Thank you all. Closing the voting as we have got
sufficient votes. Will send out tally in a separate email.
On Sat, Aug 22, 2020 at 8:34 PM Shiyan Xu
wrote:
> Submitted the PR to update testing commands
> https://github.com/apache/hudi/pull/2010
>
> +1 (non-binding)
>
> - packaging ok
> -
+1 (non-binding)
- Compiles successfully
- Ran tests on EMR with bunch of upserts/delete commits, and verified query
results through spark datasource, spark-sql, hive and presto for COW/MOR
tables
- Ran insert/bulk insert/upserts on 100GB tpcds table
- Ran release validation scripts successfully
+1 (binding)
Downloaded tar and verified compile [OK]
Run integration test locally. [OK]
Run a few tests in IDE. [OK]
Run quickstart [OK]
Verify NOTICE and LICENSE exists [OK]
Check Checksum [OK]
Check no Binary files in source release [OK]
Rat Check Passed [OK]
On Sat, Aug 22, 2020 at
It can be up to the individual to use the IDE formatter or not, as long as
there is a tool to help enforce Checkstyle rules.
For people who use IDE formatter, importing Checkstyle.xml as a format
scheme does not fully control the formatter's behavior, that's why IDE
sometimes gets in the way. But
+1(binding)
1. Ran long running structured streaming writes on fake data and verified
compactions and ingestion is happening without errors.
2. Ran both scala and python based quickstart without any errors. There was
an issue in the documented quickstart steps (not in hudi) for python
example.
+1 (binding)
- Ran the rc checks, I typically do
- Tested a smoke test on both cow, mor tables
- by running lot commits over longer period of time,
- verifying the state of the dataset
- count validation match.
On Sat, Aug 22, 2020 at 6:08 AM leesf wrote:
> +1 (binding)
> - mvn
Aah, yes. That’s right.
On Sat, Aug 22, 2020 at 2:43 AM Vinoth Chandar wrote:
> All of the remaining meta fields compress very very nicely. They have
>
> almost no overhead.
>
>
>
> On Fri, Aug 21, 2020 at 12:00 PM Abhishek Modi
>
> wrote:
>
>
>
> > @sivabalan the current plan is to only add
+1 (binding)
- mvn clean package -DskipTests OK
- ran quickstart guide OK (still get the exception ERROR
view.PriorityBasedFileSystemView: Got error running preferred function.
Trying secondary
org.apache.hudi.exception.HoodieRemoteException: 192.168.1.102:56544 failed
to respond
at
Hi vc,
Yes, this part of the practice may have different preferences for different
developers. I have never opened the IDE's automatic formatting, nor have I
used the IDE's formatting functions artificially. Because I have
participated in multiple open source communities, each open source
Good moring Balaji, Vinoth,
Thank you both for your replies. I agree that this is a topic that should come
up more often and I am surprised that so little is said about this.
The option B in your mail (writing the delete marker also in the historical
records) sounds like a good option, but
Thanks Raymond. The tests ran successfully with these commands.
Best Regards,
Gary Li
On 8/21/20, 10:18 PM, "Shiyan Xu" wrote:
I should have documented this...(which I will soon)
When run from terminal, could you please try running with maven profile like
`mvn -Punit-tests
All of the remaining meta fields compress very very nicely. They have
almost no overhead.
On Fri, Aug 21, 2020 at 12:00 PM Abhishek Modi
wrote:
> @sivabalan the current plan is to only add this for hoodie_record_key. But
> I'm hoping to make the implementation general enough to add other
Hi David,
Thanks for the detailed email. and apologies for the sudden break in
communication.
>We wanted to manipulate the commit times to rebuild the history.
yes. best not to try and change the commit times/history.
>- replay the data omitting the data of the persons who have requested to
be
>But, IMO, we can ignore the IDE here, if it breaks the code style,
checkstyle will stop building and spotless will work.
I differ here slightly. Most people reformat code using the "format code"
in the IDE. And IDEs also can reorganize the code when you save etc.
We need a solid way to not be
14 matches
Mail list logo
