You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "vicennial (via GitHub)" <gi...@apache.org> on 2024/01/10 20:20:10 UTC
[PR] [SPARK-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder [spark]
vicennial opened a new pull request, #44670:
URL: https://github.com/apache/spark/pull/44670
<!--
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.
-->
This PR makes `SparkConnectReattachExecuteHandler` fetch the `ExecuteHolder` via the`SessionHolder` which in turn refreshes it's aliveness. Further, this makes it consistent with `SparkConnectReattachExecuteHandler`.
### 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.
-->
Currently in ReattachExecute, we fetch the ExecuteHolder directly without going through the SessionHolder and hence the "aliveness" of the SessionHolder is not refreshed.
This would result in long-running queries (which do not send `ReleaseExecute` requests in specific) failing because the `SessionHolder` would expire from the cache during an active query execution.
### 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'.
-->
Yes, it fixes the bug where long-running may fail when their corresponding `SessionHolder` is expired during active query execution.
### 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.
-->
### 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.
-->
--
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-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder [spark]
Posted by "vicennial (via GitHub)" <gi...@apache.org>.
vicennial commented on code in PR #44670:
URL: https://github.com/apache/spark/pull/44670#discussion_r1448778094
##########
connector/connect/server/src/main/scala/org/apache/spark/sql/connect/service/SparkConnectReattachExecuteHandler.scala:
##########
@@ -29,25 +29,26 @@ class SparkConnectReattachExecuteHandler(
extends Logging {
def handle(v: proto.ReattachExecuteRequest): Unit = {
- val executeHolder = SparkConnectService.executionManager
- .getExecuteHolder(ExecuteKey(v.getUserContext.getUserId, v.getSessionId, v.getOperationId))
Review Comment:
Hmm, interestingly the python client tests are failing in `ReleaseExecute` https://github.com/vicennial/spark/actions/runs/7486495590/job/20377029638
```
/__w/spark/spark/python/pyspark/sql/connect/client/reattach.py:231: UserWarning: ReleaseExecute failed with exception: <_InactiveRpcError of RPC that terminated with:
status = StatusCode.INTERNAL
details = "[INVALID_HANDLE.SESSION_CLOSED] The handle da42ad8c-12d3-476f-8240-0ed36d34ebe4 is invalid. Session was closed. SQLSTATE: HY000"
debug_error_string = "UNKNOWN:Error received from peer {created_time:"2024-01-11T09:25:48.603606747+00:00", grpc_status:13, grpc_message:"[INVALID_HANDLE.SESSION_CLOSED] The handle da42ad8c-12d3-476f-8240-0ed36d34ebe4 is invalid. Session was closed. SQLSTATE: HY000"}"
>.
```
--
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-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder [spark]
Posted by "HyukjinKwon (via GitHub)" <gi...@apache.org>.
HyukjinKwon commented on PR #44670:
URL: https://github.com/apache/spark/pull/44670#issuecomment-1891563522
Merged 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
Re: [PR] [SPARK-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder [spark]
Posted by "juliuszsompolski (via GitHub)" <gi...@apache.org>.
juliuszsompolski commented on code in PR #44670:
URL: https://github.com/apache/spark/pull/44670#discussion_r1449134991
##########
python/pyspark/sql/tests/connect/client/test_reattach.py:
##########
@@ -84,10 +84,10 @@ def wait_for_responses():
# query1 and query2 could get either an:
# OPERATION_CANCELED if it happens fast - when closing the session interrupted the queries,
# and that error got pushed to the client buffers before the client got disconnected.
- # OPERATION_ABANDONED if it happens slow - when closing the session interrupted the client
- # RPCs before it pushed out the error above. The client would then get an
+ # INVALID_HANDLE.SESSION_CLOSED if it happens slow - when closing the session interrupted
+ # the client RPCs before it pushed out the error above. The client would then get an
# INVALID_CURSOR.DISCONNECTED, which it will retry with a ReattachExecute, and then get an
- # INVALID_HANDLE.OPERATION_ABANDONED.
+ # INVALID_HANDLE.SESSION_CLOSED.
Review Comment:
There is the same test case in https://github.com/apache/spark/blob/master/connector/connect/server/src/test/scala/org/apache/spark/sql/connect/service/SparkConnectServiceE2ESuite.scala#L66-L72
if it got flaky here, it probably can also be flaky there 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-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder [spark]
Posted by "vicennial (via GitHub)" <gi...@apache.org>.
vicennial commented on code in PR #44670:
URL: https://github.com/apache/spark/pull/44670#discussion_r1448774542
##########
connector/connect/server/src/main/scala/org/apache/spark/sql/connect/service/SessionHolder.scala:
##########
@@ -214,6 +214,9 @@ case class SessionHolder(userId: String, sessionId: String, session: SparkSessio
logInfo(s"Session $key accessed, time $lastAccessTimeMs.")
}
+ // Exposed for testing.
+ private[connect] def getAccessTime(): Long = lastAccessTimeMs
Review Comment:
Nice 👍
--
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-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder [spark]
Posted by "juliuszsompolski (via GitHub)" <gi...@apache.org>.
juliuszsompolski commented on code in PR #44670:
URL: https://github.com/apache/spark/pull/44670#discussion_r1448795792
##########
connector/connect/server/src/main/scala/org/apache/spark/sql/connect/service/SparkConnectReattachExecuteHandler.scala:
##########
@@ -29,25 +29,26 @@ class SparkConnectReattachExecuteHandler(
extends Logging {
def handle(v: proto.ReattachExecuteRequest): Unit = {
- val executeHolder = SparkConnectService.executionManager
- .getExecuteHolder(ExecuteKey(v.getUserContext.getUserId, v.getSessionId, v.getOperationId))
Review Comment:
hm, I'm aware of some rare issues with the scala client https://issues.apache.org/jira/browse/SPARK-45206, but python client should be fine with https://github.com/apache/spark/pull/42929#discussion_r1329071826 fix I believe...
--
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-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder [spark]
Posted by "HyukjinKwon (via GitHub)" <gi...@apache.org>.
HyukjinKwon closed pull request #44670: [SPARK-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder
URL: https://github.com/apache/spark/pull/44670
--
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-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder [spark]
Posted by "juliuszsompolski (via GitHub)" <gi...@apache.org>.
juliuszsompolski commented on PR #44670:
URL: https://github.com/apache/spark/pull/44670#issuecomment-1886846011
> Further, this makes it consistent with `SparkConnectReattachExecuteHandler`.
You mean ReleaseExecute handler?
--
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-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder [spark]
Posted by "vicennial (via GitHub)" <gi...@apache.org>.
vicennial commented on code in PR #44670:
URL: https://github.com/apache/spark/pull/44670#discussion_r1448781837
##########
connector/connect/server/src/main/scala/org/apache/spark/sql/connect/service/SparkConnectReattachExecuteHandler.scala:
##########
@@ -29,25 +29,26 @@ class SparkConnectReattachExecuteHandler(
extends Logging {
def handle(v: proto.ReattachExecuteRequest): Unit = {
- val executeHolder = SparkConnectService.executionManager
- .getExecuteHolder(ExecuteKey(v.getUserContext.getUserId, v.getSessionId, v.getOperationId))
- .getOrElse {
- if (SparkConnectService.executionManager
- .getAbandonedTombstone(
- ExecuteKey(v.getUserContext.getUserId, v.getSessionId, v.getOperationId))
- .isDefined) {
- logDebug(s"Reattach operation abandoned: ${v.getOperationId}")
- throw new SparkSQLException(
- errorClass = "INVALID_HANDLE.OPERATION_ABANDONED",
- messageParameters = Map("handle" -> v.getOperationId))
+ val sessionHolder = SparkConnectService.sessionManager
+ .getIsolatedSession(SessionKey(v.getUserContext.getUserId, v.getSessionId))
- } else {
- logDebug(s"Reattach operation not found: ${v.getOperationId}")
- throw new SparkSQLException(
- errorClass = "INVALID_HANDLE.OPERATION_NOT_FOUND",
- messageParameters = Map("handle" -> v.getOperationId))
- }
+ val executeHolder = sessionHolder.executeHolder(v.getOperationId).getOrElse {
+ if (SparkConnectService.executionManager
+ .getAbandonedTombstone(
+ ExecuteKey(v.getUserContext.getUserId, v.getSessionId, v.getOperationId))
+ .isDefined) {
+ logDebug(s"Reattach operation abandoned: ${v.getOperationId}")
+ throw new SparkSQLException(
+ errorClass = "INVALID_HANDLE.OPERATION_ABANDONED",
+ messageParameters = Map("handle" -> v.getOperationId))
+
Review Comment:
This is an existing empty line, its showing up due to whitespace changes. Will remove it anyway 👍
--
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-46660][CONNECT] ReattachExecute requests updates aliveness of SessionHolder [spark]
Posted by "juliuszsompolski (via GitHub)" <gi...@apache.org>.
juliuszsompolski commented on code in PR #44670:
URL: https://github.com/apache/spark/pull/44670#discussion_r1448658127
##########
connector/connect/server/src/main/scala/org/apache/spark/sql/connect/service/SessionHolder.scala:
##########
@@ -214,6 +214,9 @@ case class SessionHolder(userId: String, sessionId: String, session: SparkSessio
logInfo(s"Session $key accessed, time $lastAccessTimeMs.")
}
+ // Exposed for testing.
+ private[connect] def getAccessTime(): Long = lastAccessTimeMs
Review Comment:
nit: you can use `getSessionHolderInfo.lastAccessTimeMs`
##########
connector/connect/server/src/main/scala/org/apache/spark/sql/connect/service/SparkConnectReattachExecuteHandler.scala:
##########
@@ -29,25 +29,26 @@ class SparkConnectReattachExecuteHandler(
extends Logging {
def handle(v: proto.ReattachExecuteRequest): Unit = {
- val executeHolder = SparkConnectService.executionManager
- .getExecuteHolder(ExecuteKey(v.getUserContext.getUserId, v.getSessionId, v.getOperationId))
- .getOrElse {
- if (SparkConnectService.executionManager
- .getAbandonedTombstone(
- ExecuteKey(v.getUserContext.getUserId, v.getSessionId, v.getOperationId))
- .isDefined) {
- logDebug(s"Reattach operation abandoned: ${v.getOperationId}")
- throw new SparkSQLException(
- errorClass = "INVALID_HANDLE.OPERATION_ABANDONED",
- messageParameters = Map("handle" -> v.getOperationId))
+ val sessionHolder = SparkConnectService.sessionManager
+ .getIsolatedSession(SessionKey(v.getUserContext.getUserId, v.getSessionId))
- } else {
- logDebug(s"Reattach operation not found: ${v.getOperationId}")
- throw new SparkSQLException(
- errorClass = "INVALID_HANDLE.OPERATION_NOT_FOUND",
- messageParameters = Map("handle" -> v.getOperationId))
- }
+ val executeHolder = sessionHolder.executeHolder(v.getOperationId).getOrElse {
+ if (SparkConnectService.executionManager
+ .getAbandonedTombstone(
+ ExecuteKey(v.getUserContext.getUserId, v.getSessionId, v.getOperationId))
+ .isDefined) {
+ logDebug(s"Reattach operation abandoned: ${v.getOperationId}")
+ throw new SparkSQLException(
+ errorClass = "INVALID_HANDLE.OPERATION_ABANDONED",
+ messageParameters = Map("handle" -> v.getOperationId))
+
Review Comment:
nit: no empty line.
##########
connector/connect/server/src/main/scala/org/apache/spark/sql/connect/service/SparkConnectReattachExecuteHandler.scala:
##########
@@ -29,25 +29,26 @@ class SparkConnectReattachExecuteHandler(
extends Logging {
def handle(v: proto.ReattachExecuteRequest): Unit = {
- val executeHolder = SparkConnectService.executionManager
- .getExecuteHolder(ExecuteKey(v.getUserContext.getUserId, v.getSessionId, v.getOperationId))
Review Comment:
I reviewed and traced the code to see if this change would introduce any strange range conditions because SparkConnectService.executionManager.getExecuteHolder was a one step process under executionsLock, and getIsolatedSession + sessionHolder.executeHolder is a two step process, but I believe all is good.
Even if the session/execution closed in the meanwhile, then `executeHolder.runGrpcResponseSender` -> `addGrpcResponseSender` would handle detecting that the execution was concurrently closed.
--
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