ID Tokens, both RP
initiated and 3rd party initiated Single Sign on and RP
initiated signle logout.
It also adds an OpenID Connect Provider in the idp-fixture to
be used for the associated integration tests.
This is a backport of #40674
For pattern "n:localhost" PatternRule#isLocalhost() matches
any local address, loopback address.
[Note: I think for "localhost" this should not consider IP address
as a match when they are bound to network interfaces. It should just
be loopback address check unless the intent is to match all local addresses.
This class is adopted from Netty3 and I am not sure if this is intended
behavior or maybe I am missing something]
For now I have fixed this assuming the PatternRule#isLocalhost check is
correct by avoiding use of local address to check address denied.
Closes#40194
* moved hlrc parsing tests from xpack to hlrc module and removed dependency on hlrc from xpack core
* deprecated old base test class
* added deprecated jdoc tag
* split test between xpack-core part and hlrc part
* added lang-mustache test dependency, this previously came in via
hlrc dependency.
* added hlrc dependency on a qa module
* duplicated ClusterPrivilegeName class in xpack-core, since x-pack
core no longer has a dependency on hlrc.
* replace ClusterPrivilegeName usages with string literals
* moved tests to dedicated to hlrc packages in order to remove Hlrc part from the name and make sure to use imports instead of full qualified class where possible
* remove ESTestCase. from method invocation and use method directly,
because these tests indirectly extend from ESTestCase
This PR generates deprecation log entries for each Role Descriptor,
used for building a Role, when the Role Descriptor grants more privileges
for an alias compared to an index that the alias points to. This is done in
preparation for the removal of the ability to define privileges over aliases.
There is one log entry for each "role descriptor name"-"alias name" pair.
On such a notice, the administrator is expected to modify the Role Descriptor
definition so that the name pattern for index names does not cover aliases.
Caveats:
* Role Descriptors that are not used in any authorization process,
either because they are not mapped to any user or the user they are mapped to
is not used by clients, are not be checked.
* Role Descriptors are merged when building the effective Role that is used in
the authorization process. Therefore some Role Descriptors can overlap others,
so even if one matches aliases in a deprecated way, and it is reported as such,
it is not at risk from the breaking behavior in the current role mapping configuration
and index-alias configuration. It is still reported because it is a best practice to
change its definition, or remove offending aliases.
* Replace usages RandomizedTestingTask with built-in Gradle Test (#40978)
This commit replaces the existing RandomizedTestingTask and supporting code with Gradle's built-in JUnit support via the Test task type. Additionally, the previous workaround to disable all tasks named "test" and create new unit testing tasks named "unitTest" has been removed such that the "test" task now runs unit tests as per the normal Gradle Java plugin conventions.
(cherry picked from commit 323f312bbc829a63056a79ebe45adced5099f6e6)
* Fix forking JVM runner
* Don't bump shadow plugin version
This opt-out query cache has an unsafe publication issue, where the
cache is exposed to another thread (namely the cluster state update
thread) before the constructor has finished execution. This exposes the
opt-out query cache to concurrency bugs. This commit addresses this by
ensuring that the opt-out query cache is not registered as a listener
for license state changes until after the constructor has returned.
* Avoid sharing source directories as it breaks intellij
* Subprojects share main project output classes directory
* Fix jar hell
* Fix sql security with ssl integ tests
* Relax dependency ordering rule so we don't explode on cycles
This adds a new security/qa test for TLS on a basic license.
It starts a 2 node cluster with a basic license, and TLS enabled
on both HTTP and Transport, and verifies the license type, x-pack
SSL usage and SSL certificates API.
It also upgrades the cluster to a trial license and performs that
same set of checks (to ensure that clusters with basic license
and TLS enabled can be upgraded to a higher feature license)
Backport of: #40714
This change updates our version of httpclient to version 4.5.8, which
contains the fix for HTTPCLIENT-1968, which is a bug where the client
started re-writing paths that contained encoded reserved characters
with their unreserved form.
Many gradle projects specifically use the -try exclude flag, because
there are many cases where auto-closeable resource ignore is never
referenced in body of corresponding try statement. Suppressing this
warning specifically in each case that it happens using
`@SuppressWarnings("try")` would be very verbose.
This change removes `-try` from any gradle project and adds it to the
build plugin. Also this change removes exclude flags from gradle projects
that is already specified in build plugin (for example -deprecation).
Relates to #40366
It is possible to have SSL enabled but security disabled if security
was dynamically disabled by the license type (e.g. trial license).
e.g. In the following configuration:
xpack.license.self_generated.type: trial
# xpack.security not set, default to disabled on trial
xpack.security.transport.ssl.enabled: true
The security feature will be reported as
available: true
enabled: false
And in this case, SSL will be active even though security is not
enabled.
This commit causes the X-Pack feature usage to report the state of the
"ssl" features unless security was explicitly disabled in the
settings.
Backport of: #40672
This adds a new `role_templates` field to role mappings that is an
alternative to the existing roles field.
These templates are evaluated at runtime to determine which roles should be
granted to a user.
For example, it is possible to specify:
"role_templates": [
{ "template":{ "source": "_user_{{username}}" } }
]
which would mean that every user is assigned to their own role based on
their username.
You may not specify both roles and role_templates in the same role
mapping.
This commit adds support for templates to the role mapping API, the role
mapping engine, the Java high level rest client, and Elasticsearch
documentation.
Due to the lack of caching in our role mapping store, it is currently
inefficient to use a large number of templated role mappings. This will be
addressed in a future change.
Backport of: #39984, #40504
This commit introduces 2 changes to application privileges:
- The validation rules now accept a wildcard in the "suffix" of an application name.
Wildcards were always accepted in the application name, but the "valid filename" check
for the suffix incorrectly prevented the use of wildcards there.
- A role may now be defined against a wildcard application (e.g. kibana-*) and this will
be correctly treated as granting the named privileges against all named applications.
This does not allow wildcard application names in the body of a "has-privileges" check, but the
"has-privileges" check can test concrete application names against roles with wildcards.
Backport of: #40398
Replicated closed indices can't be indexed into or searched, and therefore don't need a shard with
full indexing and search capabilities allocated. We can save on a lot of heap memory for those
indices by not allocating a mapper service and caching infrastructure (which preallocates a constant
amount per instance). Before this change, a 1GB ES instance could host 250 replicated closed
metricbeat indices (each index with one shard). After this change, the same instance can host 7300
replicated closed metricbeat instances (not that this would be a recommended configuration). Most
of the remaining memory is in the cluster state and the IndexSettings object.
This refactoring is in the context of the work related to moving security
tokens to a new index. In that regard, the Token Service has to work with
token documents stored in any of the two indices, albeit only as a transient
situation. I reckoned the added complexity as unmanageable,
hence this refactoring.
This is incomplete, as it fails to address the goal of minimizing .security accesses,
but I have stopped because otherwise it would've become a full blown rewrite
(if not already). I will follow-up with more targeted PRs.
In addition to being a true refactoring, some 400 errors moved to 500. Furthermore,
more stringed validation of various return result, has been implemented, notably the
one of the token document creation.
When creating API keys we check for if API key with
the same key name already exists and fail the request if it does.
The check should have been performed with XPackSecurityUser
instead of the authenticated user. This caused the request to fail
in case of the non-super user trying to create an API key.
This commit fixes by executing search action with SECURITY_ORIGIN
so it can be executed with XPackSecurityUser.
Also fixed the Rest test to avoid using a user with `super_user` role.
Closes#40029
The setup-passwords tool gives cryptic messages in case where custom discovery providers are
used (see #33580). As the URL auto-detection logic should be seen as best effort, this commit
improves the exception message to make it clearer what needs to be done to fix the issue.
Relates #33580
`SecurityIndexManager` is hardcoded to handle only the `.security`-`.security-7` alias-index pair.
This commit removes the hardcoded bits, so that the `SecurityIndexManager` can be reused
for other indices, such as the planned security tokens index (`.security-tokens-7`).
The LDAP tests attempt to bind all interfaces,
but if for some reason an interface can't be bound
the tests will stall until the suite times out.
This modifies the tests to be a bit more lenient and allow
some binding to fail so long as at least one succeeds.
This allows the test to continue even in more antagonistic
environments.
A TLS handshake requires exchanging multiple messages to initiate a
session. If one side decides to close during the handshake, it is
supposed to send a close_notify alert (similar to closing during
application data exchange). The java SSLEngine engine throws an
exception when this happens. We currently log this at the warn level if
trace logging is not enabled. This level is too high for a valid
scenario. Additionally it happens all the time in tests (quickly closing
and opened transports). This commit changes this to be logged at the
debug level if trace is not enabled. Additionally, it extracts the
transport security exception handling to a common class.
Previously all the threads were writing the received tokens to a
HashSet. In cases with many threads, sometimes (1 every ~25 tests)
calling size() on the HashSet returned 2 even though it seemed to
contain only one String and there was no evidence from logging that
threadSecurityClient.refreshToken() ever returned a different
access or refresh token.
This commit changes the test to use a ConcurrentHashMap instead,
checking that we only received one pair of access token/refresh token
eventually. It also adds a check so that we won't take into consideration
tokens that are returned after 30s, hence not in the concurrent refresh
time window.
Fixes several errors of the token retry logic:
* not checking for backoff.hasNext() before calling backoff.next()
* checking for backoff.hasNext() without calling backoff.next()
* not preserving the context on the retry
* calling scheduleWithFixedDelay instead of schedule
Today the `GroupedActionListener` accepts a `defaults` parameter but all
callers pass an empty list. Also it is permitted to pass an empty group but
this is trappy because the delegated listener is never be called in that case.
This commit removes the `defaults` parameter and forbids an empty group.
As we are moving to single type indices,
we need to address this change in security-related indexes.
To address this, we are
- updating index templates to use preferred type name `_doc`
- updating the API calls to use preferred type name `_doc`
Upgrade impact:-
In case of an upgrade from 6.x, the security index has type
`doc` and this will keep working as there is a single type and `_doc`
works as an alias to an existing type. The change is handled in the
`SecurityIndexManager` when we load mappings and settings from
the template. Previously, we used to do a `PutIndexTemplateRequest`
with the mapping source JSON with the type name. This has been
modified to remove the type name from the source.
So in the case of an upgrade, the `doc` type is updated
whereas for fresh installs `_doc` is updated. This happens as
backend handles `_doc` as an alias to the existing type name.
An optional step is to `reindex` security index and update the
type to `_doc`.
Since we do not support the security audit log index,
that template has been deleted.
Relates: #38637
This is a backport of #39631
Co-authored-by: Jay Modi jaymode@users.noreply.github.com
This change adds support for the concurrent refresh of access
tokens as described in #36872
In short it allows subsequent client requests to refresh the same token that
come within a predefined window of 60 seconds to be handled as duplicates
of the original one and thus receive the same response with the same newly
issued access token and refresh token.
In order to support that, two new fields are added in the token document. One
contains the instant (in epoqueMillis) when a given refresh token is refreshed
and one that contains a pointer to the token document that stores the new
refresh token and access token that was created by the original refresh.
A side effect of this change, that was however also a intended enhancement
for the token service, is that we needed to stop encrypting the string
representation of the UserToken while serializing. ( It was necessary as we
correctly used a new IV for every time we encrypted a token in serialization, so
subsequent serializations of the same exact UserToken would produce
different access token strings)
This change also handles the serialization/deserialization BWC logic:
In mixed clusters we keep creating tokens in the old format and
consume only old format tokens
In upgraded clusters, we start creating tokens in the new format but
still remain able to consume old format tokens (that could have been
created during the rolling upgrade and are still valid)
When reading/writing TokensInvalidationResult objects, we take into
consideration that pre 7.1.0 these contained an integer field that carried
the attempt count
Resolves#36872
Previously, the security index could be wrongfully recreated. This might
happen if the index was interpreted as missing, as in the case of a fresh
install, but the index existed and the state did not yet recover.
This fix will return HTTP SERVICE_UNAVAILABLE (503) for requests that
try to write to the security index before the state has not been recovered yet.
This is a backport of #38382
This change adds supports for the concurrent refresh of access
tokens as described in #36872
In short it allows subsequent client requests to refresh the same token that
come within a predefined window of 60 seconds to be handled as duplicates
of the original one and thus receive the same response with the same newly
issued access token and refresh token.
In order to support that, two new fields are added in the token document. One
contains the instant (in epoqueMillis) when a given refresh token is refreshed
and one that contains a pointer to the token document that stores the new
refresh token and access token that was created by the original refresh.
A side effect of this change, that was however also a intended enhancement
for the token service, is that we needed to stop encrypting the string
representation of the UserToken while serializing. ( It was necessary as we
correctly used a new IV for every time we encrypted a token in serialization, so
subsequent serializations of the same exact UserToken would produce
different access token strings)
This change also handles the serialization/deserialization BWC logic:
- In mixed clusters we keep creating tokens in the old format and
consume only old format tokens
- In upgraded clusters, we start creating tokens in the new format but
still remain able to consume old format tokens (that could have been
created during the rolling upgrade and are still valid)
Resolves#36872
Co-authored-by: Jay Modi jaymode@users.noreply.github.com
This commit adds a simple integ test that exercises the flow:
* snapshot .security
* delete .security
* restore .security
, checking that the Native Realm works as expected.
Relates #34454
Currently there are two security tests that specifically target the
netty security transport. This PR moves the client authentication tests
into `AbstractSimpleSecurityTransportTestCase` so that the nio transport
will also be tested.
Additionally the work to build transport configurations is moved out of
the netty transport and tested independently.
This changes the name of the internal security index to ".security-7",
but supports indices that were upgraded from earlier versions and use
the ".security-6" name.
In all cases, both ".security-6" and ".security-7" are considered to
be restricted index names regardless of which name is actually in use
on the cluster.
Backport of: #39337
This change is a backport of #39252
- Fixes TokenBackwardsCompatibilityIT: Existing tests seemed to made
the assumption that in the oneThirdUpgraded stage the master node
will be on the old version and in the twoThirdsUpgraded stage, the
master node will be one of the upgraded ones. However, there is no
guarantee that the master node in any of the states will or will
not be one of the upgraded ones.
This class now tests:
- That we can generate and consume tokens before we start the
rolling upgrade.
- That we can consume tokens generated in the old cluster during
all the stages of the rolling upgrade.
- That while on a mixed cluster, when/if the master node is
upgraded, we can generate, consume and refresh a token
- That after the rolling upgrade, we can consume a token
generated in an old cluster and can invalidate it so that it
can't be used any more.
- Ensures that during the rolling upgrade, the upgraded nodes have
the same configuration as the old nodes. Specifically that the
file realm we use is explicitly named `file1`. This is needed
because while attempting to refresh a token in a mixed cluster
we might create a token hitting an old node and attempt to refresh
it hitting a new node. If the file realm name is not the same, the
refresh will be seen as being made by a "different" client, and
will, thus, fail.
- Renames the Authentication variable we check while refreshing a
token to be clientAuth in order to make the code more readable.
Some of the above were possibly causing the flakiness of #37379
Currently remote compression and ping schedule settings are dynamic.
However, we do not listen for changes. This commit adds listeners for
changes to those two settings. Additionally, when those settings change
we now close existing connections and open new ones with the settings
applied.
Fixes#37201.
This change aims to fix failures in the session factory load balancing
tests that mock failure scenarios. For these tests, we randomly shut
down ldap servers and bind a client socket to the port they were
listening on. Unfortunately, we would occasionally encounter failures
in these tests where a socket was already in use and/or the port
we expected to connect to was wrong and in fact was to one of the ldap
instances that should have been shut down.
The failures are caused by the behavior of certain operating systems
when it comes to binding ports and wildcard addresses. It is possible
for a separate application to be bound to a wildcard address and still
allow our code to bind to that port on a specific address. So when we
close the server socket and open the client socket, we are still able
to establish a connection since the other application is already
listening on that port on a wildcard address. Another variant is that
the os will allow a wildcard bind of a server socket when there is
already an application listening on that port for a specific address.
In order to do our best to prevent failures in these scenarios, this
change does the following:
1. Binds a client socket to all addresses in an awaitBusy
2. Adds assumption that we could bind all valid addresses
3. In the case that we still establish a connection to an address that
we should not be able to, try to bind and expect a failure of not
being connected
Closes#32190
In most of the places we avoid creating the `.security` index (or updating the mapping)
for read/search operations. This is more of a nit for the case of the getRole call,
that fixes a possible mapping update during a get role, and removes a dead if branch
about creating the `.security` index.
This commit attempts to remove the retention leases on the leader shards
when unfollowing an index. This is best effort, since the leader might
not be available.
Yogesh Gaikwad
Albert Zaharovits
Ioannis Kakavas
William Brafford
David Roberts
Armin Braun
Alpar Torok
Yannick Welsch
Tim Brooks
Tim Vernum
Mark Vieira
Ryan Ernst
Jason Tedor
Jay Modi
Christoph Büscher
Jim Ferenczi
Ryan Ernst
Nhat Nguyen
Przemyslaw Gomulka
Alan Woodward
Tal Levy
David Kyle
Ioannis Kakavas
Albert Zaharovits
Jim Ferenczi
Martijn van Groningen
Michael Basnight
David Turner
Tanguy Leroux