[
https://issues.apache.org/jira/browse/FLINK-41435?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
LuChengxiang updated FLINK-41435:
---------------------------------
Description:
h2. Summary
Flink periodically obtains HDFS delegation tokens and distributes them through
HadoopDelegationTokenReceiver. The receiver currently adds new credentials to
UserGroupInformation.getCurrentUser() on the thread that performs the refresh.
Long-lived Hadoop clients, especially checkpoint storage, can capture a
UGI/Subject when their DFSClient is created and use that captured identity for
later RPC connections.
If the refresh path resolves to a different current UGI from the one captured
by a long-lived client, the client can continue using its startup token even
though a new token was obtained successfully. The failure appears at the 24h
HDFS token deadline as SecretManager$InvalidToken.
h2. Observed failure
This was observed in a Flink 1.20.5 Kubernetes application-mode job using
ZooKeeper-based high availability and HDFS-backed checkpoint storage. The
primary failure was in the checkpoint storage path:
{noformat}
CheckpointCoordinator.initializeCheckpointLocation
-> FsCheckpointStorageAccess.initializeLocationForCheckpoint
-> HadoopFileSystem.mkdirs
-> DFSClient.mkdirs
{noformat}
The checkpoint storage client reported InvalidToken for the startup token after
the refresh had created a new HDFS token and the receiver reported success. The
same JVM also showed a FileSystemJobResultStore stack using the expired token
during cleanup. The job-result store is a secondary long-lived client with the
same failure mode, not the primary trigger.
Not every Kubernetes application-mode job is affected. The failure is
timing-dependent: it requires a long-lived Hadoop client to have captured one
Subject/UGI lifecycle while the refresh path later resolves to a different one.
Installation idempotency is out of scope. A diagnostic build that skipped the
second security installation still failed at the 24h HDFS token deadline.
h2. Root cause
HadoopDelegationTokenReceiver targets UserGroupInformation.getCurrentUser() of
the thread executing the refresh. That current UGI is resolved from the
thread's effective AccessControlContext and is not guaranteed to be the UGI
held by HadoopSecurityContext or the UGI captured by a long-lived
checkpoint-storage or job-result-store DFSClient.
The controlled Kubernetes probe observed this path:
{noformat}
provider thread = default-dispatcher-12
freshUGI Subject = 817147333
login/security context Subject = 1333938290
current Subject before doAs = 1333938290
current Subject inside doAs = 817147333
current Subject after doAs = 1333938290
FJP workers created during the provider's blocked RPC:
default-dispatcher-13 / 14 / 15, creator=default-dispatcher-12
later cluster-io worker:
cluster-io-thread-3, creator=default-dispatcher-15, subjectId=817147333
{noformat}
The fresh Subject is not created by HadoopSecurityContext or by the dispatcher.
It belongs to the UGI returned by
{{KerberosLoginProvider.doLoginAndReturnUGI()}}, which calls
{{UserGroupInformation.loginUserFromKeytabAndReturnUGI(principal, keytab)}}.
That call performs a Kerberos login and returns a newly constructed UGI for the
same principal; it does not replace the process login user.
During {{freshUGI.doAs(...)}}, {{fs.addDelegationTokens(...)}} performed an
HDFS delegation-token RPC and blocked in {{CompletableFuture.get()}}. Because
the caller was a ForkJoinWorkerThread, ForkJoinPool entered its managed-block
compensation path and created {{default-dispatcher-13}},
{{default-dispatcher-14}}, and {{default-dispatcher-15}} while the fresh
Subject was active. The Thread constructors copied the creator's effective
AccessControlContext, including the fresh Subject. After {{doAs}} returned,
{{getCurrentUser()}} on the creator resolved to the login Subject again, but
the newly created workers retained the inherited fresh Subject.
Later, {{default-dispatcher-15}} created {{cluster-io-thread-3}}; the probe
recorded {{subjectId=817147333}} for that worker while {{getLoginUser()}} and
{{HadoopSecurityContext}} still referred to the login Subject {{1333938290}}. A
receiver task scheduled on such a worker can therefore add credentials to the
fresh Subject rather than the UGI captured by the long-lived client. This is a
per-thread inherited security context, not a global mutation of
{{UserGroupInformation}} or the process login user.
The provider stack identifies this invocation as {{ResourceManager.onStart}} ->
{{DefaultDelegationTokenManager.start()}} -> {{startTokensUpdate()}}, i.e. the
initial token acquisition. The later scheduled refresh is submitted through the
cluster-io executor. The startup FJP expansion is observed; the same mechanism
is possible for any provider {{doAs}} on a ForkJoinWorkerThread with a blocking
HDFS RPC, but that later path has not been directly reproduced in this probe.
h2. YARN status and scope
In a controlled YARN application-mode probe, the provider also ran on a Pekko
default-dispatcher worker. The provider window showed {{poolSize=3}} /
{{active=2}} before and after, with no {{tryCompensate}} or {{createWorker}};
the current Subject after the provider returned was the login Subject, while
the provider's internal {{freshUGI}} Subject was a different value. YARN did
show ForkJoinPool expansion during the separate JobGraph-generation startup
path, so this is not evidence that YARN is immune.
This is one YARN sample; the provider's {{CompletableFuture}} state was not
instrumented, so the YARN trigger is not established. YARN is not listed as a
confirmed reproduction, but it is not claimed to be exempt.
h2. Proposed change
This issue addresses the case where a token is obtained successfully but the
refreshed credentials do not reach long-lived Hadoop clients, primarily
checkpoint storage.
* Add an accessor for the UGI held by HadoopSecurityContext.
* In HadoopDelegationTokenReceiver, add refreshed credentials to the
security-context UGI, the current UGI, and the login UGI. Deduplicate the
target UGI references by object identity; this prevents duplicate
{{addCredentials}} calls for the same wrapper, but does not merge distinct UGI
wrappers that reference the same Subject. The security-context UGI is the
primary target; the other two preserve non-secured deployments and previous
behavior.
* Add tests covering (a) a refresh thread whose current user differs from the
login user, and (b) a security-context UGI distinct from the current/login
user; verify that all applicable targets receive the credentials.
Out of scope for this issue:
* Preventing HDFS client initialization or blocking RPCs from occurring inside
{{freshUGI.doAs(...)}} or the provider's ForkJoinPool worker.
* Closing or rebinding FileSystem instances created under a transient fresh
Subject; this is a separate client-lifecycle cleanup.
* SecurityUtils.install idempotency.
The proposed receiver change does not rely solely on the refresh thread's
inherited Subject. It does not by itself guarantee that a client already
created under a different Subject is rebound. That remains a separate lifecycle
concern.
h2. Testing
* A controlled Kubernetes application-mode probe recorded the chain from
{{fs.addDelegationTokens()}} through Hadoop ipc {{Client.getRpcResponse()}},
{{CompletableFuture.get()}}, {{ForkJoinPool.managedBlock()}},
{{tryCompensate()}}, and Pekko {{ThreadFactory.newThread()}}. The creator and
the new dispatcher workers had the same fresh Subject identity.
* The same probe recorded {{cluster-io-thread-3}} created by
{{default-dispatcher-15}} with {{subjectId=817147333}}, while the login and
security-context Subjects remained {{1333938290}}.
* Trigger multiple post-startup token refreshes in a long-running job and
verify which executor runs the provider, whether new ForkJoinPool workers are
created, and which Subject the receiver's current user resolves to.
* The controlled YARN application-mode probe showed no {{tryCompensate}} or
{{createWorker}} in the provider window. The separate JobGraph-generation
startup path did create FJP workers, confirming that YARN can enter the same
FJP mechanism under different startup work; the provider's
{{CompletableFuture}} blocking state remains unverified.
* A deployed application-mode job should survive the 24h token deadline and
continue checkpointing without SecretManager$InvalidToken. Verify that the
checkpoint-storage DFSClient receives the new token, and verify the
job-result-store client separately because it can capture a different UGI.
h2. Acceptance criteria
* A token refresh updates the UGI used by checkpoint storage and other
long-lived Hadoop clients.
* Refresh delivers credentials to the security-context UGI regardless of which
Subject a pooled dispatcher or cluster-io worker inherited.
* A long-running application-mode job survives the 24h HDFS token deadline.
was:
h2. Background
Flink periodically obtains HDFS delegation tokens and distributes them through
HadoopDelegationTokenReceiver. The receiver currently adds new credentials to
UserGroupInformation.getCurrentUser() on the thread that performs the refresh.
Long-lived Hadoop clients, especially checkpoint storage and
FileSystemJobResultStore, capture a UGI/Subject when their DFSClient is created
and use that captured identity for later RPC connections.
If the refresh path resolves to a different current UGI from the one captured
by a long-lived client, the client can continue using the startup token even
though a new token was obtained successfully. The failure appears at the 24h
HDFS token deadline as SecretManager$InvalidToken.
h2. Observed failure
This was observed in a Kubernetes application-mode job using ZooKeeper-based
high availability and HDFS-backed checkpoint storage. The primary failure was
in the checkpoint storage path:
{noformat}
CheckpointCoordinator.initializeCheckpointLocation
-> FsCheckpointStorageAccess.initializeLocationForCheckpoint
-> HadoopFileSystem.mkdirs
-> DFSClient.mkdirs
{noformat}
The checkpoint storage client reported InvalidToken for the startup token after
the refresh had created a new HDFS token and the receiver reported success. The
same JVM also showed a FileSystemJobResultStore stack using the expired token
during cleanup. The job-result store is therefore a secondary long-lived client
with the same failure mode, not the primary trigger.
Not every Kubernetes application-mode job is affected. The failure is
timing-dependent: it requires a long-lived Hadoop client to have captured one
Subject/UGI lifecycle while the refresh path later resolves to a different one.
The double security installation is not the root cause: a diagnostic build that
skipped the second security installation still failed at the 24h HDFS token
deadline. This issue therefore focuses on the receiver's `current` UGI
behavior; installation idempotency remains out of scope.
h2. New finding: how a fresh Subject reaches a pooled Pekko worker
A controlled Kubernetes application-mode probe instrumented the token provider,
the HDFS RPC path, and Pekko's dispatcher thread factory. The observed chain
was:
{noformat}
HadoopFSDelegationTokenProvider.obtainDelegationTokens()
-> freshUGI.doAs(...)
-> fs.addDelegationTokens(...)
-> HDFS delegation-token RPC
-> Client.getRpcResponse()
-> CompletableFuture.get()
-> ForkJoinPool.managedBlock()
-> tryCompensate()
-> createWorker()
-> Pekko ThreadFactory.newThread()
{noformat}
This happened while a Pekko default-dispatcher worker was executing
freshUGI.doAs(...). The HDFS RPC was waiting on a CompletableFuture. Because
the caller was a ForkJoinWorkerThread, JDK ForkJoinPool entered its
managed-block compensation path and created additional dispatcher workers while
the fresh Subject was active.
h3. Origin of the fresh Subject
The fresh Subject is not created by HadoopSecurityContext or by the dispatcher.
It belongs to the UGI returned by
{{KerberosLoginProvider.doLoginAndReturnUGI()}}, which calls
{{UserGroupInformation.loginUserFromKeytabAndReturnUGI(principal, keytab)}}
when a principal is configured. That call performs a Kerberos login and returns
a newly constructed UGI for the same principal; it does not replace the process
login user. In the probe, the provider observed:
{noformat}
provider thread = default-dispatcher-12
freshUGI Subject = 817147333
login/security context Subject = 1333938290
current Subject before doAs = 1333938290
current Subject inside doAs = 817147333
current Subject after doAs = 1333938290
FJP workers created during the provider's blocked RPC:
default-dispatcher-13 / 14 / 15, creator=default-dispatcher-12
later cluster-io worker:
cluster-io-thread-3, creator=default-dispatcher-15, subjectId=817147333
{noformat}
In this probe the provider was invoked by {{ResourceManager.onStart}} ->
{{DefaultDelegationTokenManager.start()}} -> {{startTokensUpdate()}}, i.e. the
initial token acquisition, not the later scheduled refresh.
While the provider was executing {{freshUGI.doAs(...)}},
{{fs.addDelegationTokens(...)}} performed an HDFS delegation-token RPC and
blocked in {{CompletableFuture.get()}}. Because the caller was a
ForkJoinWorkerThread, ForkJoinPool entered its managed-block compensation path
and created {{default-dispatcher-13}}, {{default-dispatcher-14}}, and
{{default-dispatcher-15}} with {{default-dispatcher-12}} as their creator. The
Thread constructors copied the creator's effective AccessControlContext,
including the fresh Subject. After {{doAs}} returned, {{getCurrentUser()}} on
the creator resolved to the login Subject again, but the newly created workers
retained the inherited fresh Subject.
Later, {{default-dispatcher-15}} created {{cluster-io-thread-3}}. The probe
recorded {{subjectId=817147333}} for that worker, while {{getLoginUser()}} and
{{HadoopSecurityContext}} still referred to the login Subject {{1333938290}}.
This is the concrete propagation path from the provider's {{freshUGI}} to a
worker that later resolves {{UserGroupInformation.getCurrentUser()}} to the
fresh Subject.
This is a per-thread inherited security context, not a global mutation of
{{UserGroupInformation}} or the process login user.
{{UserGroupInformation.getCurrentUser()}} returns the current user including
any {{doAs}} scope on the current stack. {{Subject.doAs}} constructs a
Subject-bound {{AccessControlContext}} and makes it effective through
{{AccessController.doPrivileged}}; during the action,
{{AccessController.getContext()}} observes that inner context. A {{Thread}}
created inside this window stores the context returned by
{{AccessController.getContext()}} in its {{inheritedAccessControlContext}}, so
the pooled worker can keep the inner fresh Subject after {{doAs}} returns.
The probe observed this HDFS RPC path while the provider was running on a
ForkJoinPool worker. The resulting dispatcher workers inherited the fresh
Subject, so a later UserGroupInformation.getCurrentUser() call on a pooled
worker can resolve to that fresh Subject rather than the security-context or
login Subject. This is the `current` UGI used by the receiver.
This is not a startup-only condition: the token provider executes on each
refresh, and a later blocking HDFS RPC on a ForkJoinPool worker while
freshUGI.doAs() is active can create another worker inheriting the fresh
Subject. The startup path is where the expansion was observed; the inherited
Subject persists on the worker and can be used by later refresh or receiver
tasks.
h2. YARN status
In a controlled YARN application-mode probe, refresh-time current and login
Subject identities were equal. The provider did run on a Pekko dispatcher
worker, but all provider invocations reported no newly created Pekko threads.
The normal YARN application-mode startup path tested so far did not reproduce
the fresh-Subject worker creation observed on Kubernetes.
This absence of new workers is not proof that YARN never enters the
managedBlock/createWorker path. A YARN AM-level full-stack capture has not been
completed, so the platform-specific trigger is not yet established. YARN is
therefore not listed as a confirmed reproduction, but it is also not claimed to
be immune. The code path is available on YARN if a token-provider call creates
a new HDFS client and blocks on a ForkJoinWorkerThread; that blocking-client
path still needs to be confirmed there.
h2. Version and code-path notes
The 1.18 path had synchronous FileSystemJobResultStore I/O, so the
worker/Subject interaction was not triggered in the deployments observed. 1.15
did not have the unified delegation-token manager path. These version
differences may explain the observed frequency; they do not prove that those
versions are immune.
FLINK-27204 (1.19+) moved FileSystemJobResultStore I/O onto the shared
cluster-io executor. DefaultDelegationTokenManager uses the same executor for
periodic token refresh. The provider and receiver are invoked on the refresh
path; the security-context UGI, login UGI, and current UGI can diverge.
h2. Root cause
HadoopDelegationTokenReceiver targets UserGroupInformation.getCurrentUser() of
the thread executing the refresh. That current UGI is resolved from the
thread's effective AccessControlContext and is not guaranteed to be the UGI
held by HadoopSecurityContext or the UGI captured by a long-lived
checkpoint-storage or job-result-store DFSClient.
In the reproduced Kubernetes path, the freshUGI.doAs() window caused a new
Pekko dispatcher worker to inherit the fresh Subject. Consequently, the
receiver could execute with a current UGI that differed from the
security-context and login UGIs. The second security installation is not
required for this mismatch, and install idempotency therefore does not address
it.
h2. Proposed change
This issue addresses the case where a token is obtained successfully but the
refreshed credentials do not reach long-lived Hadoop clients, primarily
checkpoint storage.
* Add an accessor for the UGI held by HadoopSecurityContext.
* In HadoopDelegationTokenReceiver, add refreshed credentials to the
security-context UGI, the current UGI, and the login UGI, deduplicated by
object identity. The security-context UGI is the primary target; the other two
preserve non-secured deployments and previous behavior.
* Add tests covering (a) a refresh thread whose current user differs from the
login user, and (b) a security-context UGI distinct from the current/login
user; verify that all applicable targets receive the credentials.
Out of scope for this issue:
* Preventing HDFS client initialization or blocking RPCs from occurring inside
freshUGI.doAs() or the provider's ForkJoinPool worker.
* Closing or rebinding FileSystem instances created under a transient fresh
Subject; this is a separate client-lifecycle cleanup.
* SecurityUtils.install idempotency.
The proposed receiver change does not rely solely on the refresh thread's
inherited Subject. It does not by itself guarantee that a client already
created under a different Subject is rebound. That remains a separate lifecycle
concern.
h2. Testing
* A controlled Kubernetes application-mode probe recorded the full chain from
fs.addDelegationTokens() through Hadoop ipc Client.getRpcResponse(),
CompletableFuture.get(), ForkJoinPool.managedBlock(), tryCompensate(), and
Pekko ThreadFactory.newThread(). The creator and the new dispatcher workers had
the same fresh Subject identity.
* A later receiver invocation on a cluster-io worker observed that inherited
fresh Subject while the login and security-context Subjects remained the
original login Subject.
* Trigger multiple post-startup token refreshes in a long-running job and
verify which executor runs the provider, whether new ForkJoinPool workers are
created, and which Subject the receiver's current user resolves to.
* A JVM probe confirmed that a fixed thread pool created under Subject A but
first submitted under Subject B runs subsequent tasks under Subject B.
* A controlled Kubernetes application-mode probe instrumented
ExecutorThreadFactory. The first cluster-io worker was created by
DefaultLeaderElectionService-leadershipOperationExecutor-thread-1, and the
first jobmanager-io worker by a cluster-io worker; both inherited the same
Kerberos Subject at creation time. This is a supporting inheritance check, not
evidence for the ForkJoinPool compensation path.
* Receiver tests as described above.
* The controlled YARN application-mode probe observed no new Pekko dispatcher
workers in the tested normal startup path. A YARN AM-level full-stack capture
remains to be completed.
* A deployed application-mode job should survive the 24h token deadline and
continue checkpointing without SecretManager$InvalidToken. Verify that the
checkpoint-storage DFSClient receives the new token, and verify the
job-result-store client separately because it can capture a different UGI.
h2. Acceptance criteria
* A token refresh updates the UGI used by checkpoint storage and other
long-lived Hadoop clients.
* Refresh delivers credentials to the security-context UGI regardless of which
Subject a pooled dispatcher or cluster-io worker inherited.
* A long-running application-mode job survives the 24h HDFS token deadline.
> Hadoop delegation token refresh can target a different UGI than Hadoop clients
> ------------------------------------------------------------------------------
>
> Key: FLINK-41435
> URL: https://issues.apache.org/jira/browse/FLINK-41435
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Configuration
> Affects Versions: 2.0.0, 1.19.0, 1.20.0, 2.1.0, 2.2.0, 2.3.0, 2.4.0
> Environment: Flink 1.19.0 and later (1.19.x, 1.20.x, 2.0 through
> 2.4), plus master (2.5-SNAPSHOT); Kubernetes and standalone application mode
> with Kerberos and HDFS delegation tokens. The mismatch is more likely when
> FileSystemJobResultStore and the token manager share the cluster-io executor
> (FLINK-27204).
> Reporter: LuChengxiang
> Priority: Major
>
> h2. Summary
> Flink periodically obtains HDFS delegation tokens and distributes them
> through HadoopDelegationTokenReceiver. The receiver currently adds new
> credentials to UserGroupInformation.getCurrentUser() on the thread that
> performs the refresh. Long-lived Hadoop clients, especially checkpoint
> storage, can capture a UGI/Subject when their DFSClient is created and use
> that captured identity for later RPC connections.
> If the refresh path resolves to a different current UGI from the one captured
> by a long-lived client, the client can continue using its startup token even
> though a new token was obtained successfully. The failure appears at the 24h
> HDFS token deadline as SecretManager$InvalidToken.
> h2. Observed failure
> This was observed in a Flink 1.20.5 Kubernetes application-mode job using
> ZooKeeper-based high availability and HDFS-backed checkpoint storage. The
> primary failure was in the checkpoint storage path:
> {noformat}
> CheckpointCoordinator.initializeCheckpointLocation
> -> FsCheckpointStorageAccess.initializeLocationForCheckpoint
> -> HadoopFileSystem.mkdirs
> -> DFSClient.mkdirs
> {noformat}
> The checkpoint storage client reported InvalidToken for the startup token
> after the refresh had created a new HDFS token and the receiver reported
> success. The same JVM also showed a FileSystemJobResultStore stack using the
> expired token during cleanup. The job-result store is a secondary long-lived
> client with the same failure mode, not the primary trigger.
> Not every Kubernetes application-mode job is affected. The failure is
> timing-dependent: it requires a long-lived Hadoop client to have captured one
> Subject/UGI lifecycle while the refresh path later resolves to a different
> one.
> Installation idempotency is out of scope. A diagnostic build that skipped the
> second security installation still failed at the 24h HDFS token deadline.
> h2. Root cause
> HadoopDelegationTokenReceiver targets UserGroupInformation.getCurrentUser()
> of the thread executing the refresh. That current UGI is resolved from the
> thread's effective AccessControlContext and is not guaranteed to be the UGI
> held by HadoopSecurityContext or the UGI captured by a long-lived
> checkpoint-storage or job-result-store DFSClient.
> The controlled Kubernetes probe observed this path:
> {noformat}
> provider thread = default-dispatcher-12
> freshUGI Subject = 817147333
> login/security context Subject = 1333938290
> current Subject before doAs = 1333938290
> current Subject inside doAs = 817147333
> current Subject after doAs = 1333938290
> FJP workers created during the provider's blocked RPC:
> default-dispatcher-13 / 14 / 15, creator=default-dispatcher-12
> later cluster-io worker:
> cluster-io-thread-3, creator=default-dispatcher-15, subjectId=817147333
> {noformat}
> The fresh Subject is not created by HadoopSecurityContext or by the
> dispatcher. It belongs to the UGI returned by
> {{KerberosLoginProvider.doLoginAndReturnUGI()}}, which calls
> {{UserGroupInformation.loginUserFromKeytabAndReturnUGI(principal, keytab)}}.
> That call performs a Kerberos login and returns a newly constructed UGI for
> the same principal; it does not replace the process login user.
> During {{freshUGI.doAs(...)}}, {{fs.addDelegationTokens(...)}} performed an
> HDFS delegation-token RPC and blocked in {{CompletableFuture.get()}}. Because
> the caller was a ForkJoinWorkerThread, ForkJoinPool entered its managed-block
> compensation path and created {{default-dispatcher-13}},
> {{default-dispatcher-14}}, and {{default-dispatcher-15}} while the fresh
> Subject was active. The Thread constructors copied the creator's effective
> AccessControlContext, including the fresh Subject. After {{doAs}} returned,
> {{getCurrentUser()}} on the creator resolved to the login Subject again, but
> the newly created workers retained the inherited fresh Subject.
> Later, {{default-dispatcher-15}} created {{cluster-io-thread-3}}; the probe
> recorded {{subjectId=817147333}} for that worker while {{getLoginUser()}} and
> {{HadoopSecurityContext}} still referred to the login Subject {{1333938290}}.
> A receiver task scheduled on such a worker can therefore add credentials to
> the fresh Subject rather than the UGI captured by the long-lived client. This
> is a per-thread inherited security context, not a global mutation of
> {{UserGroupInformation}} or the process login user.
> The provider stack identifies this invocation as {{ResourceManager.onStart}}
> -> {{DefaultDelegationTokenManager.start()}} -> {{startTokensUpdate()}}, i.e.
> the initial token acquisition. The later scheduled refresh is submitted
> through the cluster-io executor. The startup FJP expansion is observed; the
> same mechanism is possible for any provider {{doAs}} on a
> ForkJoinWorkerThread with a blocking HDFS RPC, but that later path has not
> been directly reproduced in this probe.
> h2. YARN status and scope
> In a controlled YARN application-mode probe, the provider also ran on a Pekko
> default-dispatcher worker. The provider window showed {{poolSize=3}} /
> {{active=2}} before and after, with no {{tryCompensate}} or {{createWorker}};
> the current Subject after the provider returned was the login Subject, while
> the provider's internal {{freshUGI}} Subject was a different value. YARN did
> show ForkJoinPool expansion during the separate JobGraph-generation startup
> path, so this is not evidence that YARN is immune.
> This is one YARN sample; the provider's {{CompletableFuture}} state was not
> instrumented, so the YARN trigger is not established. YARN is not listed as a
> confirmed reproduction, but it is not claimed to be exempt.
> h2. Proposed change
> This issue addresses the case where a token is obtained successfully but the
> refreshed credentials do not reach long-lived Hadoop clients, primarily
> checkpoint storage.
> * Add an accessor for the UGI held by HadoopSecurityContext.
> * In HadoopDelegationTokenReceiver, add refreshed credentials to the
> security-context UGI, the current UGI, and the login UGI. Deduplicate the
> target UGI references by object identity; this prevents duplicate
> {{addCredentials}} calls for the same wrapper, but does not merge distinct
> UGI wrappers that reference the same Subject. The security-context UGI is the
> primary target; the other two preserve non-secured deployments and previous
> behavior.
> * Add tests covering (a) a refresh thread whose current user differs from the
> login user, and (b) a security-context UGI distinct from the current/login
> user; verify that all applicable targets receive the credentials.
> Out of scope for this issue:
> * Preventing HDFS client initialization or blocking RPCs from occurring
> inside {{freshUGI.doAs(...)}} or the provider's ForkJoinPool worker.
> * Closing or rebinding FileSystem instances created under a transient fresh
> Subject; this is a separate client-lifecycle cleanup.
> * SecurityUtils.install idempotency.
> The proposed receiver change does not rely solely on the refresh thread's
> inherited Subject. It does not by itself guarantee that a client already
> created under a different Subject is rebound. That remains a separate
> lifecycle concern.
> h2. Testing
> * A controlled Kubernetes application-mode probe recorded the chain from
> {{fs.addDelegationTokens()}} through Hadoop ipc {{Client.getRpcResponse()}},
> {{CompletableFuture.get()}}, {{ForkJoinPool.managedBlock()}},
> {{tryCompensate()}}, and Pekko {{ThreadFactory.newThread()}}. The creator and
> the new dispatcher workers had the same fresh Subject identity.
> * The same probe recorded {{cluster-io-thread-3}} created by
> {{default-dispatcher-15}} with {{subjectId=817147333}}, while the login and
> security-context Subjects remained {{1333938290}}.
> * Trigger multiple post-startup token refreshes in a long-running job and
> verify which executor runs the provider, whether new ForkJoinPool workers are
> created, and which Subject the receiver's current user resolves to.
> * The controlled YARN application-mode probe showed no {{tryCompensate}} or
> {{createWorker}} in the provider window. The separate JobGraph-generation
> startup path did create FJP workers, confirming that YARN can enter the same
> FJP mechanism under different startup work; the provider's
> {{CompletableFuture}} blocking state remains unverified.
> * A deployed application-mode job should survive the 24h token deadline and
> continue checkpointing without SecretManager$InvalidToken. Verify that the
> checkpoint-storage DFSClient receives the new token, and verify the
> job-result-store client separately because it can capture a different UGI.
> h2. Acceptance criteria
> * A token refresh updates the UGI used by checkpoint storage and other
> long-lived Hadoop clients.
> * Refresh delivers credentials to the security-context UGI regardless of
> which Subject a pooled dispatcher or cluster-io worker inherited.
> * A long-running application-mode job survives the 24h HDFS token deadline.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)