You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "viirya (via GitHub)" <gi...@apache.org> on 2025/12/05 23:29:03 UTC
[PR] [SPARK-XXXXX][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
viirya opened a new pull request, #53358:
URL: https://github.com/apache/spark/pull/53358
<!--
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
'common/utils/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.
-->
This patch adds a safety check into `ObservationManager.tryComplete` to avoid Observation blocking.
### 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.
-->
We got reports that for some corner cases `Observation.get` will be blocked forever. It is not deadlock case after investigation. If the `CollectMetricsExec` operator was optimized away, e.g., the executed plan was optimized to have some empty relation propagation on top of plan tree of `CollectMetricsExec`, Spark won't fulfill the promise in `Observation` and `get` calls will be blocked forever.
### Does this PR introduce _any_ user-facing change?
<!--
Note that it means *any* user-facing change including all aspects such as new features, bug fixes, or other behavior changes. Documentation-only updates are not considered user-facing changes.
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'.
-->
Existing tests
### 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.
-->
No
### Was this patch authored or co-authored using generative AI tooling?
<!--
If generative AI tooling has been used in the process of authoring this patch, please include the
phrase: 'Generated-by: ' followed by the name of the tool and its version.
If no, write 'No'.
Please refer to the [ASF Generative Tooling Guidance](https://www.apache.org/legal/generative-tooling.html) for details.
-->
No
--
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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "viirya (via GitHub)" <gi...@apache.org>.
viirya commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594427339
##########
sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala:
##########
@@ -2878,6 +2878,24 @@ class DatasetSuite extends QueryTest
checkDataset(Seq(seqMutableSet).toDS(), seqMutableSet)
checkDataset(Seq(mapMutableSet).toDS(), mapMutableSet)
}
+
+ test("SPARK-54620: Observation should not blocking forever") {
+ val observation = Observation("row_count")
+
+ var df = Seq.empty[(Int, Int)].toDF("v1", "v2")
+ df = df.observe(observation,
+ functions.count(functions.lit(1)).alias("record_cnt"))
+ df = df.repartition($"v1")
+ .select($"v1" + 1 as "v1", $"v2" + 1 as "v2")
+ .join(
+ Seq((1, 2), (3, 4)).toDF("v1", "v2").repartition($"v2"),
+ Seq("v1"),
+ "inner")
+ df.collect()
+
+ val metrics = observation.get
Review Comment:
For now, I choose to not change empty relation propagation logic but let Spark return empty map.
--
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
Re: [PR] [SPARK-XXXXX][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "dongjoon-hyun (via GitHub)" <gi...@apache.org>.
dongjoon-hyun commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594230178
##########
sql/api/src/main/scala/org/apache/spark/sql/Observation.scala:
##########
@@ -78,7 +78,11 @@ class Observation(val name: String) {
@throws[InterruptedException]
def get: Map[String, Any] = {
val row = getRow
- row.getValuesMap(row.schema.map(_.name))
+ if (row.schema == null) {
Review Comment:
1. We still wait indefinitely via `Duration.Inf`, don't we?
https://github.com/apache/spark/blob/4baf94d83f510c5a095a4383e7f53e4bd488fedb/sql/api/src/main/scala/org/apache/spark/sql/Observation.scala#L132-L134
2. When `row != null` and `row.schema == null`?
--
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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "viirya (via GitHub)" <gi...@apache.org>.
viirya commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594426128
##########
sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala:
##########
@@ -2878,6 +2878,24 @@ class DatasetSuite extends QueryTest
checkDataset(Seq(seqMutableSet).toDS(), seqMutableSet)
checkDataset(Seq(mapMutableSet).toDS(), mapMutableSet)
}
+
+ test("SPARK-54620: Observation should not blocking forever") {
+ val observation = Observation("row_count")
+
+ var df = Seq.empty[(Int, Int)].toDF("v1", "v2")
+ df = df.observe(observation,
+ functions.count(functions.lit(1)).alias("record_cnt"))
+ df = df.repartition($"v1")
+ .select($"v1" + 1 as "v1", $"v2" + 1 as "v2")
+ .join(
+ Seq((1, 2), (3, 4)).toDF("v1", "v2").repartition($"v2"),
+ Seq("v1"),
+ "inner")
+ df.collect()
+
+ val metrics = observation.get
Review Comment:
This call is blocked forever for now as Spark adds `EmptyRelation` on the AQE plan. So alternatively, we can tweak `PropagateEmptyRelationBase` and `AQEPropagateEmptyRelation` to unblock it. But it also means we miss some more cases we can do short-circuit evaluation.
--
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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "viirya (via GitHub)" <gi...@apache.org>.
viirya commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594434400
##########
sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala:
##########
@@ -2878,6 +2878,24 @@ class DatasetSuite extends QueryTest
checkDataset(Seq(seqMutableSet).toDS(), seqMutableSet)
checkDataset(Seq(mapMutableSet).toDS(), mapMutableSet)
}
+
+ test("SPARK-54620: Observation should not blocking forever") {
+ val observation = Observation("row_count")
+
+ var df = Seq.empty[(Int, Int)].toDF("v1", "v2")
+ df = df.observe(observation,
+ functions.count(functions.lit(1)).alias("record_cnt"))
+ df = df.repartition($"v1")
+ .select($"v1" + 1 as "v1", $"v2" + 1 as "v2")
+ .join(
+ Seq((1, 2), (3, 4)).toDF("v1", "v2").repartition($"v2"),
+ Seq("v1"),
+ "inner")
+ df.collect()
+
+ val metrics = observation.get
Review Comment:
For example, this is the comparison between two query plans:
```
== Physical Plan ==
AdaptiveSparkPlan isFinalPlan=true
+- == Final Plan ==
ResultQueryStage 2
+- EmptyRelation [plan_id=10867]
+- Project [v1#12199, v2#12200, v2#12211]
+- EmptyRelation Join Inner, (v1#12199 = v1#12210)
```
```
== Physical Plan ==
AdaptiveSparkPlan isFinalPlan=true
+- == Final Plan ==
ResultQueryStage 3
+- *(3) Project [v1#12199, v2#12200, v2#12211]
+- *(3) BroadcastHashJoin [v1#12199], [v1#12210], Inner, BuildLeft, false
:- BroadcastQueryStage 2
: +- BroadcastExchange HashedRelationBroadcastMode(List(cast(input[0, int, false] as bigint)),false), [plan_id=10929]
: +- *(2) Project [(v1#12195 + 1) AS v1#12199, (v2#12196 + 1) AS v2#12200]
: +- AQEShuffleRead coalesced
: +- ShuffleQueryStage 0
: +- Exchange hashpartitioning(v1#12195, 5), REPARTITION_BY_COL, [plan_id=10868]
: +- CollectMetrics row_count, [count(1) AS record_cnt#12197L]
: +- LocalTableScan <empty>, [v1#12195, v2#12196]
+- AQEShuffleRead coalesced
+- ShuffleQueryStage 1
+- Exchange hashpartitioning(v2#12211, 5), REPARTITION_BY_COL, [plan_id=10898]
+- *(1) Project [_1#12203 AS v1#12210, _2#12204 AS v2#12211]
+- *(1) LocalTableScan [_1#12203, _2#12204]
```
--
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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "dongjoon-hyun (via GitHub)" <gi...@apache.org>.
dongjoon-hyun closed pull request #53358: [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking
URL: https://github.com/apache/spark/pull/53358
--
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
Re: [PR] [SPARK-XXXXX][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "viirya (via GitHub)" <gi...@apache.org>.
viirya commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594267597
##########
sql/api/src/main/scala/org/apache/spark/sql/Observation.scala:
##########
@@ -78,7 +78,11 @@ class Observation(val name: String) {
@throws[InterruptedException]
def get: Map[String, Any] = {
val row = getRow
- row.getValuesMap(row.schema.map(_.name))
+ if (row.schema == null) {
Review Comment:
> 1. We still wait indefinitely via `Duration.Inf`, don't we?
No, since we update it with empty row by `setMetricsAndNotify`.
> 2. When `row != null` and `row.schema == null`?
Okay.
--
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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "viirya (via GitHub)" <gi...@apache.org>.
viirya commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594427339
##########
sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala:
##########
@@ -2878,6 +2878,24 @@ class DatasetSuite extends QueryTest
checkDataset(Seq(seqMutableSet).toDS(), seqMutableSet)
checkDataset(Seq(mapMutableSet).toDS(), mapMutableSet)
}
+
+ test("SPARK-54620: Observation should not blocking forever") {
+ val observation = Observation("row_count")
+
+ var df = Seq.empty[(Int, Int)].toDF("v1", "v2")
+ df = df.observe(observation,
+ functions.count(functions.lit(1)).alias("record_cnt"))
+ df = df.repartition($"v1")
+ .select($"v1" + 1 as "v1", $"v2" + 1 as "v2")
+ .join(
+ Seq((1, 2), (3, 4)).toDF("v1", "v2").repartition($"v2"),
+ Seq("v1"),
+ "inner")
+ df.collect()
+
+ val metrics = observation.get
Review Comment:
This is a trade-off. For now, I choose to not change empty relation propagation logic but let Spark return empty map. Because users can easily check the metric key on the returned map to know if there is collected metric or not. It also doesn't change current working behavior because these cases will block `Observation.get` call forever.
--
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
Re: [PR] [SPARK-XXXXX][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "dongjoon-hyun (via GitHub)" <gi...@apache.org>.
dongjoon-hyun commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594230178
##########
sql/api/src/main/scala/org/apache/spark/sql/Observation.scala:
##########
@@ -78,7 +78,11 @@ class Observation(val name: String) {
@throws[InterruptedException]
def get: Map[String, Any] = {
val row = getRow
- row.getValuesMap(row.schema.map(_.name))
+ if (row.schema == null) {
Review Comment:
1. We still wait indefinitely via `Duration.Inf`, isn't it?
https://github.com/apache/spark/blob/4baf94d83f510c5a095a4383e7f53e4bd488fedb/sql/api/src/main/scala/org/apache/spark/sql/Observation.scala#L132-L134
2. When `row != null` and `row.schema == null`?
--
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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "viirya (via GitHub)" <gi...@apache.org>.
viirya commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594427339
##########
sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala:
##########
@@ -2878,6 +2878,24 @@ class DatasetSuite extends QueryTest
checkDataset(Seq(seqMutableSet).toDS(), seqMutableSet)
checkDataset(Seq(mapMutableSet).toDS(), mapMutableSet)
}
+
+ test("SPARK-54620: Observation should not blocking forever") {
+ val observation = Observation("row_count")
+
+ var df = Seq.empty[(Int, Int)].toDF("v1", "v2")
+ df = df.observe(observation,
+ functions.count(functions.lit(1)).alias("record_cnt"))
+ df = df.repartition($"v1")
+ .select($"v1" + 1 as "v1", $"v2" + 1 as "v2")
+ .join(
+ Seq((1, 2), (3, 4)).toDF("v1", "v2").repartition($"v2"),
+ Seq("v1"),
+ "inner")
+ df.collect()
+
+ val metrics = observation.get
Review Comment:
This is a trade-off. For now, I choose to not change empty relation propagation logic but let Spark return empty map. Because users can easily check the metric key on the returned map to know if there is collected metric or not.
--
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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "viirya (via GitHub)" <gi...@apache.org>.
viirya commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594426128
##########
sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala:
##########
@@ -2878,6 +2878,24 @@ class DatasetSuite extends QueryTest
checkDataset(Seq(seqMutableSet).toDS(), seqMutableSet)
checkDataset(Seq(mapMutableSet).toDS(), mapMutableSet)
}
+
+ test("SPARK-54620: Observation should not blocking forever") {
+ val observation = Observation("row_count")
+
+ var df = Seq.empty[(Int, Int)].toDF("v1", "v2")
+ df = df.observe(observation,
+ functions.count(functions.lit(1)).alias("record_cnt"))
+ df = df.repartition($"v1")
+ .select($"v1" + 1 as "v1", $"v2" + 1 as "v2")
+ .join(
+ Seq((1, 2), (3, 4)).toDF("v1", "v2").repartition($"v2"),
+ Seq("v1"),
+ "inner")
+ df.collect()
+
+ val metrics = observation.get
Review Comment:
This call is blocked forever for now as Spark adds `EmptyRelation` on the AQE plan. So alternatively, we can tweak `PropagateEmptyRelationBase` and `AQEPropagateEmptyRelation` to unblock it. But it also means we reduce some more cases we can do short-circuit evaluation.
--
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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "viirya (via GitHub)" <gi...@apache.org>.
viirya commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594427339
##########
sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala:
##########
@@ -2878,6 +2878,24 @@ class DatasetSuite extends QueryTest
checkDataset(Seq(seqMutableSet).toDS(), seqMutableSet)
checkDataset(Seq(mapMutableSet).toDS(), mapMutableSet)
}
+
+ test("SPARK-54620: Observation should not blocking forever") {
+ val observation = Observation("row_count")
+
+ var df = Seq.empty[(Int, Int)].toDF("v1", "v2")
+ df = df.observe(observation,
+ functions.count(functions.lit(1)).alias("record_cnt"))
+ df = df.repartition($"v1")
+ .select($"v1" + 1 as "v1", $"v2" + 1 as "v2")
+ .join(
+ Seq((1, 2), (3, 4)).toDF("v1", "v2").repartition($"v2"),
+ Seq("v1"),
+ "inner")
+ df.collect()
+
+ val metrics = observation.get
Review Comment:
This is a trade-off. For now, I choose to not change empty relation propagation logic but let Spark return empty map. Because users can easily check the metric key on the returned map to know if there is collected metric or not. It also doesn't change current working behavior but only fixing blocking cases.
--
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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "dongjoon-hyun (via GitHub)" <gi...@apache.org>.
dongjoon-hyun commented on PR #53358:
URL: https://github.com/apache/spark/pull/53358#issuecomment-3620585062
Merged to master/4.1 for Apache Spark 4.1.0 RC3 for 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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "cloud-fan (via GitHub)" <gi...@apache.org>.
cloud-fan commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2610449979
##########
sql/core/src/main/scala/org/apache/spark/sql/classic/ObservationManager.scala:
##########
@@ -56,10 +56,22 @@ private[sql] class ObservationManager(session: SparkSession) {
val allMetrics = qe.observedMetrics
qe.logical.foreach {
case c: CollectMetrics =>
- allMetrics.get(c.name).foreach { metrics =>
+ val keyExists = observations.containsKey((c.name, c.dataframeId))
+ val metrics = allMetrics.get(c.name)
+ if (keyExists && metrics.isEmpty) {
+ // If the key exists but no metrics were collected, it means for some reason the metrics
+ // could not be collected. This can happen e.g., if the CollectMetricsExec was optimized
+ // away.
val observation = observations.remove((c.name, c.dataframeId))
if (observation != null) {
- observation.setMetricsAndNotify(metrics)
+ observation.setMetricsAndNotify(Row.empty)
+ }
+ } else {
+ metrics.foreach { metrics =>
Review Comment:
`metrics.foreach { metric`
##########
sql/core/src/main/scala/org/apache/spark/sql/classic/ObservationManager.scala:
##########
@@ -56,10 +56,22 @@ private[sql] class ObservationManager(session: SparkSession) {
val allMetrics = qe.observedMetrics
qe.logical.foreach {
case c: CollectMetrics =>
- allMetrics.get(c.name).foreach { metrics =>
+ val keyExists = observations.containsKey((c.name, c.dataframeId))
+ val metrics = allMetrics.get(c.name)
+ if (keyExists && metrics.isEmpty) {
+ // If the key exists but no metrics were collected, it means for some reason the metrics
+ // could not be collected. This can happen e.g., if the CollectMetricsExec was optimized
+ // away.
val observation = observations.remove((c.name, c.dataframeId))
if (observation != null) {
- observation.setMetricsAndNotify(metrics)
+ observation.setMetricsAndNotify(Row.empty)
+ }
+ } else {
+ metrics.foreach { metrics =>
Review Comment:
`metrics.foreach { metric =>`
--
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
Re: [PR] [SPARK-54620][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "viirya (via GitHub)" <gi...@apache.org>.
viirya commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594427339
##########
sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala:
##########
@@ -2878,6 +2878,24 @@ class DatasetSuite extends QueryTest
checkDataset(Seq(seqMutableSet).toDS(), seqMutableSet)
checkDataset(Seq(mapMutableSet).toDS(), mapMutableSet)
}
+
+ test("SPARK-54620: Observation should not blocking forever") {
+ val observation = Observation("row_count")
+
+ var df = Seq.empty[(Int, Int)].toDF("v1", "v2")
+ df = df.observe(observation,
+ functions.count(functions.lit(1)).alias("record_cnt"))
+ df = df.repartition($"v1")
+ .select($"v1" + 1 as "v1", $"v2" + 1 as "v2")
+ .join(
+ Seq((1, 2), (3, 4)).toDF("v1", "v2").repartition($"v2"),
+ Seq("v1"),
+ "inner")
+ df.collect()
+
+ val metrics = observation.get
Review Comment:
This is a trade-off. For now, I choose to not change empty relation propagation logic but let Spark return empty map.
--
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
Re: [PR] [SPARK-XXXXX][SQL] Add safety check in ObservationManager to avoid Observation blocking [spark]
Posted by "viirya (via GitHub)" <gi...@apache.org>.
viirya commented on code in PR #53358:
URL: https://github.com/apache/spark/pull/53358#discussion_r2594268984
##########
sql/api/src/main/scala/org/apache/spark/sql/Observation.scala:
##########
@@ -78,7 +78,11 @@ class Observation(val name: String) {
@throws[InterruptedException]
def get: Map[String, Any] = {
val row = getRow
- row.getValuesMap(row.schema.map(_.name))
+ if (row.schema == null) {
Review Comment:
Oh, I updated to `row == null || row.schema == null`.
--
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