Add release notes for 0.289#23513
Conversation
tdcmeehan
requested review from
a team,
elharo and
steveburnett
as code owners
|
@karteekmurthys |
"Add a REST API in Presto C++ worker to fetch worker stats in Prometheus Data Format." Is this sufficient? |
steveburnett
left a comment
steveburnett
left a comment
There was a problem hiding this comment.
Usual pull branch, local doc build, review local doc build.
What I did in this review:
- Added missing PR numbers to release note entries.
- Minor formatting of property names following Formatting in the Release Notes entries.
- Nit of punctuation, moving the
.from after the PR # to the end of the sentence. Edit applied throughout for consistency with the 0.288, 0.287, and 0.286 release notes. - A very few changes to the text of entries.
What I did not do:
- Review any of the Missing Release Notes items.
- Revise all of the entries that do not conform to the Order of changes in the Release Notes Guidelines. Specifically, the items in a set in the General Changes section beginning with
Make,Pass,Switch,The Icebergshould be rewritten. Others are welcome to, or I can address these in a second review after the changes in this review are applied. - Because of the previous point, I did not sort the entries in General Changes to follow the Order of changes in the Release Notes Guidelines. Again, that can be addressed in a later review pass. I largely wanted to address formatting issues in this first pass.
tdcmeehan
force-pushed
the
release-notes-0.289
branch
2 times, most recently
from
34ea7e9 to
d936a48
Compare
steveburnett
left a comment
steveburnett
left a comment
There was a problem hiding this comment.
Pull updated branch, local doc build, review local doc build.
What I did in this review:
- Checked all links to PR numbers to verify the linked PR had something to do with the release note entry. (Caught two typos in the PR numbers this way.)
- Checked all links in the release note entries.
- Reviewed the current draft as a whole again, looking for any nits in formatting.
tdcmeehan
force-pushed
the
release-notes-0.289
branch
from
d936a48 to
71e3673
Compare
|
In addition to my current open review above, one more thing: the release note entry for PR #23174 appears to be missing from the latest update. Was it intentionally removed? #23174 (Author: Zac Blanco): [Iceberg] Add table refresh retry configurations The Iceberg connector when configured with Hive, can now configure table metadata refresh timeouts. See documentation for details. It should be revised from its current wording if it is to be included. Maybe something like the following: Add configuration for retries of table metadata refresh timeouts, for the Iceberg connector when configured with Hive. :pr: |
I excluded this because I deemed it non-user facing and mostly for testing. @ZacBlanco do you agree? |
tdcmeehan
force-pushed
the
release-notes-0.289
branch
2 times, most recently
from
7570cfa to
b5fda53
Compare
|
@rschlussel does this need to be backported to 0.288?
|
|
Does this need to be backported to 0.288? |
yes, that should be backported if the other nan changes are in 0.288 |
steveburnett
left a comment
steveburnett
left a comment
There was a problem hiding this comment.
Only a few minor comments.
|
@pgupta2 I changed the release note for the query progress events:
Please check this and let me know what you think:
|
tdcmeehan
force-pushed
the
release-notes-0.289
branch
3 times, most recently
from
b75fd0b to
b801bc8
Compare
steveburnett
left a comment
steveburnett
left a comment
There was a problem hiding this comment.
One formatting nit I should have caught earlier, I think this should be everything.
tdcmeehan
force-pushed
the
release-notes-0.289
branch
from
b801bc8 to
3ebd0f4
Compare
That PR originated from a user from the community who was having issues with operations timing out and taking too long to fail, causing a hanging behavior. The configurations were added to the documentation because they may be desirable for a user to configure to decrease the amount of time spent "hanging". Personally I would keep it in the release notes, but don't have strong opinions. |
steveburnett
left a comment
steveburnett
left a comment
There was a problem hiding this comment.
LGTM! (docs)
Pull updated branch, local doc build, looks good. Thanks!
|
@ZacBlanco how does this sound?
|
Is that link right? The configuration properties added are at |
Oops, how about this:
I think the documentation you added will be anchored to the Glue section. |
tdcmeehan
force-pushed
the
release-notes-0.289
branch
2 times, most recently
from
3944af0 to
b01c4ae
Compare
|
I have incorporated @steveburnett and @elharo's feedback, and am now waiting on @pgupta2 on this item: #23513 (comment) Additionally, @swapsmagic it would be helpful if you could comment on #23513 (comment) |
@swapsmagic @pgupta2 ping |
@tdcmeehan : The changes looks good to me. Thanks. |
I don't think it's necessary, give the feature was introduced recently and disable by default. So unless someone started using it there is no need to backport it. |
8 participants
Missing Release Notes
Abe Varghese
Amit Dutta
Auden Woolfson
Bikramjeet Vig
Deepak Majeti
Elliotte Rusty Harold
Jialiang Tan
Jimmy Lu
Karteekmurthys
Ke
Kevin Wilfong
Miguel Blanco Godón
Pratik Joseph Dabre
Yiqun (Ethan) Zhang
abhinavmuk04
aditi-pandit
misterjpapa
wypb
Extracted Release Notes
hive.skip_empty_files:pr:22727.22936.array_distinct,array_except,array_intersect,array_union,set_union, andset_aggwould return an error if the input was an array of rows with null fields or array of arrays with null elements. These functions now use full "IS DISTINCT FROM" semantics for comparison.22949.control.reuse-tableandtest.reuse-tableconfiguration properties for the Presto Verifier to reuse the output tables of the source query for control and test.22890.hive.legacy-timestamp-bucketingand session propertyhive.legacy_timestamp_bucketingto use the original hash function for Timestamp column, which is not hive compatible. :pr:22890.23027.warn-on-possible-nansand session propertywarn_on_possible_nansto produce a warning on division operations or comparison operations involving double or real types. Division operations are common causes of accidental creation of nans, and the semantics of comparison operations involving nans changed considerably in the most recent Presto release.use-new-nan-definitionis false :pr:23060.remove_cross_join_with_constant_single_row_input:pr:23081.23104.23127.23142.hive.file-status-cache.max-retained-sizeconfiguration property.hive.file-status-cache-sizeis now deprecated :pr:23176.23178./clients/dbeaverdocumentation :pr:23189.cache.velox.full-persistence-intervalto a non-zero value. :pr:23192./clients/supersetdocumentation :pr:23194.event.query-progress-publish-intervalto specify the time interval at which progress events should be generated. Default is every 1 minute. :pr:23195.23200.enable_verbose_history_based_optimizer_runtime_statsto track latency of HBO optimizer :pr:23241.metadata_previous_versions_maxandmetadata_delete_after_committo maintain the previous metadata files :pr:23260.23260.remove_orphan_files#23267 (Author: wangd): [Iceberg] Support procedureremove_orphan_filesremove_orphan_filesto remove orphan files that are not referenced in any metadata files for Iceberg. :pr:23267.23312.23319).equalValuesAreIdenticaltoType. Override this method to returnfalsewhen the values of the type may have more than one canonical representation.enforce_history_based_optimizer_register_timeoutto enforce the maximum time HBO query registration can take :pr:23354.23356.hive.parquet.writer.version = PARQUET_2_0for old behavior. :pr:23369.12345.23404.23411.ip_prefix_collapse... :pr:23445.23447.23453.All Commits
_<number>(wangd)remove_orphan_files(wangd)iceberg.file-format(wangd)SecurityConfigfromSystemSessionProperties(wangd)TableWriterNode.UpdateTarget(wangd)HiveSplitManager(wangd)