Adar Dembo has posted comments on this change. ( )

Change subject: add release notes for 1.9.0

Patch Set 3:

File docs/release_notes.adoc:
PS3, Line 58: link:TODO[administrative documentation]
This and a few other links are marked as TODO. What's the remaining work here?
PS3, Line 62: An link:[official
            :   repository] has been created for Apache Kudu Docker artifacts
Will this be populated prior to the release? It looks empty right now.
PS3, Line 64: People wanting to integrate with Kudu
Nit: maybe "Kudu integrators"?
PS3, Line 64: unit
Nit: drop 'unit'; you can do a Java test at any level.
PS3, Line 67: This functionality
            :   ships as part of the `kudu-test-utils` Maven module in the 
            :   class. In order for the `MiniCluster` to automatically find the 
            :   artifacts when starting up, the `kudu-binary` module must be 
added as a test
            :   time dependency, along with the `kudu-test-utils` module in the 
            :   Maven or Gradle build.
Too much implementation detail, I think. Can we link to a more detailed page 
and/or example instead?
PS3, Line 80: cut by a factor of its replication factor.
I think I wrote this, but can you think of a clearer way to say it? What we're 
trying to say is that if you had previously overridden that gflag, your 
effective max size for a new table of RF=3 is 1/3 the value of that gflag. "Cut 
by a factor of ..." isn't great because "factor of" suggests an additional 
scalar applied to the RF.
PS3, Line 84:   order (see KUDU-1400).
For this and other JIRAs, could we insert direct links to the JIRAs themselves? 
That's what we did for prior release notes.
PS3, Line 85: is
PS3, Line 91: * The Kudu-Spark example can now work against a Kudu cluster with 
a single
            :   tablet server and accepts a custom replication factor as a 
What makes this noteworthy?
PS3, Line 93: As a part of this upgrade `spark-avro`
            :   was migrated from the Databricks implementation to the 
implementation now
            :   included in Apache Spark. This has been done in an 
API-compatible way.
This seems like an implementation detail; why do Kudu users care about this 
bit? What does it mean for them?
PS3, Line 102: is
has been
PS3, Line 105: * The amount of server-side logging has been greatly reduced for 
             :   consensus implementation and background processes.
May want to add a tiny bit of color here. Just to talk about how we removed 
logging that didn't seem useful and was unnecessarily verbose.
PS3, Line 107: will now more obviously depict
now more obviously depicts
PS3, Line 109: is added
was added

Below too.
PS3, Line 112: supporting
             :   very basic predicates
A bit of color to explain the format in which predicates are described? Or too 
much detail?
PS3, Line 121: will now
             :   detect and report
now detects and reports
PS3, Line 128: The `--cmeta_force_fsync` flag is introduced to fsync Kudu's 
             :   metadata more agressively
The new ... flag may be used to fsync Kudu's consensus metadata more 
PS3, Line 130: improve its handling of
improve its durability in the face of

To view, visit
To unsubscribe, visit

Gerrit-Project: kudu
Gerrit-Branch: branch-1.9.x
Gerrit-MessageType: comment
Gerrit-Change-Id: I733dfae39c06f15f7f55ae823678caf6ca433bfc
Gerrit-Change-Number: 12389
Gerrit-PatchSet: 3
Gerrit-Owner: Andrew Wong <>
Gerrit-Reviewer: Adar Dembo <>
Gerrit-Reviewer: Alexey Serbin <>
Gerrit-Reviewer: Andrew Wong <>
Gerrit-Reviewer: Grant Henke <>
Gerrit-Reviewer: Hao Hao <>
Gerrit-Reviewer: Kudu Jenkins (120)
Gerrit-Reviewer: Mike Percy <>
Gerrit-Reviewer: Todd Lipcon <>
Gerrit-Reviewer: Will Berkeley <>
Gerrit-Comment-Date: Wed, 13 Feb 2019 19:45:29 +0000
Gerrit-HasComments: Yes

Reply via email to