I agree, we should not work around the testcase but rather understand
and fix the root cause.
Closure cleaner should have null'ed out the references and allowed it
to be serialized.
Regards,
Mridul
On Sun, Aug 5, 2018 at 8:38 PM Wenchen Fan wrote:
>
> It seems to me that the closure cleaner fail
Answering one of missed question:
> I am not sure how were you planning to expose the state key groups at
api level and if it would be transparent.
I was thinking about introducing new configuration: it may look like adding
unnecessary configuration, but I thought it would help elasticity
("adap
Great help from the community!
On Sun, Aug 5, 2018 at 6:17 PM Xiao Li wrote:
> FYI, the new hints have been merged. They will be available in the
> upcoming release (Spark 2.4).
>
> *John Zhuge*, thanks for your work! Really appreciate it! Please submit
> more PRs and help the community improve
https://issues.apache.org/jira/browse/SPARK-24940
The PR has been merged to 2.4.0.
On Sun, Aug 5, 2018 at 6:06 PM Koert Kuipers wrote:
> lukas,
> what is the jira ticket for this? i would like to follow it's activity.
> thanks!
> koert
>
> On Wed, Jul 25, 2018 at 5:32 PM, lukas nalezenec wrote
It seems to me that the closure cleaner fails to clean up something. The
failed test case defines a serializable class inside the test case, and the
class doesn't refer to anything in the outer class. Ideally it can be
serialized after cleaning up the closure.
This is somehow a very weird way to d
FYI, the new hints have been merged. They will be available in the upcoming
release (Spark 2.4).
*John Zhuge*, thanks for your work! Really appreciate it! Please submit
more PRs and help the community improve Spark. : )
Xiao
2018-08-05 21:06 GMT-04:00 Koert Kuipers :
> lukas,
> what is the jira
lukas,
what is the jira ticket for this? i would like to follow it's activity.
thanks!
koert
On Wed, Jul 25, 2018 at 5:32 PM, lukas nalezenec wrote:
> Hi,
> Yes, This feature is planned - Spark should be soon able to repartition
> output by size.
> Lukas
>
>
> Dne st 25. 7. 2018 23:26 uživatel F
Hi Assaf,
No idea (and don't remember I've ever wondered about it before), but why
not doing this (untested):
trait MySparkTestTrait {
lazy val spark: SparkSession = SparkSession.builder().getOrCreate() //
<-- you sure you don't need master?
import spark.implicits._
}
Wouldn't that import wo
Makes sense, not sure if closure cleaning is related to the last one for
example or others. The last one is a bit weird, unless I am missing
something about the LegacyAccumulatorWrapper logic.
Stavros
On Sun, Aug 5, 2018 at 10:23 PM, Sean Owen wrote:
> Yep that's what I did. There are more fail
Yep that's what I did. There are more failures with different resolutions.
I'll open a JIRA and PR and ping you, to make sure that the changes are all
reasonable, and not an artifact of missing something about closure cleaning
in 2.12.
In the meantime having a 2.12 build up and running for master
Hi Sean,
I run a quick build so the failing tests seem to be:
- SPARK-17644: After one stage is aborted for too many failed
attempts, subsequent stagesstill behave correctly on fetch failures
*** FAILED ***
A job with one fetch failure should eventually succeed
(DAGSchedulerSuite.scala:2422)
Hi all,
I have been playing a bit with SQLImplicits and noticed that it is an
abstract class. I was wondering why is that? It has no constructor.
Because of it being an abstract class it means that adding a test trait
cannot extend it and still be a trait.
Consider the following:
trait MySp
Shane et al - could we get a test job in Jenkins to test the Scala 2.12
build? I don't think I have the access or expertise for it, though I could
probably copy and paste a job. I think we just need to clone the, say,
master Maven Hadoop 2.7 job, and add two steps: run
"./dev/change-scala-version.s
Yes it's a resaonable argument, that putting N more external integration
modules on the default spark-submit classpath might bring in more
third-party dependencies that clash or something. I think the convenience
factor isn't a big deal; users can also just write a dependence on said
module in thei
"coalesce" looks like working: I misunderstood it as an efficient version
of "repartition" which does shuffle, so expected it would trigger shuffle.
My proposal would be covered as using "coalesce": thanks Joseph for
correction. Let me abandon the proposal.
We may still miss for now is documentati
15 matches
Mail list logo