Github user hvanhovell commented on the issue:
https://github.com/apache/spark/pull/15722
LGTM. Merging to master/2.1. 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
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/15722
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/69793/
Test PASSed.
---
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/15722
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
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/15722
**[Test build #69793 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/69793/consoleFull)**
for PR 15722 at commit
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/15722
**[Test build #69793 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/69793/consoleFull)**
for PR 15722 at commit
Github user hvanhovell commented on the issue:
https://github.com/apache/spark/pull/15722
retest this please
---
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
Github user jiexiong commented on the issue:
https://github.com/apache/spark/pull/15722
@hvanhovell, thanks for your suggestion. I will change the PR description
as suggested.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as
Github user hvanhovell commented on the issue:
https://github.com/apache/spark/pull/15722
@jiexiong PR descriptions are used in git commit messages, and should be
clear and concise. The fix LGTM, but the description should be improved for
future reference. How about we change it into
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/15722
Merged build finished. Test FAILed.
---
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
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/15722
Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/69689/
Test FAILed.
---
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/15722
**[Test build #69689 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/69689/consoleFull)**
for PR 15722 at commit
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/15722
**[Test build #69689 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/69689/consoleFull)**
for PR 15722 at commit
Github user hvanhovell commented on the issue:
https://github.com/apache/spark/pull/15722
retest this please
---
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
Github user jiexiong commented on the issue:
https://github.com/apache/spark/pull/15722
Please retest!
---
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
Github user sitalkedia commented on the issue:
https://github.com/apache/spark/pull/15722
jenkins retest this please.
---
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
Github user dongjoon-hyun commented on the issue:
https://github.com/apache/spark/pull/15722
Retest this please.
---
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
Github user jiexiong commented on the issue:
https://github.com/apache/spark/pull/15722
@hvanhovell , I have already updated the description and explained how the
PR fixed it. Could you please take another look?
---
If your project is set up for it, you can reply to this email and
Github user hvanhovell commented on the issue:
https://github.com/apache/spark/pull/15722
ping @jiexiong
---
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,
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/15722
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
Github user hvanhovell commented on the issue:
https://github.com/apache/spark/pull/15722
@jiexiong could you improve the PR description and add a description of
what is actually causing this bug (a summary of the discussions in the PR would
probably suffice), and how this PR fixes
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/15722
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/68338/
Test PASSed.
---
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/15722
**[Test build #68338 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/68338/consoleFull)**
for PR 15722 at commit
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/15722
**[Test build #68338 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/68338/consoleFull)**
for PR 15722 at commit
Github user hvanhovell commented on the issue:
https://github.com/apache/spark/pull/15722
retest this please
---
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
Github user sitalkedia commented on the issue:
https://github.com/apache/spark/pull/15722
The test case failure does not seem to be related to this change. May be
flaky test?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/15722
**[Test build #3399 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/NewSparkPullRequestBuilder/3399/consoleFull)**
for PR 15722 at commit
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/15722
**[Test build #3399 has
started](https://amplab.cs.berkeley.edu/jenkins/job/NewSparkPullRequestBuilder/3399/consoleFull)**
for PR 15722 at commit
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/15722
**[Test build #3396 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/NewSparkPullRequestBuilder/3396/consoleFull)**
for PR 15722 at commit
Github user davies commented on the issue:
https://github.com/apache/spark/pull/15722
@sitalkedia I see your point, even the longArray only take a small factor
of memory of the executor, but still could make other concurrent tasks fail to
acquire enough initial memory, that make
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/15722
**[Test build #3396 has
started](https://amplab.cs.berkeley.edu/jenkins/job/NewSparkPullRequestBuilder/3396/consoleFull)**
for PR 15722 at commit
Github user sitalkedia commented on the issue:
https://github.com/apache/spark/pull/15722
@davies - Yes, we dumped the logging and confirmed that the OOM is because
we are not freeing the `LongArray` while reseting the `BytesToBytesMap`. The
job which used to fail because of OOM runs
Github user davies commented on the issue:
https://github.com/apache/spark/pull/15722
@jiexiong The longArray will not grow indefinitely, it only grow when the
number of keys reach 50% of it's size. Another assumption is that the memory
used by longArray should be much smaller than
Github user sitalkedia commented on the issue:
https://github.com/apache/spark/pull/15722
@davies - We fixed a similar issue with `UnsafeExternalSorter` in
SPARK-14363.
Basically following scenario is leading to OOM - Lets say we have total 4G
of memory available that is shared
Github user rxin commented on the issue:
https://github.com/apache/spark/pull/15722
@jiexiong can you update the pr description to include more information
about the OOM and remove the facebook internal links? Thanks.
---
If your project is set up for it, you can reply to this
Github user mridulm commented on the issue:
https://github.com/apache/spark/pull/15722
I agree with @jiexiong's analysis : @davies to support the scenario you
mentioned, we should not be free'ing the pages.
---
If your project is set up for it, you can reply to this email and have
Github user jiexiong commented on the issue:
https://github.com/apache/spark/pull/15722
Here is my understanding: after spilling, it would call reset() to release
the memory. In the reset() function, it deletes all the memory pages, but it
did not release any memory from longArray().
Github user davies commented on the issue:
https://github.com/apache/spark/pull/15722
@jiexiong Do you have a theory why this will cause OOM? To me, the current
code will use more memory than needed but less allocation, why it will cause
OOM?
---
If your project is set up for it,
37 matches
Mail list logo