Revert "Preserve case for RowType's field name and JSON content when … #21864
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
When process is terminated with core wait for termination
The current join nodes are in presto-main. Moving them requires moving a lot of objets. Since we don't need to allow the connectos to optimize the joins for now we can just add an adapter node that will be converted back to the internal node type when returned by connector optimzers.
Support for $path and $data_sequence_number hidden columns. The first is generally useful and the second is a requirement for implementing equality deletes as a join.
The current equality delete implementation applies deletes at the split level. Since equality deletes often apply to a lot of files, the current implementation ends up opening the delete files #splits * #delete_files times. This commit implements equality deletes as a join. A connector optimizer is added to apply the appropriate join(s).
…ffer manager Add http request active check when fetch data from output buffer manager in Velox. The active check is based on whether the http callstate has been destroyed or the associated request has expired. This is to avoid arbitrary output buffer to load data into a destination buffer which has set notify but the associated client request has expired. This helps to accelerate the shuffle for query with scale writer which uses arbitrary output buffer. Unit test is added to verify this behavior.
Same as in TaskResource
Point the original endpoint to /v1/task/async
Add task runtime stats to collect memory reclaim stats during a task execution. This helps to debug how much time spent on task memory reclaim of a slow query execution. If the memory reclaim is triggered by the task itself, then the time will also be counted in task cpu execution time. If not, it is counted in the task's scheduled time but not execution time as the task will be stopped before memory reclamation.
To reduce likelihood of running into port allocation conflicts
partitions are a coordinator-only parameter, therefore can be ignored when serializing
…ck for too long Also detect deadlock or starving and signal alerts if these happen.
- Implement HTTP request retries in PrestoSparkHttpTaskClient once and reuse for all kind of Task requests - Do not retry PrestoException and invalid responses by the server as they usually indicate programming errors and not communication errors (rely SimpleHttpResponseHandler to mimic Presto coordinator logic)
Avoid using scheduled executors to run callbacks as it may run out of threads and have none to schedule tasks
ScheduledThreadPoolExecutor is a fixed size thread pool and it never grows
Presto doesn't have PartitionedOutputNode and assigns its source node's plan node id to PartitionedOutputOperator. so Presto UI is expecting PartitionedOutputOperator having same node id as its source operator. However, Velox has PartitionedOutputNode whose plan node is is set to "root". To comply with original assumption, fix partitionedOutputNode plan node id to be "root.<source plan node ID>" and parse them back to source plan node ID in task stats reporting. Issue: #21741
We want to use runtimestats in DirectoryLister implementation to measure different runtime metrics
This was missed in the recent dependency update.
This has been moved to the corresponding Velox setup scripts
This was required for the antlr dependency which is now removed.
RE2 is installed in the Velox centos setup script
The six package is not used in Prestissimo
…``CAST``" This reverts commit d5a31b3.
NikhilCollooru
requested review from
shrinidhijoshi,
a team and
sdruzkin
as code owners
NikhilCollooru
requested review from
presto-oss
and removed request for
a team
|
Codenotify: Notifying subscribers in CODENOTIFY files for diff c192fea...25c5208.
|
|
@NikhilCollooru looks like you reverted a bunch of other PRs with this 4 line PR. Was it intentional? |
|
@sdruzkin doesn't look like this PR was merged eventually. Are you seeing your commits reverted on master? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Reverts the PR 21602
#21602