Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/21490
Merged build finished. Test PASSed.
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/21490
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/94525/
Test PASSed.
---
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/21490
**[Test build #94525 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/94525/testReport)**
for PR 21490 at commit
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/21490
**[Test build #94525 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/94525/testReport)**
for PR 21490 at commit
Github user jose-torres commented on the issue:
https://github.com/apache/spark/pull/21490
(sorry, I've been busy with Spark Summit)
The problem I see is that fault tolerance might not be cleanly separable
from query stop tolerance. If a user stops the query at the wrong
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/21490
Can one of the admins verify this patch?
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional
Github user arunmahadevan commented on the issue:
https://github.com/apache/spark/pull/21490
@jerryshao , was not aware of your PR before raising this, so looks like we
both encountered similar issues. I think the intention is similar here and
attempts to start from last saved offset
Github user jerryshao commented on the issue:
https://github.com/apache/spark/pull/21490
@arunmahadevan I have already proposed a fix for this issue
https://github.com/apache/spark/pull/20958, but seems it is not so necessary to
fix based on the comments, and pending to review.
---
Github user arunmahadevan commented on the issue:
https://github.com/apache/spark/pull/21490
The patch does not attempt to make the micro-batch text-socket source fault
tolerant and I suppose that may not be very trivial.
This just makes the source work when query is
Github user jose-torres commented on the issue:
https://github.com/apache/spark/pull/21490
The proposed strategy still won't work correctly, though. If a batch fails
before commit, the restart won't replay the same records that were in that
batch, but the new ones incoming from the
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/21490
Merged build finished. Test PASSed.
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/21490
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/91453/
Test PASSed.
---
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/21490
**[Test build #91453 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/91453/testReport)**
for PR 21490 at commit
Github user arunmahadevan commented on the issue:
https://github.com/apache/spark/pull/21490
If "checkpointLocation" is set and the SS application is restarted, it
throws the below exception because the currentOffset and lastOffsetCommitted
are initialized to -1.
```
Github user jose-torres commented on the issue:
https://github.com/apache/spark/pull/21490
What's the motivation for this change? Does something go wrong if they
aren't initialized correctly? (And note that we document that the socket
source doesn't support fault recovery at all -
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/21490
**[Test build #91453 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/91453/testReport)**
for PR 21490 at commit
16 matches
Mail list logo