You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by zsxwing <gi...@git.apache.org> on 2017/04/27 20:48:22 UTC
[GitHub] spark pull request #17789: [SPARK-19525][CORE]Add RDD checkpoint compression...
GitHub user zsxwing opened a pull request:
https://github.com/apache/spark/pull/17789
[SPARK-19525][CORE]Add RDD checkpoint compression support
## What changes were proposed in this pull request?
This PR adds RDD checkpoint compression support and add a new config `spark.checkpoint.compress` to enable/disable it. Credit goes to @aramesh117
Closes #17024
## How was this patch tested?
The new unit test.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/zsxwing/spark pr17024
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/17789.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #17789
----
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by zsxwing <gi...@git.apache.org>.
Github user zsxwing commented on the issue:
https://github.com/apache/spark/pull/17789
> A question I had even with the earlier PR was - should we add the extension to either the directory or the file indicating compression type ?
Shuffle and cache files don't have an extension. I think it's better to be consistent in the whole code path.
> would it be a good idea to separate the codecs for compressing checkpoints and for network communication?
Save as above. Shuffle and cache files use the same codec.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by zsxwing <gi...@git.apache.org>.
Github user zsxwing commented on the issue:
https://github.com/apache/spark/pull/17789
> did this PR was for compression to be enabled for spark streaming usecase.
Streaming checkpoint includes two parts:
- DStream graph and metadata
- RDD checkpoints
Right now the first one is compressed. This PR is for RDD checkpoints.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by aramesh117 <gi...@git.apache.org>.
Github user aramesh117 commented on the issue:
https://github.com/apache/spark/pull/17789
@zsxwing Sorry for the delay! Thank you so much for your review and I saw a bit of your patch - it looks very nice. I have just one question - would it be a good idea to separate the codecs for compressing checkpoints and for network communication? I see that you reuse the same codec. If we wanted to isolate the effect of the RDD checkpointing, we would not be able to do that easily if these are coupled.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by zsxwing <gi...@git.apache.org>.
Github user zsxwing commented on the issue:
https://github.com/apache/spark/pull/17789
cc @mridulm since you reviewed the initial PR.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by AmplabJenkins <gi...@git.apache.org>.
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/17789
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/76242/
Test PASSed.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by mridulm <gi...@git.apache.org>.
Github user mridulm commented on the issue:
https://github.com/apache/spark/pull/17789
@zsxwing They are compressed ? Interesting ... I never played with spark streaming unfortunately, so did not know !
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by aramesh117 <gi...@git.apache.org>.
Github user aramesh117 commented on the issue:
https://github.com/apache/spark/pull/17789
@mridulm and @zsxwing thank you so much! This will help us out a lot! Much appreciated. :)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by zsxwing <gi...@git.apache.org>.
Github user zsxwing commented on the issue:
https://github.com/apache/spark/pull/17789
Streaming checkpoints are on HDFS but don't have an extension :)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by mridulm <gi...@git.apache.org>.
Github user mridulm commented on the issue:
https://github.com/apache/spark/pull/17789
I thought the main reason @aramesh117 did this PR was for compression to be enabled for spark streaming usecase.
If compression is already enabled, then am I missing something here ?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by mridulm <gi...@git.apache.org>.
Github user mridulm commented on the issue:
https://github.com/apache/spark/pull/17789
Ah interesting, thanks for clarifying ... weird that first was compressed and not second. But if there is an expectation some of the data is compressed already; perhaps we are being consistent now and there is no need to add extension (unless we uniformly do it everywhere).
LGTM, thanks for the change @zsxwing and @aramesh117 !
@zsxwing would it be possible for you to merge this ? I am facing some env issues right now. Thanks
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark pull request #17789: [SPARK-19525][CORE]Add RDD checkpoint compression...
Posted by mridulm <gi...@git.apache.org>.
Github user mridulm commented on a diff in the pull request:
https://github.com/apache/spark/pull/17789#discussion_r113822807
--- Diff: core/src/main/scala/org/apache/spark/rdd/ReliableCheckpointRDD.scala ---
@@ -169,7 +177,12 @@ private[spark] object ReliableCheckpointRDD extends Logging {
val bufferSize = env.conf.getInt("spark.buffer.size", 65536)
val fileOutputStream = if (blockSize < 0) {
- fs.create(tempOutputPath, false, bufferSize)
+ val fileStream = fs.create(tempOutputPath, false, bufferSize)
+ if (env.conf.get(CHECKPOINT_COMPRESS)) {
+ CompressionCodec.createCodec(env.conf).compressedOutputStream(fileStream)
--- End diff --
A question I had even with the earlier PR was - should we add the extension to either the directory or the file indicating compression type ?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by mridulm <gi...@git.apache.org>.
Github user mridulm commented on the issue:
https://github.com/apache/spark/pull/17789
Shuffle and cache files are not on hdfs :-) They do not potentially survive the application or be consumed OOB for recovery/inspection.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by zsxwing <gi...@git.apache.org>.
Github user zsxwing commented on the issue:
https://github.com/apache/spark/pull/17789
In addition, I agree that having an extension and separating the codecs are good ideas. But they should be done in other PRs to not introduce multiple features in a large PR.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by SparkQA <gi...@git.apache.org>.
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/17789
**[Test build #76242 has finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/76242/testReport)** for PR 17789 at commit [`dab44ed`](https://github.com/apache/spark/commit/dab44edce2cb9b49b44983fc944fbc1ebbf67ce7).
* This patch passes all tests.
* This patch merges cleanly.
* This patch adds no public classes.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark pull request #17789: [SPARK-19525][CORE]Add RDD checkpoint compression...
Posted by asfgit <gi...@git.apache.org>.
Github user asfgit closed the pull request at:
https://github.com/apache/spark/pull/17789
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by AmplabJenkins <gi...@git.apache.org>.
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/17789
Merged build finished. Test PASSed.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by mridulm <gi...@git.apache.org>.
Github user mridulm commented on the issue:
https://github.com/apache/spark/pull/17789
To add, for non streaming usecases, this will definitely help - but was this a recent change for streaming ? (probably after @aramesh117 make the PR ?)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by mridulm <gi...@git.apache.org>.
Github user mridulm commented on the issue:
https://github.com/apache/spark/pull/17789
Sounds good on doing it in separate PR - I am not too worried about shuffle/blockdata/etc btw - since they are private to application execution - checkpoint's tend to also be perused for other purposes (whether by design or not) since they are on hdfs; and with this PR it will become necessary to either guess the codec by iterating over supported codec's or pass it through other means for consumers.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by SparkQA <gi...@git.apache.org>.
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/17789
**[Test build #76242 has started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/76242/testReport)** for PR 17789 at commit [`dab44ed`](https://github.com/apache/spark/commit/dab44edce2cb9b49b44983fc944fbc1ebbf67ce7).
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by zsxwing <gi...@git.apache.org>.
Github user zsxwing commented on the issue:
https://github.com/apache/spark/pull/17789
yes. See https://github.com/apache/spark/blob/master/streaming/src/main/scala/org/apache/spark/streaming/Checkpoint.scala#L138
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] spark issue #17789: [SPARK-19525][CORE]Add RDD checkpoint compression suppor...
Posted by zsxwing <gi...@git.apache.org>.
Github user zsxwing commented on the issue:
https://github.com/apache/spark/pull/17789
Thanks, @mridulm @aramesh117 Merging to master and 2.2.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org