-
Notifications
You must be signed in to change notification settings - Fork 13.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[FLINK-15145][config] Change TM memory configuration default values for FLIP-49. #10860
[FLINK-15145][config] Change TM memory configuration default values for FLIP-49. #10860
Conversation
Thanks a lot for your contribution to the Apache Flink project. I'm the @flinkbot. I help the community Automated ChecksLast check on commit 5f9cd22 (Wed Jan 15 09:40:04 UTC 2020) Warnings:
Mention the bot in a comment to re-run the automated checks. Review Progress
Please see the Pull Request Review Guide for a full explanation of the review process. The Bot is tracking the review progress through labels. Labels are applied according to the order of the review items. For consensus, approval by a Flink committer of PMC member is required Bot commandsThe @flinkbot bot supports the following commands:
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the PR @xintongsong
It looks good to me, just added suggestions for a bit more details in comments
and reorganising process and Flink memory sizes.
EDIT: we probably want to mention the legacy option taskmanager.heap.size
only in migration guide.
# The total process memory size for the TaskManager. | ||
# | ||
# Note this accounts for all memory usages of a TaskManager process, including JVM metaspace and other overheads. | ||
# To exclude JVM metaspace and other overheads, please use total flink memory size (taskmanager.memory.flink.size) instead. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
# To exclude JVM metaspace and other overheads, please use total flink memory size (taskmanager.memory.flink.size) instead. |
@@ -42,9 +42,12 @@ jobmanager.rpc.port: 6123 | |||
jobmanager.heap.size: 1024m | |||
|
|||
|
|||
# The heap size for the TaskManager JVM | |||
# The total process memory size for the TaskManager. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
# The total process memory size for the TaskManager. | |
# The total process memory size for the TaskExecutor. |
I think this the most recent term for task manager after FLIP-6
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In the code base it is correct that we call it TaskExecutor
. However, we never fully transitioned from TaskManager
to TaskExecutor
in our documentation. Given that the memory config options are prefixed with taskmanager
, I would suggest to stick with TaskManager
for the time being.
# The heap size for the TaskManager JVM | ||
# The total process memory size for the TaskManager. | ||
# | ||
# Note this accounts for all memory usages of a TaskManager process, including JVM metaspace and other overheads. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
# Note this accounts for all memory usages of a TaskManager process, including JVM metaspace and other overheads. | |
# Note this accounts for all memory usage within the TaskExecutor process, including JVM metaspace and other overhead. | |
taskmanager.memory.process.size: 1568m |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Same here with TaskExecutor
-> TaskManager
.
…in default flink-conf.yaml. - Correct the description of process memory size. - Provide the alternative flink memory size.
…fault flink-conf.yaml to 1568MB.
…ault value to 192MB.
5f9cd22
to
f57ca41
Compare
Thanks for the review, @azagrebin. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks a lot for providing this PR @xintongsong. The changes look good to me. The only thing I would suggest to change is the term TaskExecutor
into TaskManager
because that is also the prefix of the newly introduced memory configuration options.
@@ -42,10 +42,16 @@ jobmanager.rpc.port: 6123 | |||
jobmanager.heap.size: 1024m | |||
|
|||
|
|||
# The heap size for the TaskManager JVM | |||
# The total process memory size for the TaskExecutor. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I would suggest to use the term TaskExecutor
for time being because we have never really changed the name in the documetation.
# The heap size for the TaskManager JVM | ||
# The total process memory size for the TaskExecutor. | ||
# | ||
# Note this accounts for all memory usage within the TaskExecutor process, including JVM metaspace and other overhead. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Same here with TaskExecutor
and TaskManager
.
…y size in default flink-conf.yaml.
Thank you @tillrohrmann. |
Thanks @xintongsong for addressing my comments so quickly. Merging this PR now. |
…t value to 96MB. This closes #10860.
What is the purpose of the change
This PR updates FLIP-49 TM memory sizes configuration default values according to the outcome of tuning with real jobs.
Brief change log
Verifying this change
This change is a trivial rework / code cleanup without any test coverage.
Does this pull request potentially affect one of the following parts:
@Public(Evolving)
: (yes)Documentation