Github user shivaram commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-50371592
Updated the JIRA -- Closing this issue
---
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 do
Github user shivaram closed the pull request at:
https://github.com/apache/spark/pull/1471
---
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 en
Github user mateiz commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-50226768
Sure, you can modify the existing one.
---
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
Github user shivaram commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-50210528
Yeah I will close this PR -- Should I just modify SPARK-2563 for the Socket
re-opening issue or do you think a new JIRA is better ?
---
If your project is set up for i
Github user mateiz commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-50209890
If this PR doesn't help by the way, make sure to close it too so it doesn't
stay in the list.
---
If your project is set up for it, you can reply to this email and have y
Github user mateiz commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-50209854
I see, got it. It sounds like we should open a JIRA for creating a new
socket then. It's pretty strange that you can't reuse the same one in Java, but
I guess that's how i
Github user shivaram commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-50208195
@mateiz So I looked at this more closely today -- It turns out these
retries don't help much with Connection timed out exceptions. If the connection
attempt times out, t
Github user mateiz commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-50206285
@shivaram what did you think about sleeping longer after each attempt? Does
each attempt already take some time to time out? Otherwise we are only sleeping
1 ms.
---
If
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-50191160
QA results for PR 1471:- This patch PASSES unit tests.- This patch
merges cleanly- This patch adds no public classesFor more
information see test
ouptut:https://amplab.c
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-50184891
QA tests have started for PR 1471. This patch merges cleanly. View
progress:
https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/17192/consoleFull
---
If
Github user shivaram commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-49911839
Actually hold off on merging this -- I found that this patch doesn't
completely solve the problem. The issue I think is that `finishConnect` throws
an IOException [1] if
Github user mateiz commented on a diff in the pull request:
https://github.com/apache/spark/pull/1471#discussion_r15302438
--- Diff: docs/configuration.md ---
@@ -755,6 +755,13 @@ Apart from these, the following properties are also
available, and may be useful
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-49384160
QA results for PR 1471:- This patch PASSES unit tests.- This patch
merges cleanly- This patch adds no public classesFor more
information see test
ouptut:https://amplab.c
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1471#issuecomment-49378078
QA tests have started for PR 1471. This patch merges cleanly. View
progress:
https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/16793/consoleFull
---
If
GitHub user shivaram opened a pull request:
https://github.com/apache/spark/pull/1471
[SPARK-2563] Make connection retries configurable
In a large EC2 cluster, I often see the first shuffle stage in a job fail
due to connection timeout exceptions. This patch makes the number of conn
15 matches
Mail list logo