You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "rkhachatryan (via GitHub)" <gi...@apache.org> on 2026/01/19 10:43:05 UTC
[PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
rkhachatryan opened a new pull request, #27441:
URL: https://github.com/apache/flink/pull/27441
(no comment)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "flinkbot (via GitHub)" <gi...@apache.org>.
flinkbot commented on PR #27441:
URL: https://github.com/apache/flink/pull/27441#issuecomment-3767683539
<!--
Meta data
{
"version" : 1,
"metaDataEntries" : [ {
"hash" : "e10a7c13557c4869b0fc3855eb9215bb60031d83",
"status" : "UNKNOWN",
"url" : "TBD",
"triggerID" : "e10a7c13557c4869b0fc3855eb9215bb60031d83",
"triggerType" : "PUSH"
} ]
}-->
## CI report:
* e10a7c13557c4869b0fc3855eb9215bb60031d83 UNKNOWN
<details>
<summary>Bot commands</summary>
The @flinkbot bot supports the following commands:
- `@flinkbot run azure` re-run the last Azure build
</details>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "davidradl (via GitHub)" <gi...@apache.org>.
davidradl commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2704757134
##########
docs/layouts/shortcodes/generated/checkpointing_configuration.html:
##########
@@ -122,6 +122,12 @@
<td>Duration</td>
<td>The minimal pause between checkpointing attempts. This setting defines how soon thecheckpoint coordinator may trigger another checkpoint after it becomes possible to triggeranother checkpoint with respect to the maximum number of concurrent checkpoints(see <code class="highlighter-rouge">execution.checkpointing.max-concurrent-checkpoints</code>).<br /><br />If the maximum number of concurrent checkpoints is set to one, this setting makes effectively sure that a minimum amount of time passes where no checkpoint is in progress at all.</td>
</tr>
+ <tr>
+ <td><h5>execution.checkpointing.pause-checkpoints-if-tasks-not-running</h5></td>
+ <td style="word-wrap: break-word;">false</td>
+ <td>Boolean</td>
+ <td>When enabled, checkpoints will be paused as long as some tasks are not running (and not finished), even if the job status is running. This help to prevent an additional delay on job startup</td>
Review Comment:
I am struggling with these sentences.
- nit it would be simpler to say _if any tasks_ rather than _as long as some tasks_
- What does it mean to say a task is not running, but has a status of running? I think we need to talk about non blocking output, maybe more words outside of the config definition to say how to recognise non blocking jobs - as this relates to the subset of jobs we are effecting with this option.
- nit: This help -> This helps
- the text says "prevent an additional delay on job startup". But the logic does not seem startup specific? Can we get into any other non startup related scenarios where this might be helpful
- Can we guide the user as to when they should consider using this option and any downside to using it.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "rkhachatryan (via GitHub)" <gi...@apache.org>.
rkhachatryan commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2712870774
##########
flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/CheckpointCoordinatorDeActivator.java:
##########
@@ -18,35 +18,33 @@
package org.apache.flink.runtime.checkpoint;
-import org.apache.flink.api.common.JobID;
import org.apache.flink.api.common.JobStatus;
+import org.apache.flink.runtime.execution.ExecutionState;
+import org.apache.flink.runtime.executiongraph.ExecutionAttemptID;
+import org.apache.flink.runtime.executiongraph.ExecutionStateUpdateListener;
import org.apache.flink.runtime.executiongraph.JobStatusListener;
-import static org.apache.flink.util.Preconditions.checkNotNull;
-
/**
* This actor listens to changes in the JobStatus and activates or deactivates the periodic
* checkpoint scheduler.
*/
-public class CheckpointCoordinatorDeActivator implements JobStatusListener {
-
- private final CheckpointCoordinator coordinator;
- private final boolean allTasksOutputNonBlocking;
+public interface CheckpointCoordinatorDeActivator
+ extends JobStatusListener, ExecutionStateUpdateListener {
- public CheckpointCoordinatorDeActivator(
+ static CheckpointCoordinatorDeActivator forJobStatus(
CheckpointCoordinator coordinator, boolean allTasksOutputNonBlocking) {
- this.coordinator = checkNotNull(coordinator);
- this.allTasksOutputNonBlocking = allTasksOutputNonBlocking;
+ return (jobId, newJobStatus, timestamp) -> {
+ if (newJobStatus == JobStatus.RUNNING && allTasksOutputNonBlocking) {
Review Comment:
The check is in `CheckpointCoordinator.createActivatorDeactivator`, but as mentioned above, I'm going to remove this option.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "davidradl (via GitHub)" <gi...@apache.org>.
davidradl commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2704757134
##########
docs/layouts/shortcodes/generated/checkpointing_configuration.html:
##########
@@ -122,6 +122,12 @@
<td>Duration</td>
<td>The minimal pause between checkpointing attempts. This setting defines how soon thecheckpoint coordinator may trigger another checkpoint after it becomes possible to triggeranother checkpoint with respect to the maximum number of concurrent checkpoints(see <code class="highlighter-rouge">execution.checkpointing.max-concurrent-checkpoints</code>).<br /><br />If the maximum number of concurrent checkpoints is set to one, this setting makes effectively sure that a minimum amount of time passes where no checkpoint is in progress at all.</td>
</tr>
+ <tr>
+ <td><h5>execution.checkpointing.pause-checkpoints-if-tasks-not-running</h5></td>
+ <td style="word-wrap: break-word;">false</td>
+ <td>Boolean</td>
+ <td>When enabled, checkpoints will be paused as long as some tasks are not running (and not finished), even if the job status is running. This help to prevent an additional delay on job startup</td>
Review Comment:
I am struggling with this sentence.
- nit it would be simpler to say _if any tasks_ rather than _as long as some tasks_
- What does it mean to say a task is not running, but has a status of running?
- nit: This help -> This helps
##########
docs/layouts/shortcodes/generated/checkpointing_configuration.html:
##########
@@ -122,6 +122,12 @@
<td>Duration</td>
<td>The minimal pause between checkpointing attempts. This setting defines how soon thecheckpoint coordinator may trigger another checkpoint after it becomes possible to triggeranother checkpoint with respect to the maximum number of concurrent checkpoints(see <code class="highlighter-rouge">execution.checkpointing.max-concurrent-checkpoints</code>).<br /><br />If the maximum number of concurrent checkpoints is set to one, this setting makes effectively sure that a minimum amount of time passes where no checkpoint is in progress at all.</td>
</tr>
+ <tr>
+ <td><h5>execution.checkpointing.pause-checkpoints-if-tasks-not-running</h5></td>
+ <td style="word-wrap: break-word;">false</td>
+ <td>Boolean</td>
+ <td>When enabled, checkpoints will be paused as long as some tasks are not running (and not finished), even if the job status is running. This help to prevent an additional delay on job startup</td>
Review Comment:
I am struggling with these sentences.
- nit it would be simpler to say _if any tasks_ rather than _as long as some tasks_
- What does it mean to say a task is not running, but has a status of running?
- nit: This help -> This helps
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "davidradl (via GitHub)" <gi...@apache.org>.
davidradl commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2704757134
##########
docs/layouts/shortcodes/generated/checkpointing_configuration.html:
##########
@@ -122,6 +122,12 @@
<td>Duration</td>
<td>The minimal pause between checkpointing attempts. This setting defines how soon thecheckpoint coordinator may trigger another checkpoint after it becomes possible to triggeranother checkpoint with respect to the maximum number of concurrent checkpoints(see <code class="highlighter-rouge">execution.checkpointing.max-concurrent-checkpoints</code>).<br /><br />If the maximum number of concurrent checkpoints is set to one, this setting makes effectively sure that a minimum amount of time passes where no checkpoint is in progress at all.</td>
</tr>
+ <tr>
+ <td><h5>execution.checkpointing.pause-checkpoints-if-tasks-not-running</h5></td>
+ <td style="word-wrap: break-word;">false</td>
+ <td>Boolean</td>
+ <td>When enabled, checkpoints will be paused as long as some tasks are not running (and not finished), even if the job status is running. This help to prevent an additional delay on job startup</td>
Review Comment:
I am struggling with these sentences.
- nit it would be simpler to say _if any tasks_ rather than _as long as some tasks_
- What does it mean to say a task is not running, but has a status of running?
- nit: This help -> This helps
- the text says "prevent an additional delay on job startup". But the logic does not seem startup specific? Can we get into any other non startup related scenarios where this might be helpful
- Can we guide the user as to when they should consider using this option.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "1996fanrui (via GitHub)" <gi...@apache.org>.
1996fanrui commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2712921226
##########
tools/maven/checkstyle.xml:
##########
@@ -61,7 +61,7 @@ This file is based on the checkstyle file of Apache Beam.
-->
<module name="FileLength">
- <property name="max" value="3100"/>
+ <property name="max" value="4000"/>
Review Comment:
It does not seem a good idea. Generally, it is the time to refactor or something like this when reaching this limitation.
##########
flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/CheckpointCoordinatorDeActivator.java:
##########
@@ -18,35 +18,33 @@
package org.apache.flink.runtime.checkpoint;
-import org.apache.flink.api.common.JobID;
import org.apache.flink.api.common.JobStatus;
+import org.apache.flink.runtime.execution.ExecutionState;
+import org.apache.flink.runtime.executiongraph.ExecutionAttemptID;
+import org.apache.flink.runtime.executiongraph.ExecutionStateUpdateListener;
import org.apache.flink.runtime.executiongraph.JobStatusListener;
-import static org.apache.flink.util.Preconditions.checkNotNull;
-
/**
* This actor listens to changes in the JobStatus and activates or deactivates the periodic
* checkpoint scheduler.
*/
-public class CheckpointCoordinatorDeActivator implements JobStatusListener {
-
- private final CheckpointCoordinator coordinator;
- private final boolean allTasksOutputNonBlocking;
+public interface CheckpointCoordinatorDeActivator
+ extends JobStatusListener, ExecutionStateUpdateListener {
- public CheckpointCoordinatorDeActivator(
+ static CheckpointCoordinatorDeActivator forJobStatus(
CheckpointCoordinator coordinator, boolean allTasksOutputNonBlocking) {
- this.coordinator = checkNotNull(coordinator);
- this.allTasksOutputNonBlocking = allTasksOutputNonBlocking;
+ return (jobId, newJobStatus, timestamp) -> {
+ if (newJobStatus == JobStatus.RUNNING && allTasksOutputNonBlocking) {
+ // start the checkpoint scheduler if there is no blocking edge
+ coordinator.startCheckpointScheduler();
+ } else {
+ // anything else should stop the trigger for now
+ coordinator.stopCheckpointScheduler();
+ }
Review Comment:
`forJobStatus` is only called once, and `allTasksOutputNonBlocking` is false. so `forJobStatus` can be simplified.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "davidradl (via GitHub)" <gi...@apache.org>.
davidradl commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2704825685
##########
flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/CheckpointCoordinator.java:
##########
@@ -2146,18 +2148,24 @@ private void restoreStateToCoordinators(
// job status listener that schedules / cancels periodic checkpoints
// ------------------------------------------------------------------------
- public JobStatusListener createActivatorDeactivator(boolean allTasksOutputNonBlocking) {
+ public CheckpointCoordinatorDeActivator createActivatorDeactivator(
+ boolean allTasksOutputNonBlocking) {
synchronized (lock) {
if (shutdown) {
throw new IllegalArgumentException("Checkpoint coordinator is shut down");
}
- if (jobStatusListener == null) {
- jobStatusListener =
- new CheckpointCoordinatorDeActivator(this, allTasksOutputNonBlocking);
+ if (deActivator == null) {
+ if (pauseCheckpointsIfTasksNotRunning && allTasksOutputNonBlocking) {
+ deActivator = new ExecutionTrackingCheckpointCoordinatorDeActivator(this);
Review Comment:
Maybe a comment to indicate why this if body results in a pause.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "rkhachatryan (via GitHub)" <gi...@apache.org>.
rkhachatryan commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2712857043
##########
docs/layouts/shortcodes/generated/checkpointing_configuration.html:
##########
@@ -122,6 +122,12 @@
<td>Duration</td>
<td>The minimal pause between checkpointing attempts. This setting defines how soon thecheckpoint coordinator may trigger another checkpoint after it becomes possible to triggeranother checkpoint with respect to the maximum number of concurrent checkpoints(see <code class="highlighter-rouge">execution.checkpointing.max-concurrent-checkpoints</code>).<br /><br />If the maximum number of concurrent checkpoints is set to one, this setting makes effectively sure that a minimum amount of time passes where no checkpoint is in progress at all.</td>
</tr>
+ <tr>
+ <td><h5>execution.checkpointing.pause-checkpoints-if-tasks-not-running</h5></td>
+ <td style="word-wrap: break-word;">false</td>
+ <td>Boolean</td>
+ <td>When enabled, checkpoints will be paused as long as some tasks are not running (and not finished), even if the job status is running. This help to prevent an additional delay on job startup</td>
Review Comment:
We discussed this PR with @1996fanrui offline and agreed to remove the option because the new behavior is the only desirable (until FLIP-457). After FLIP-457, this logic will be changed.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "rkhachatryan (via GitHub)" <gi...@apache.org>.
rkhachatryan commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2719088704
##########
flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/CheckpointCoordinatorDeActivator.java:
##########
@@ -18,35 +18,33 @@
package org.apache.flink.runtime.checkpoint;
-import org.apache.flink.api.common.JobID;
import org.apache.flink.api.common.JobStatus;
+import org.apache.flink.runtime.execution.ExecutionState;
+import org.apache.flink.runtime.executiongraph.ExecutionAttemptID;
+import org.apache.flink.runtime.executiongraph.ExecutionStateUpdateListener;
import org.apache.flink.runtime.executiongraph.JobStatusListener;
-import static org.apache.flink.util.Preconditions.checkNotNull;
-
/**
* This actor listens to changes in the JobStatus and activates or deactivates the periodic
* checkpoint scheduler.
*/
-public class CheckpointCoordinatorDeActivator implements JobStatusListener {
-
- private final CheckpointCoordinator coordinator;
- private final boolean allTasksOutputNonBlocking;
+public interface CheckpointCoordinatorDeActivator
+ extends JobStatusListener, ExecutionStateUpdateListener {
- public CheckpointCoordinatorDeActivator(
+ static CheckpointCoordinatorDeActivator forJobStatus(
CheckpointCoordinator coordinator, boolean allTasksOutputNonBlocking) {
- this.coordinator = checkNotNull(coordinator);
- this.allTasksOutputNonBlocking = allTasksOutputNonBlocking;
+ return (jobId, newJobStatus, timestamp) -> {
+ if (newJobStatus == JobStatus.RUNNING && allTasksOutputNonBlocking) {
+ // start the checkpoint scheduler if there is no blocking edge
+ coordinator.startCheckpointScheduler();
+ } else {
+ // anything else should stop the trigger for now
+ coordinator.stopCheckpointScheduler();
+ }
Review Comment:
Right, I've simplified the code, PTAL.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "davidradl (via GitHub)" <gi...@apache.org>.
davidradl commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2704816254
##########
flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/CheckpointCoordinatorDeActivator.java:
##########
@@ -18,35 +18,33 @@
package org.apache.flink.runtime.checkpoint;
-import org.apache.flink.api.common.JobID;
import org.apache.flink.api.common.JobStatus;
+import org.apache.flink.runtime.execution.ExecutionState;
+import org.apache.flink.runtime.executiongraph.ExecutionAttemptID;
+import org.apache.flink.runtime.executiongraph.ExecutionStateUpdateListener;
import org.apache.flink.runtime.executiongraph.JobStatusListener;
-import static org.apache.flink.util.Preconditions.checkNotNull;
-
/**
* This actor listens to changes in the JobStatus and activates or deactivates the periodic
* checkpoint scheduler.
*/
-public class CheckpointCoordinatorDeActivator implements JobStatusListener {
-
- private final CheckpointCoordinator coordinator;
- private final boolean allTasksOutputNonBlocking;
+public interface CheckpointCoordinatorDeActivator
+ extends JobStatusListener, ExecutionStateUpdateListener {
- public CheckpointCoordinatorDeActivator(
+ static CheckpointCoordinatorDeActivator forJobStatus(
CheckpointCoordinator coordinator, boolean allTasksOutputNonBlocking) {
- this.coordinator = checkNotNull(coordinator);
- this.allTasksOutputNonBlocking = allTasksOutputNonBlocking;
+ return (jobId, newJobStatus, timestamp) -> {
+ if (newJobStatus == JobStatus.RUNNING && allTasksOutputNonBlocking) {
Review Comment:
I an curious - should we be checking the new config option here? I thought this is the condition we wanted to pause checkpointing for? Am I missing something?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "davidradl (via GitHub)" <gi...@apache.org>.
davidradl commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2704757134
##########
docs/layouts/shortcodes/generated/checkpointing_configuration.html:
##########
@@ -122,6 +122,12 @@
<td>Duration</td>
<td>The minimal pause between checkpointing attempts. This setting defines how soon thecheckpoint coordinator may trigger another checkpoint after it becomes possible to triggeranother checkpoint with respect to the maximum number of concurrent checkpoints(see <code class="highlighter-rouge">execution.checkpointing.max-concurrent-checkpoints</code>).<br /><br />If the maximum number of concurrent checkpoints is set to one, this setting makes effectively sure that a minimum amount of time passes where no checkpoint is in progress at all.</td>
</tr>
+ <tr>
+ <td><h5>execution.checkpointing.pause-checkpoints-if-tasks-not-running</h5></td>
+ <td style="word-wrap: break-word;">false</td>
+ <td>Boolean</td>
+ <td>When enabled, checkpoints will be paused as long as some tasks are not running (and not finished), even if the job status is running. This help to prevent an additional delay on job startup</td>
Review Comment:
I am struggling with these sentences.
- nit it would be simpler to say _if any tasks_ rather than _as long as some tasks_
- What does it mean to say a task is not running, but has a status of running?
- nit: This help -> This helps
- the text says "prevent an additional delay on job startup". But the logic does not seem startup specific? Can we get into any other non startup related scenarios where this might be helpful
- Can we guide the user as to when they should consider using this option and any downside to using it.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "davidradl (via GitHub)" <gi...@apache.org>.
davidradl commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2704757134
##########
docs/layouts/shortcodes/generated/checkpointing_configuration.html:
##########
@@ -122,6 +122,12 @@
<td>Duration</td>
<td>The minimal pause between checkpointing attempts. This setting defines how soon thecheckpoint coordinator may trigger another checkpoint after it becomes possible to triggeranother checkpoint with respect to the maximum number of concurrent checkpoints(see <code class="highlighter-rouge">execution.checkpointing.max-concurrent-checkpoints</code>).<br /><br />If the maximum number of concurrent checkpoints is set to one, this setting makes effectively sure that a minimum amount of time passes where no checkpoint is in progress at all.</td>
</tr>
+ <tr>
+ <td><h5>execution.checkpointing.pause-checkpoints-if-tasks-not-running</h5></td>
+ <td style="word-wrap: break-word;">false</td>
+ <td>Boolean</td>
+ <td>When enabled, checkpoints will be paused as long as some tasks are not running (and not finished), even if the job status is running. This help to prevent an additional delay on job startup</td>
Review Comment:
I am struggling with these sentences.
- nit it would be simpler to say _if any tasks_ rather than _as long as some tasks_
- What does it mean to say a task is not running, but has a status of running? I think we need to talk about non blocking output, maybe more words outside of the config definition to say how to recognise non blocking jobs - as these seem to be the subset of jobs we are effecting here.
- nit: This help -> This helps
- the text says "prevent an additional delay on job startup". But the logic does not seem startup specific? Can we get into any other non startup related scenarios where this might be helpful
- Can we guide the user as to when they should consider using this option and any downside to using it.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "rkhachatryan (via GitHub)" <gi...@apache.org>.
rkhachatryan merged PR #27441:
URL: https://github.com/apache/flink/pull/27441
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [FLINK-38940] Start checkpointing when all tasks are running and job is running [flink]
Posted by "rkhachatryan (via GitHub)" <gi...@apache.org>.
rkhachatryan commented on code in PR #27441:
URL: https://github.com/apache/flink/pull/27441#discussion_r2719087880
##########
tools/maven/checkstyle.xml:
##########
@@ -61,7 +61,7 @@ This file is based on the checkstyle file of Apache Beam.
-->
<module name="FileLength">
- <property name="max" value="3100"/>
+ <property name="max" value="4000"/>
Review Comment:
This actually wasn't necessary, removed this change.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org