GitHub user blrunner opened a pull request:
https://github.com/apache/tajo/pull/959
TAJO-1905: Insert clause to partitioned table fails on S3
Currently, Tajo output committer works as following:
* Each task write output to a temp directory.
* ``FileTablespace::commitTable`` renames first successful task's temp
directory to final destination.
But above approach will occurs FileNotFoundException because of eventual
consistency of S3. To resolve it, I implemented output committer for S3 and the
committer works as following:
* Each task write output to local disk instead of S3 (in CTAS statement or
INERT statement)
* ``S3TableSpace::commitTable`` copies first successful task's temp
directory to S3.
This PR depends on https://github.com/apache/tajo/pull/952. CTAS statement
and INSERT statement for partition table ran successfully with this PR. For the
reference, I was inspired by Netflix integrating spark
slide(http://www.slideshare.net/piaozhexiu/netflix-integrating-spark-at-petabyte-scale-53391704).
To resolve this issue basically, each task need to write output to final
destination and we need to implement pluggable output committer. But this way
looks like a long time work. I think that this PR may be an interim work for
the pluggable output committer.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/blrunner/tajo TAJO-1905
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/tajo/pull/959.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #959
----
----
---
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 enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
