Github user srowen commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-103157190
I think this has timed out. Would you close this PR for now?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as
Github user liyezhang556520 closed the pull request at:
https://github.com/apache/spark/pull/3825
---
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
Github user liyezhang556520 commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68684212
@JoshRosen , If we want to use the supervision mechanism. We need to add
another actor level as parent of the current Master actor. I don't know if that
is
Github user liyezhang556520 commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68828355
@markhamstra , thanks for reminder, I'll update this PR by making a try to
introduce the supervision.
---
If your project is set up for it, you can reply to
Github user markhamstra commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68721665
@liyezhang556520 That's been done already in the DAGScheduler. If we need
another level of supervision for Master or other actors, we should consider
whether these
Github user JoshRosen commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68648457
@markhamstra From that page:
Depending on the nature of the work to be supervised and the nature of
the failure, the supervisor has a choice of the following
Github user markhamstra commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68650323
@JoshRosen your thinking is that Master will be in good shape even though
an exception has been thrown? If you can guarantee that, then resuming the
actor while
Github user JoshRosen commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68429462
I'd rather use fewer Akka features than more, since this will make it
easier to replace Akka with our own RPC layer in the future. Therefore, I'd
much prefer to not
Github user markhamstra commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68445280
It doesn't seem to me that usage of the newer Akka persistence API is
called for, but it does seem that wrapping the `receive` in a try-catch is
trying to do the job
Github user liyezhang556520 commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68424069
Hi @JoshRosen , it'll be a little weird to wrap the `receive` or
`receiveWithLogging` method with try-catch blocks. And also this is conflict
with the fault
Github user liyezhang556520 commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68428197
I think a better way is to use the Akka's
[persistence](http://doc.akka.io/docs/akka/snapshot/scala/persistence.html)
feature, recover the actor's state when
GitHub user liyezhang556520 opened a pull request:
https://github.com/apache/spark/pull/3825
[SPARK-4991][CORE] Worker should reconnect to Master when Master actor
restart
This is a following JIRA of
[SPARK-4989](https://issues.apache.org/jira/browse/SPARK-4991). when Master
akka
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68241590
[Test build #24859 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/24859/consoleFull)
for PR 3825 at commit
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68246470
[Test build #24859 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/24859/consoleFull)
for PR 3825 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68246474
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
Github user JoshRosen commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68284859
Wouldn't it be better to ensure that actors like Master and DAGScheduler
never die due to uncaught exceptions?
---
If your project is set up for it, you can reply to
Github user JoshRosen commented on the pull request:
https://github.com/apache/spark/pull/3825#issuecomment-68286943
More specifically, I guess I'm suggesting that we modify wrap the `receive`
and `receiveWithLogging` methods of our actors with try-catch blocks to log any
exceptions
17 matches
Mail list logo