You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "vinodkc (via GitHub)" <gi...@apache.org> on 2025/04/01 23:47:53 UTC
[PR] [SPARK-51685][SS] Excessive Info logging from RocksDb operations casing too big executor stderr files [spark]
vinodkc opened a new pull request, #50483:
URL: https://github.com/apache/spark/pull/50483
<!--
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.
-->
Long-running structured streaming applications with RocksDb statestore is failing after some time due to "No space left on device" error.
Checked the executor logs and noticed that the volume of executor logs is crazy. Around 1GB of data for 1 minute of run.
Each info log entry from `RocksDBFileManager` and `RocksDB` classes were printing more than 4000 lines
### 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.
-->
Info logs printing all checkpoint file paths. These verbos info logs affects streaming application disk space as most users keep Info as the default log level.
So it is better to use Debug log to log such details.
### 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'.
-->
No, some verbos info level log entries changed to debug level
### 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.
-->
Manaully verified the executor logs using info and debug log levels
### 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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on PR #50483:
URL: https://github.com/apache/spark/pull/50483#issuecomment-2776853829
@HeartSaVioR - PTAL, thanks !
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "Kimahriman (via GitHub)" <gi...@apache.org>.
Kimahriman commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2025474678
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDB.scala:
##########
@@ -1485,11 +1485,12 @@ class RocksDB(
override def log(infoLogLevel: InfoLogLevel, logMsg: String) = {
// Map DB log level to log4j levels
// Warn is mapped to info because RocksDB warn is too verbose
+ // Info is mapped to debug because RocksDB info is too verbose
Review Comment:
Particularly because it appears to spit out the statistics at some regular interval for every opened state store instance on the executor, which can accumulate to a lot and then drown out the logs of what the executor is actually doing
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027590088
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -290,8 +290,8 @@ class RocksDBFileManager(
colFamilyIdMapping, colFamilyTypeMapping, maxColumnFamilyId)
val metadataFile = localMetadataFile(checkpointDir)
metadata.writeToFile(metadataFile)
- logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}:\n" +
- log"${MDC(LogKeys.METADATA_JSON, metadata.prettyJson)}")
+ logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}")
Review Comment:
Done
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027610531
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -290,8 +290,8 @@ class RocksDBFileManager(
colFamilyIdMapping, colFamilyTypeMapping, maxColumnFamilyId)
val metadataFile = localMetadataFile(checkpointDir)
metadata.writeToFile(metadataFile)
- logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}:\n" +
- log"${MDC(LogKeys.METADATA_JSON, metadata.prettyJson)}")
+ logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}")
Review Comment:
Done
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "siying (via GitHub)" <gi...@apache.org>.
siying commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2035895751
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDB.scala:
##########
@@ -1485,11 +1485,12 @@ class RocksDB(
override def log(infoLogLevel: InfoLogLevel, logMsg: String) = {
// Map DB log level to log4j levels
// Warn is mapped to info because RocksDB warn is too verbose
+ // Info is mapped to debug because RocksDB info is too verbose
Review Comment:
@anishshri-db sorry for the late respond. Stats dump period can be adjusted using https://github.com/facebook/rocksdb/blob/6d802639f7dc35bf765dbe1ed6b3942e4d76375d/java/src/main/java/org/rocksdb/Options.java#L764
The default is 10 minutes, which perhaps is indeed too frequent.. I think we can easily adjust to one hour.
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2025256690
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -843,14 +843,16 @@ class RocksDBFileManager(
}
zout.close() // so that any error in closing also cancels the output stream
logInfo(log"Zipped ${MDC(LogKeys.NUM_BYTES, totalBytes)} bytes (before compression) to " +
- log"${MDC(LogKeys.FILE_NAME, filesStr)}")
+ log"${MDC(LogKeys.FILE_NAME, dfsZipFile)}")
+ logDebug(log"Files are:\n\t${MDC(LogKeys.FILE_NAME, fileListStr)}")
Review Comment:
Move whole line to debug
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -865,7 +867,8 @@ class RocksDBFileManager(
lazy val files = Option(Utils.recursiveList(dir)).getOrElse(Array.empty).map { f =>
s"${f.getAbsolutePath} - ${f.length()} bytes"
}
- logInfo(msg + log" - ${MDC(LogKeys.NUM_FILES, files.length)} files\n\t" +
+ logInfo(msg + log" - ${MDC(LogKeys.NUM_FILES, files.length)} files")
Review Comment:
Same here
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2025256245
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -290,8 +290,8 @@ class RocksDBFileManager(
colFamilyIdMapping, colFamilyTypeMapping, maxColumnFamilyId)
val metadataFile = localMetadataFile(checkpointDir)
metadata.writeToFile(metadataFile)
- logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}:\n" +
- log"${MDC(LogKeys.METADATA_JSON, metadata.prettyJson)}")
+ logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}")
Review Comment:
Is this really an issue ? How often did you see 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
Re: [PR] [SPARK-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "siying (via GitHub)" <gi...@apache.org>.
siying commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2035895751
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDB.scala:
##########
@@ -1485,11 +1485,12 @@ class RocksDB(
override def log(infoLogLevel: InfoLogLevel, logMsg: String) = {
// Map DB log level to log4j levels
// Warn is mapped to info because RocksDB warn is too verbose
+ // Info is mapped to debug because RocksDB info is too verbose
Review Comment:
@anishshri-db sorry for the late respond. Stats period can be adjusted using https://github.com/facebook/rocksdb/blob/6d802639f7dc35bf765dbe1ed6b3942e4d76375d/java/src/main/java/org/rocksdb/Options.java#L764
The default is 10 minutes, which perhaps is indeed too frequent.. I think we can easily adjust to one hour.
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2025478460
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDB.scala:
##########
@@ -1485,11 +1485,12 @@ class RocksDB(
override def log(infoLogLevel: InfoLogLevel, logMsg: String) = {
// Map DB log level to log4j levels
// Warn is mapped to info because RocksDB warn is too verbose
+ // Info is mapped to debug because RocksDB info is too verbose
Review Comment:
@siying - is there some way to disable the periodic stats reporting in RocksDB ?
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027611246
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -340,8 +340,8 @@ class RocksDBFileManager(
// Copy the necessary immutable files
val metadataFile = localMetadataFile(localDir)
val metadata = RocksDBCheckpointMetadata.readFromFile(metadataFile)
- logInfo(log"Read metadata for version ${MDC(LogKeys.VERSION_NUM, version)}:\n" +
- log"${MDC(LogKeys.METADATA_JSON, metadata.prettyJson)}")
+ logInfo(log"Read metadata for version ${MDC(LogKeys.VERSION_NUM, version)}")
Review Comment:
Changed to Debug log
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2025257321
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -340,8 +340,8 @@ class RocksDBFileManager(
// Copy the necessary immutable files
val metadataFile = localMetadataFile(localDir)
val metadata = RocksDBCheckpointMetadata.readFromFile(metadataFile)
- logInfo(log"Read metadata for version ${MDC(LogKeys.VERSION_NUM, version)}:\n" +
- log"${MDC(LogKeys.METADATA_JSON, metadata.prettyJson)}")
+ logInfo(log"Read metadata for version ${MDC(LogKeys.VERSION_NUM, version)}")
Review Comment:
Is this really an issue ?
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027584730
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -290,8 +290,8 @@ class RocksDBFileManager(
colFamilyIdMapping, colFamilyTypeMapping, maxColumnFamilyId)
val metadataFile = localMetadataFile(checkpointDir)
metadata.writeToFile(metadataFile)
- logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}:\n" +
- log"${MDC(LogKeys.METADATA_JSON, metadata.prettyJson)}")
+ logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}")
Review Comment:
@vinodkc - if this is not very verbose, lets revert to original and same for the line below ?
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027590088
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -290,8 +290,8 @@ class RocksDBFileManager(
colFamilyIdMapping, colFamilyTypeMapping, maxColumnFamilyId)
val metadataFile = localMetadataFile(checkpointDir)
metadata.writeToFile(metadataFile)
- logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}:\n" +
- log"${MDC(LogKeys.METADATA_JSON, metadata.prettyJson)}")
+ logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}")
Review Comment:
Done
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027591760
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -340,8 +340,8 @@ class RocksDBFileManager(
// Copy the necessary immutable files
val metadataFile = localMetadataFile(localDir)
val metadata = RocksDBCheckpointMetadata.readFromFile(metadataFile)
- logInfo(log"Read metadata for version ${MDC(LogKeys.VERSION_NUM, version)}:\n" +
- log"${MDC(LogKeys.METADATA_JSON, metadata.prettyJson)}")
+ logInfo(log"Read metadata for version ${MDC(LogKeys.VERSION_NUM, version)}")
Review Comment:
Yes, sometimes it was printing the content of metadata.prettyJson, which is verbose
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2025926556
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDB.scala:
##########
@@ -1485,11 +1485,12 @@ class RocksDB(
override def log(infoLogLevel: InfoLogLevel, logMsg: String) = {
// Map DB log level to log4j levels
// Warn is mapped to info because RocksDB warn is too verbose
+ // Info is mapped to debug because RocksDB info is too verbose
Review Comment:
@anishshri-db ,
>Which specific logs did we see most often here ?
Mainly `Hard Linking <filename>.log`
eg: 25/03/19 10:38:55 INFO RocksDB [StateStoreId(opId=0,partId=265,name=default), queryRunId=30d232ab-0d2a-454d-a217-5b2debfe10e9, queryId=4c8ca990-9fd4-4b69-8f90-ce933b3f4ef3]: [NativeRocksDB-1] [ilities/checkpoint/checkpoint_impl.cc:127] Hard Linking 009723.log
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2025258480
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDB.scala:
##########
@@ -1485,11 +1485,12 @@ class RocksDB(
override def log(infoLogLevel: InfoLogLevel, logMsg: String) = {
// Map DB log level to log4j levels
// Warn is mapped to info because RocksDB warn is too verbose
+ // Info is mapped to debug because RocksDB info is too verbose
Review Comment:
Some RocksDB info logs have actually been useful for us for debugging. Which specific logs did we see most often here ?
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027595910
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -290,8 +290,8 @@ class RocksDBFileManager(
colFamilyIdMapping, colFamilyTypeMapping, maxColumnFamilyId)
val metadataFile = localMetadataFile(checkpointDir)
metadata.writeToFile(metadataFile)
- logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}:\n" +
- log"${MDC(LogKeys.METADATA_JSON, metadata.prettyJson)}")
+ logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}")
Review Comment:
It was printing the content of metadata.prettyJson, which is verbose,
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027597377
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -290,8 +290,8 @@ class RocksDBFileManager(
colFamilyIdMapping, colFamilyTypeMapping, maxColumnFamilyId)
val metadataFile = localMetadataFile(checkpointDir)
metadata.writeToFile(metadataFile)
- logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}:\n" +
- log"${MDC(LogKeys.METADATA_JSON, metadata.prettyJson)}")
+ logInfo(log"Written metadata for version ${MDC(LogKeys.VERSION_NUM, version)}")
Review Comment:
Ok then lets not split the line and move to DEBUG
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027589200
##########
core/src/main/scala/org/apache/spark/util/Utils.scala:
##########
@@ -3085,7 +3085,8 @@ private[spark] object Utils
entry = in.getNextEntry()
}
in.close() // so that any error in closing does not get ignored
- logInfo(log"Unzipped from ${MDC(PATH, dfsZipFile)}\n\t${MDC(PATHS, files.mkString("\n\t"))}")
+ logInfo(log"Unzipped from ${MDC(PATH, dfsZipFile)}")
Review Comment:
Done
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "Kimahriman (via GitHub)" <gi...@apache.org>.
Kimahriman commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2025291601
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDB.scala:
##########
@@ -1485,11 +1485,12 @@ class RocksDB(
override def log(infoLogLevel: InfoLogLevel, logMsg: String) = {
// Map DB log level to log4j levels
// Warn is mapped to info because RocksDB warn is too verbose
+ // Info is mapped to debug because RocksDB info is too verbose
Review Comment:
This part is the biggest pain for us and I've been trying to find ways to filter these out. All the statistics are incredibly verbose, and when you are running with multiple cores per executor and they all just get intermingled it's basically impossible to find anything useful. +1 to this being a thing to opt-in to debug logs for debugging purposes
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027590852
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -843,14 +843,16 @@ class RocksDBFileManager(
}
zout.close() // so that any error in closing also cancels the output stream
logInfo(log"Zipped ${MDC(LogKeys.NUM_BYTES, totalBytes)} bytes (before compression) to " +
- log"${MDC(LogKeys.FILE_NAME, filesStr)}")
+ log"${MDC(LogKeys.FILE_NAME, dfsZipFile)}")
+ logDebug(log"Files are:\n\t${MDC(LogKeys.FILE_NAME, fileListStr)}")
Review Comment:
Done
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2027590367
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDBFileManager.scala:
##########
@@ -865,7 +867,8 @@ class RocksDBFileManager(
lazy val files = Option(Utils.recursiveList(dir)).getOrElse(Array.empty).map { f =>
s"${f.getAbsolutePath} - ${f.length()} bytes"
}
- logInfo(msg + log" - ${MDC(LogKeys.NUM_FILES, files.length)} files\n\t" +
+ logInfo(msg + log" - ${MDC(LogKeys.NUM_FILES, files.length)} files")
Review Comment:
Done
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR closed pull request #50483: [SPARK-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files
URL: https://github.com/apache/spark/pull/50483
--
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-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on PR #50483:
URL: https://github.com/apache/spark/pull/50483#issuecomment-2777330340
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
Re: [PR] [SPARK-51685][SS] Excessive Info logging from RocksDb operations causing too big executor stderr files [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #50483:
URL: https://github.com/apache/spark/pull/50483#discussion_r2025255349
##########
core/src/main/scala/org/apache/spark/util/Utils.scala:
##########
@@ -3085,7 +3085,8 @@ private[spark] object Utils
entry = in.getNextEntry()
}
in.close() // so that any error in closing does not get ignored
- logInfo(log"Unzipped from ${MDC(PATH, dfsZipFile)}\n\t${MDC(PATHS, files.mkString("\n\t"))}")
+ logInfo(log"Unzipped from ${MDC(PATH, dfsZipFile)}")
Review Comment:
Lets just move this line to `DEBUG`. We can't corelate easily if we split the lines
--
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