You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "ericm-db (via GitHub)" <gi...@apache.org> on 2023/07/19 00:25:57 UTC
[GitHub] [spark] ericm-db opened a new pull request, #42066: Maintenance thread pool optional
ericm-db opened a new pull request, #42066:
URL: https://github.com/apache/spark/pull/42066
<!--
Thanks for sending a pull request! Here are some tips for you:
1. If this is your first time, please read our contributor guidelines: https://spark.apache.org/contributing.html
2. Ensure you have added or run the appropriate tests for your PR: https://spark.apache.org/developer-tools.html
3. If the PR is unfinished, add '[WIP]' in your PR title, e.g., '[WIP][SPARK-XXXX] Your PR title ...'.
4. Be sure to keep the PR description updated to reflect all changes.
5. Please write your PR title to summarize what this PR proposes.
6. If possible, provide a concise example to reproduce the issue for a faster review.
7. If you want to add a new configuration, please read the guideline first for naming configurations in
'core/src/main/scala/org/apache/spark/internal/config/ConfigEntry.scala'.
8. If you want to add or modify an error type or message, please read the guideline first in
'core/src/main/resources/error/README.md'.
-->
### What changes were proposed in this pull request?
<!--
Please clarify what changes you are proposing. The purpose of this section is to outline the changes and how this PR fixes the issue.
If possible, please consider writing useful notes for better and faster reviews in your PR. See the examples below.
1. If you refactor some codes with changing classes, showing the class hierarchy will help reviewers.
2. If you fix some SQL features, you can provide some references of other DBMSes.
3. If there is design documentation, please add the link.
4. If there is a discussion in the mailing list, please add the link.
-->
### Why are the changes needed?
<!--
Please clarify why the changes are needed. For instance,
1. If you propose a new API, clarify the use case for a new API.
2. If you fix a bug, you can clarify why it is a bug.
-->
### Does this PR introduce _any_ user-facing change?
<!--
Note that it means *any* user-facing change including all aspects such as the documentation fix.
If yes, please clarify the previous behavior and the change this PR proposes - provide the console output, description and/or an example to show the behavior difference if possible.
If possible, please also clarify if this is a user-facing change compared to the released Spark versions or within the unreleased branches such as master.
If no, write 'No'.
-->
### How was this patch tested?
<!--
If tests were added, say they were added here. Please make sure to add some test cases that check the changes thoroughly including negative and positive cases if possible.
If it was tested in a way different from regular unit tests, please clarify how you tested step by step, ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future.
If tests were not added, please describe why they were not added and/or why it was difficult to add.
If benchmark tests were added, please run the benchmarks in GitHub Actions for the consistent environment, and the instructions could accord to: https://spark.apache.org/developer-tools.html#github-workflow-benchmarks.
-->
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280179534
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -636,17 +693,44 @@ object StateStore extends Logging {
if (SparkEnv.get == null) {
throw new IllegalStateException("SparkEnv not active, cannot do maintenance on StateStores")
}
- loadedProviders.synchronized { loadedProviders.toSeq }.foreach { case (id, provider) =>
- try {
- provider.doMaintenance()
- if (!verifyIfStoreInstanceActive(id)) {
- unload(id)
- logInfo(s"Unloaded $provider")
- }
- } catch {
- case NonFatal(e) =>
- logWarning(s"Error managing $provider, stopping management thread")
- throw e
+ loadedProviders.synchronized {
+ loadedProviders.toSeq
+ }.foreach { case (id, provider) =>
+ // check exception
+ if (threadPoolException.get() != null) {
+ val exception = threadPoolException.get()
+ logWarning("Error in maintenanceThreadPool", exception)
+ throw exception
+ }
+ if (processThisPartition(id)) {
Review Comment:
Can we log INFO message in the `else` case if we are not scheduling maintenance for a given partition if one is ongoing ?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] ericm-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "ericm-db (via GitHub)" <gi...@apache.org>.
ericm-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280181850
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -440,6 +441,16 @@ object StateStore extends Logging {
@GuardedBy("loadedProviders")
private val schemaValidated = new mutable.HashMap[StateStoreProviderId, Option[Throwable]]()
+ private val maintenanceThreadPoolLock = new Object
+
+ private val threadPoolException = new AtomicReference[Throwable](null)
Review Comment:
Yeah this was an approach outlined by Jungtaek which we also use in async progress tracking.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280065414
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala:
##########
@@ -1852,6 +1852,17 @@ object SQLConf {
.createWithDefault(
"org.apache.spark.sql.execution.streaming.state.HDFSBackedStateStoreProvider")
+ val NUM_STATE_STORE_MAINTENANCE_THREADS =
+ buildConf("spark.sql.streaming.stateStore.numStateStoreMaintenanceThreads")
+ .internal()
+ .doc("Number of threads in the thread pool that perform clean up and snapshotting tasks " +
+ "for stateful streaming queries. The default value is 2 so that this thread pool " +
Review Comment:
nit: the explanation of default value is no longer true. need to update.
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -587,6 +622,8 @@ object StateStore extends Logging {
/** Stop maintenance thread and reset the maintenance task */
def stopMaintenanceTask(): Unit = loadedProviders.synchronized {
if (maintenanceTask != null) {
+ maintenanceThreadPool.stop()
Review Comment:
1. Let's cleanup relevant variables as well, `threadPoolException`, and `maintenancePartitions`. It should be necessary to re-initialize maintenanceTask.
2. For the full safety, I'd have two if statements, to handle the cleanup of `maintenanceThreadPool` and `maintenanceTask` separately.
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -628,17 +687,44 @@ object StateStore extends Logging {
if (SparkEnv.get == null) {
throw new IllegalStateException("SparkEnv not active, cannot do maintenance on StateStores")
}
- loadedProviders.synchronized { loadedProviders.toSeq }.foreach { case (id, provider) =>
- try {
- provider.doMaintenance()
- if (!verifyIfStoreInstanceActive(id)) {
- unload(id)
- logInfo(s"Unloaded $provider")
- }
- } catch {
- case NonFatal(e) =>
- logWarning(s"Error managing $provider, stopping management thread")
- throw e
+ loadedProviders.synchronized {
+ loadedProviders.toSeq
+ }.foreach { case (id, provider) =>
+ // check exception
+ if (threadPoolException.get() != null) {
Review Comment:
First of all, this is not thread-safe although we use AtomicReference, as we are dealing with two different operations.
Second, if we reset the exception here, other partition(s) will likely still schedule the maintenance task, which may not be something we desire to. We may want to skip scheduling others once the error was found.
Third, exception could be "null" (race condition can happen - imagine two threads executing L695 concurrently) and it could lead to other issue on throwing null.
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -602,22 +639,44 @@ object StateStore extends Logging {
}
/** Start the periodic maintenance task if not already started and if Spark active */
- private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit =
+ private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit = {
+ val numMaintenanceThreads = storeConf.numStateStoreMaintenanceThreads
loadedProviders.synchronized {
if (SparkEnv.get != null && !isMaintenanceRunning) {
maintenanceTask = new MaintenanceTask(
storeConf.maintenanceInterval,
- task = { doMaintenance() },
- onError = { loadedProviders.synchronized {
+ task = {
+ doMaintenance()
+ },
+ onError = {
+ loadedProviders.synchronized {
logInfo("Stopping maintenance task since an error was encountered.")
stopMaintenanceTask()
+ // SPARK-44504 - Unload explicitly to force closing underlying DB instance
Review Comment:
nit: this looks odd, the change should be already done. Could you please rebase with the latest master?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] ericm-db commented on pull request #42066: [SPARK-44480][SS] Add option for thread pool to perform maintenance for RocksDB/HDFS State Store Providers
Posted by "ericm-db (via GitHub)" <gi...@apache.org>.
ericm-db commented on PR #42066:
URL: https://github.com/apache/spark/pull/42066#issuecomment-1641413593
> @ericm-db I forgot you had a PR. #42012 Does this PR supersede the old PR? It looks like so, but just to double confirm.
@HeartSaVioR yes, this does supersede the old PR, thanks for checking. Will change the title accordingly
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] ericm-db commented on pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "ericm-db (via GitHub)" <gi...@apache.org>.
ericm-db commented on PR #42066:
URL: https://github.com/apache/spark/pull/42066#issuecomment-1642470993
> CI looks failing. Could you please look into it? https://github.com/ericm-db/spark/actions/runs/5595467996/jobs/10231376237
@HeartSaVioR looks like I was setting the runnable maintenance task incorrectly. It should be fine now
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1281035204
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -636,17 +691,46 @@ object StateStore extends Logging {
if (SparkEnv.get == null) {
throw new IllegalStateException("SparkEnv not active, cannot do maintenance on StateStores")
}
- loadedProviders.synchronized { loadedProviders.toSeq }.foreach { case (id, provider) =>
- try {
- provider.doMaintenance()
- if (!verifyIfStoreInstanceActive(id)) {
- unload(id)
- logInfo(s"Unloaded $provider")
- }
- } catch {
- case NonFatal(e) =>
- logWarning(s"Error managing $provider, stopping management thread")
- throw e
+ loadedProviders.synchronized {
+ loadedProviders.toSeq
+ }.foreach { case (id, provider) =>
+ // check exception
+ if (threadPoolException.get() != null) {
+ val exception = threadPoolException.get()
+ logWarning("Error in maintenanceThreadPool", exception)
+ throw exception
+ }
+ if (processThisPartition(id)) {
+ maintenanceThreadPool.execute(() => {
+ val startTime = System.currentTimeMillis()
+ try {
+ provider.doMaintenance()
+ if (!verifyIfStoreInstanceActive(id)) {
+ unload(id)
+ logInfo(s"Unloaded $provider")
+ }
+ } catch {
+ case NonFatal(e) =>
+ logWarning(s"Error managing $provider, stopping management thread", e)
+ threadPoolException.set(e)
+ throw e
Review Comment:
yea I guess this exception would be swallowed by the thread pool utils
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280124681
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -607,25 +650,42 @@ object StateStore extends Logging {
}
/** Start the periodic maintenance task if not already started and if Spark active */
- private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit =
+ private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit = {
+ val numMaintenanceThreads = storeConf.numStateStoreMaintenanceThreads
loadedProviders.synchronized {
if (SparkEnv.get != null && !isMaintenanceRunning) {
maintenanceTask = new MaintenanceTask(
storeConf.maintenanceInterval,
- task = { doMaintenance() },
- onError = { loadedProviders.synchronized {
+ task = {
+ doMaintenance()
+ },
+ onError = {
+ loadedProviders.synchronized {
logInfo("Stopping maintenance task since an error was encountered.")
stopMaintenanceTask()
- // SPARK-44504 - Unload explicitly to force closing underlying DB instance
- // and releasing allocated resources, especially for RocksDBStateStoreProvider.
loadedProviders.keySet.foreach { key => unload(key) }
loadedProviders.clear()
}
}
)
+ maintenanceThreadPool = new MaintenanceThreadPool(
+ numThreads = numMaintenanceThreads
Review Comment:
We can just pass `numMaintenanceThreads` right ? we only have 1 argument for the class ?
```
maintenanceThreadPool = new MaintenanceThreadPool(numMaintenanceThreads)
``` ?
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -607,25 +650,42 @@ object StateStore extends Logging {
}
/** Start the periodic maintenance task if not already started and if Spark active */
- private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit =
+ private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit = {
+ val numMaintenanceThreads = storeConf.numStateStoreMaintenanceThreads
loadedProviders.synchronized {
if (SparkEnv.get != null && !isMaintenanceRunning) {
maintenanceTask = new MaintenanceTask(
storeConf.maintenanceInterval,
- task = { doMaintenance() },
- onError = { loadedProviders.synchronized {
+ task = {
+ doMaintenance()
+ },
+ onError = {
+ loadedProviders.synchronized {
logInfo("Stopping maintenance task since an error was encountered.")
stopMaintenanceTask()
- // SPARK-44504 - Unload explicitly to force closing underlying DB instance
- // and releasing allocated resources, especially for RocksDBStateStoreProvider.
loadedProviders.keySet.foreach { key => unload(key) }
loadedProviders.clear()
}
}
)
+ maintenanceThreadPool = new MaintenanceThreadPool(
+ numThreads = numMaintenanceThreads
Review Comment:
We can just pass `numMaintenanceThreads` right ? we only have 1 argument for the class ?
```
maintenanceThreadPool = new MaintenanceThreadPool(numMaintenanceThreads)
```
?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280125591
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -607,25 +650,42 @@ object StateStore extends Logging {
}
/** Start the periodic maintenance task if not already started and if Spark active */
- private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit =
+ private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit = {
+ val numMaintenanceThreads = storeConf.numStateStoreMaintenanceThreads
loadedProviders.synchronized {
if (SparkEnv.get != null && !isMaintenanceRunning) {
maintenanceTask = new MaintenanceTask(
storeConf.maintenanceInterval,
- task = { doMaintenance() },
- onError = { loadedProviders.synchronized {
+ task = {
+ doMaintenance()
+ },
+ onError = {
+ loadedProviders.synchronized {
logInfo("Stopping maintenance task since an error was encountered.")
stopMaintenanceTask()
- // SPARK-44504 - Unload explicitly to force closing underlying DB instance
Review Comment:
Looks like some conflict happened. Could you please revert to the origin 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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1267562812
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala:
##########
@@ -1852,6 +1852,24 @@ object SQLConf {
.createWithDefault(
"org.apache.spark.sql.execution.streaming.state.HDFSBackedStateStoreProvider")
+ val ENABLE_STATE_STORE_MAINTENANCE_THREAD_POOL =
+ buildConf("spark.sql.streaming.stateStore.enableStateStoreMaintenanceThreadPool")
+ .internal()
+ .doc("Whether background maintenance tasks are executed by single thread or thread pool")
Review Comment:
Shall we provide the context to inform the reason the old mechanism is set to default? Something along the line (please refine the sentence): the thread pool is consuming more resource than single thread, which might affect the performance of the query. Enable this if you observe the issue with the maintenance task.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR commented on pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on PR #42066:
URL: https://github.com/apache/spark/pull/42066#issuecomment-1641586354
CI looks failing. Could you please look into it?
https://github.com/ericm-db/spark/actions/runs/5595467996/jobs/10231376237
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR commented on pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on PR #42066:
URL: https://github.com/apache/spark/pull/42066#issuecomment-1660967631
Thanks! Merging to master.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280177057
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -440,6 +441,16 @@ object StateStore extends Logging {
@GuardedBy("loadedProviders")
private val schemaValidated = new mutable.HashMap[StateStoreProviderId, Option[Throwable]]()
+ private val maintenanceThreadPoolLock = new Object
+
+ private val threadPoolException = new AtomicReference[Throwable](null)
Review Comment:
Can we add a comment here ? @HeartSaVioR - is this the preferred method to bubble up exceptions from the thread pool ?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1281027560
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -636,17 +691,46 @@ object StateStore extends Logging {
if (SparkEnv.get == null) {
throw new IllegalStateException("SparkEnv not active, cannot do maintenance on StateStores")
}
- loadedProviders.synchronized { loadedProviders.toSeq }.foreach { case (id, provider) =>
- try {
- provider.doMaintenance()
- if (!verifyIfStoreInstanceActive(id)) {
- unload(id)
- logInfo(s"Unloaded $provider")
- }
- } catch {
- case NonFatal(e) =>
- logWarning(s"Error managing $provider, stopping management thread")
- throw e
+ loadedProviders.synchronized {
+ loadedProviders.toSeq
+ }.foreach { case (id, provider) =>
+ // check exception
+ if (threadPoolException.get() != null) {
+ val exception = threadPoolException.get()
+ logWarning("Error in maintenanceThreadPool", exception)
+ throw exception
+ }
+ if (processThisPartition(id)) {
+ maintenanceThreadPool.execute(() => {
+ val startTime = System.currentTimeMillis()
+ try {
+ provider.doMaintenance()
+ if (!verifyIfStoreInstanceActive(id)) {
+ unload(id)
+ logInfo(s"Unloaded $provider")
+ }
+ } catch {
+ case NonFatal(e) =>
+ logWarning(s"Error managing $provider, stopping management thread", e)
+ threadPoolException.set(e)
+ throw e
Review Comment:
In this single threaded case, its caught by the maintenance thread itself right ? we are executing everything within the context of a single thread
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR closed pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR closed pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
URL: https://github.com/apache/spark/pull/42066
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280180687
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -636,17 +693,44 @@ object StateStore extends Logging {
if (SparkEnv.get == null) {
throw new IllegalStateException("SparkEnv not active, cannot do maintenance on StateStores")
}
- loadedProviders.synchronized { loadedProviders.toSeq }.foreach { case (id, provider) =>
- try {
- provider.doMaintenance()
- if (!verifyIfStoreInstanceActive(id)) {
- unload(id)
- logInfo(s"Unloaded $provider")
- }
- } catch {
- case NonFatal(e) =>
- logWarning(s"Error managing $provider, stopping management thread")
- throw e
+ loadedProviders.synchronized {
+ loadedProviders.toSeq
+ }.foreach { case (id, provider) =>
+ // check exception
+ if (threadPoolException.get() != null) {
+ val exception = threadPoolException.get()
+ logWarning("Error in maintenanceThreadPool", exception)
+ throw exception
+ }
+ if (processThisPartition(id)) {
+ maintenanceThreadPool.execute(() => {
+ val startTime = System.currentTimeMillis()
+ try {
+ provider.doMaintenance()
+ if (!verifyIfStoreInstanceActive(id)) {
+ unload(id)
+ logInfo(s"Unloaded $provider")
+ }
+ } catch {
+ case NonFatal(e) =>
+ logWarning(s"Error managing $provider, stopping management thread", e)
+ threadPoolException.set(e)
+ throw e
+ } finally {
+ val endTime = System.currentTimeMillis()
+ val duration = endTime - startTime
Review Comment:
Nit: I guess we can just do
```
val duration = System.currentTimeMillis() - startTime
```
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1281021882
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -636,17 +691,46 @@ object StateStore extends Logging {
if (SparkEnv.get == null) {
throw new IllegalStateException("SparkEnv not active, cannot do maintenance on StateStores")
}
- loadedProviders.synchronized { loadedProviders.toSeq }.foreach { case (id, provider) =>
- try {
- provider.doMaintenance()
- if (!verifyIfStoreInstanceActive(id)) {
- unload(id)
- logInfo(s"Unloaded $provider")
- }
- } catch {
- case NonFatal(e) =>
- logWarning(s"Error managing $provider, stopping management thread")
- throw e
+ loadedProviders.synchronized {
+ loadedProviders.toSeq
+ }.foreach { case (id, provider) =>
+ // check exception
+ if (threadPoolException.get() != null) {
+ val exception = threadPoolException.get()
+ logWarning("Error in maintenanceThreadPool", exception)
+ throw exception
+ }
+ if (processThisPartition(id)) {
+ maintenanceThreadPool.execute(() => {
+ val startTime = System.currentTimeMillis()
+ try {
+ provider.doMaintenance()
+ if (!verifyIfStoreInstanceActive(id)) {
+ unload(id)
+ logInfo(s"Unloaded $provider")
+ }
+ } catch {
+ case NonFatal(e) =>
+ logWarning(s"Error managing $provider, stopping management thread", e)
+ threadPoolException.set(e)
+ throw e
Review Comment:
is this exception caught by a higher level thread pool utility function ? Will it just log this exception in that case ?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR commented on pull request #42066: [SPARK-44480][SS] Add option for thread pool to perform maintenance for RocksDB/HDFS State Store Providers
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on PR #42066:
URL: https://github.com/apache/spark/pull/42066#issuecomment-1641409801
Btw, PR title from #42012 seems to be more proper.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280123789
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -607,25 +650,42 @@ object StateStore extends Logging {
}
/** Start the periodic maintenance task if not already started and if Spark active */
- private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit =
+ private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit = {
+ val numMaintenanceThreads = storeConf.numStateStoreMaintenanceThreads
loadedProviders.synchronized {
if (SparkEnv.get != null && !isMaintenanceRunning) {
maintenanceTask = new MaintenanceTask(
storeConf.maintenanceInterval,
- task = { doMaintenance() },
- onError = { loadedProviders.synchronized {
+ task = {
+ doMaintenance()
+ },
+ onError = {
+ loadedProviders.synchronized {
logInfo("Stopping maintenance task since an error was encountered.")
stopMaintenanceTask()
- // SPARK-44504 - Unload explicitly to force closing underlying DB instance
Review Comment:
Is it intentional to remove this ?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280123962
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -607,25 +650,42 @@ object StateStore extends Logging {
}
/** Start the periodic maintenance task if not already started and if Spark active */
- private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit =
+ private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit = {
+ val numMaintenanceThreads = storeConf.numStateStoreMaintenanceThreads
loadedProviders.synchronized {
if (SparkEnv.get != null && !isMaintenanceRunning) {
maintenanceTask = new MaintenanceTask(
storeConf.maintenanceInterval,
- task = { doMaintenance() },
- onError = { loadedProviders.synchronized {
+ task = {
Review Comment:
Can we revert to the original formatting ? There is no change here, correct ?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] ericm-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "ericm-db (via GitHub)" <gi...@apache.org>.
ericm-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280181850
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -440,6 +441,16 @@ object StateStore extends Logging {
@GuardedBy("loadedProviders")
private val schemaValidated = new mutable.HashMap[StateStoreProviderId, Option[Throwable]]()
+ private val maintenanceThreadPoolLock = new Object
+
+ private val threadPoolException = new AtomicReference[Throwable](null)
Review Comment:
Yeah this was an approach outlined by Jungtaek which we also use in microbatch pipelining
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -472,9 +483,33 @@ object StateStore extends Logging {
def isRunning: Boolean = !future.isDone
}
+ /**
+ * Thread Pool that runs maintenance on partitions that are scheduled by
+ * MaintenanceTask periodically
+ */
+ class MaintenanceThreadPool(numThreads: Int) {
+ private val threadPool = ThreadUtils.newDaemonFixedThreadPool(
+ numThreads, "state-store-maintenance-thread")
+
+ def execute(runnable: Runnable): Unit = {
+ threadPool.execute(runnable)
+ }
+
+ def isRunning(): Boolean = {
Review Comment:
Nowhere, can remove 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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] ericm-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "ericm-db (via GitHub)" <gi...@apache.org>.
ericm-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1281023328
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -636,17 +691,46 @@ object StateStore extends Logging {
if (SparkEnv.get == null) {
throw new IllegalStateException("SparkEnv not active, cannot do maintenance on StateStores")
}
- loadedProviders.synchronized { loadedProviders.toSeq }.foreach { case (id, provider) =>
- try {
- provider.doMaintenance()
- if (!verifyIfStoreInstanceActive(id)) {
- unload(id)
- logInfo(s"Unloaded $provider")
- }
- } catch {
- case NonFatal(e) =>
- logWarning(s"Error managing $provider, stopping management thread")
- throw e
+ loadedProviders.synchronized {
+ loadedProviders.toSeq
+ }.foreach { case (id, provider) =>
+ // check exception
+ if (threadPoolException.get() != null) {
+ val exception = threadPoolException.get()
+ logWarning("Error in maintenanceThreadPool", exception)
+ throw exception
+ }
+ if (processThisPartition(id)) {
+ maintenanceThreadPool.execute(() => {
+ val startTime = System.currentTimeMillis()
+ try {
+ provider.doMaintenance()
+ if (!verifyIfStoreInstanceActive(id)) {
+ unload(id)
+ logInfo(s"Unloaded $provider")
+ }
+ } catch {
+ case NonFatal(e) =>
+ logWarning(s"Error managing $provider, stopping management thread", e)
+ threadPoolException.set(e)
+ throw e
Review Comment:
This behavior replicates the exception handling of the single-threaded maintenance task, where we throw the exception
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280125141
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStoreConf.scala:
##########
@@ -27,6 +27,12 @@ class StateStoreConf(
def this() = this(new SQLConf)
+ /**
+ * Size of MaintenanceThreadPool to perform maintenance tasks
+ * for StateStore
Review Comment:
Does this not fit on a single line above ?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] ericm-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "ericm-db (via GitHub)" <gi...@apache.org>.
ericm-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1281029564
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -636,17 +691,46 @@ object StateStore extends Logging {
if (SparkEnv.get == null) {
throw new IllegalStateException("SparkEnv not active, cannot do maintenance on StateStores")
}
- loadedProviders.synchronized { loadedProviders.toSeq }.foreach { case (id, provider) =>
- try {
- provider.doMaintenance()
- if (!verifyIfStoreInstanceActive(id)) {
- unload(id)
- logInfo(s"Unloaded $provider")
- }
- } catch {
- case NonFatal(e) =>
- logWarning(s"Error managing $provider, stopping management thread")
- throw e
+ loadedProviders.synchronized {
+ loadedProviders.toSeq
+ }.foreach { case (id, provider) =>
+ // check exception
+ if (threadPoolException.get() != null) {
+ val exception = threadPoolException.get()
+ logWarning("Error in maintenanceThreadPool", exception)
+ throw exception
+ }
+ if (processThisPartition(id)) {
+ maintenanceThreadPool.execute(() => {
+ val startTime = System.currentTimeMillis()
+ try {
+ provider.doMaintenance()
+ if (!verifyIfStoreInstanceActive(id)) {
+ unload(id)
+ logInfo(s"Unloaded $provider")
+ }
+ } catch {
+ case NonFatal(e) =>
+ logWarning(s"Error managing $provider, stopping management thread", e)
+ threadPoolException.set(e)
+ throw e
Review Comment:
Right yeah, I meant in terms of an exception being thrown during maintenance of the StateStore, the thread will set the exception, throw it, and then the main scheduling thread will see this exception is not null, and then throw it, stopping the StateStore
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] ericm-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "ericm-db (via GitHub)" <gi...@apache.org>.
ericm-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280125367
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -607,25 +650,42 @@ object StateStore extends Logging {
}
/** Start the periodic maintenance task if not already started and if Spark active */
- private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit =
+ private def startMaintenanceIfNeeded(storeConf: StateStoreConf): Unit = {
+ val numMaintenanceThreads = storeConf.numStateStoreMaintenanceThreads
loadedProviders.synchronized {
if (SparkEnv.get != null && !isMaintenanceRunning) {
maintenanceTask = new MaintenanceTask(
storeConf.maintenanceInterval,
- task = { doMaintenance() },
- onError = { loadedProviders.synchronized {
+ task = {
+ doMaintenance()
+ },
+ onError = {
+ loadedProviders.synchronized {
logInfo("Stopping maintenance task since an error was encountered.")
stopMaintenanceTask()
- // SPARK-44504 - Unload explicitly to force closing underlying DB instance
Review Comment:
Yeah I thought we wanted to remove it because it was essentially a TODO?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR commented on pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on PR #42066:
URL: https://github.com/apache/spark/pull/42066#issuecomment-1660982471
(I changed the description of the PR slightly to include the context for newly added config and default value.)
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] anishshri-db commented on a diff in pull request #42066: [SPARK-44480][SS] Use thread pool to perform maintenance activity for hdfs/rocksdb state store providers
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #42066:
URL: https://github.com/apache/spark/pull/42066#discussion_r1280178229
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala:
##########
@@ -472,9 +483,33 @@ object StateStore extends Logging {
def isRunning: Boolean = !future.isDone
}
+ /**
+ * Thread Pool that runs maintenance on partitions that are scheduled by
+ * MaintenanceTask periodically
+ */
+ class MaintenanceThreadPool(numThreads: Int) {
+ private val threadPool = ThreadUtils.newDaemonFixedThreadPool(
+ numThreads, "state-store-maintenance-thread")
+
+ def execute(runnable: Runnable): Unit = {
+ threadPool.execute(runnable)
+ }
+
+ def isRunning(): Boolean = {
Review Comment:
Where is this used ?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR commented on pull request #42066: [SPARK-44480][SS] Add option for thread pool to perform maintenance for RocksDB/HDFS State Store Providers
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on PR #42066:
URL: https://github.com/apache/spark/pull/42066#issuecomment-1641409258
@ericm-db I forgot you had a PR. Does this PR supersede the old PR? It looks like so, but just to double confirm.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org