@
> kezhenxu94
I mean the new bug fix PR can be merged into the release branch, more people
can participate in release branch.
Best Regards
---
DolphinScheduler(Incubator) Committer
Xiaochun Liu 刘小春
liuxiaoc...@apache.org
---
> 2020年5月13日 下午2:06,kezhenxu94@apache 写道
Yeah, I agree with you.
We had better reduce PR during release, shouldn't integrate the new feature
into the dev branch, and we must ensure that release and dev branch should
be consistent.
If so, it will not be difficult to merge in the future.
Xiaochun Liu 于2020年5月13日周三 下午1:43写道:
>
> Hi, all
>
I’m totally confused why do you create a release branch if you’re going to
continuously merge new codes into THIS release branch? IMO, “release branch” is
for releasing and it aims to freeze the codes in some degree, new codes can be
still merged into the “dev” branch, which doesn’t block the
Hi!
In frontend, everyone has creation permission of the data source.
This can easily lead to repeated creation of data sources, and there are
different naming rules.
I think, creating a data source is a low-frequency operation.
So, permission should be revoked, only admin has create permission.
Hi, all
We encountered the following problems in the release branch management:
1. The release participation rate is low, it is difficult for community
developers, or it is not easy to participate in the release process. It is
mainly because everyone still fixes the bug by default, and the new
Until now ,DAG is one physical concept, which means that DAG is
heavy class , and introduces other classes and stored in a big json field.
On the bigdata platform , DAG is not a good concept for Schedulers, which can
be abandoned.
If DAG was abandoned, JOB and DEPENDENCY will simply the
archi
hi ,all ,for now the workers and masters will bing to a fixed port .
I think we need to replace it with random port, and store the current port
number
to zookeeper through heartbeat report.
Do you have any suggestions?
In addition to the job can be plugged in, is there any other content that
can be plugged in? If so, we need to consider the structural design in
advance, because this is a very important design transformation, we can
spend more time to think more
JUN GAO 于2020年5月13日周三 上午10:11写道:
> @hgaol There
@hgaol There are some discuss about DS Plugin . But
this is an early design and can only be used as a reference . I hope it
can help your design.
https://github.com/apache/incubator-dolphinscheduler/issues/201
I see the model `dolphinscheduler-plugin-api` . When I want to add a
ClickHouse Job
Hi @Gang Li, I agree with you that the plug-in module should be a common module
for all plug-in design. Currently I just create the SPI and plugin interface
for alert plug-in. If need more type for future, it’s very easy to define other
plugins.
Currently the alert service can detect the plug-i
we can refer to presto plug-in, etc.
1. Design the SPI and define the interface that the plugin needs to
implement
2. All plugins create new models to implement the interface in SPI.
The benefits are as follows:
1. For developers, they only need to care about the interface in SPI, and
do not need t
*> A better way is that user do not need to change the source code to
support new alert method. Instead, they just need to implement some
interface, and compliance with the convention, then package all(include
dependencies) in a single jar file, and put the jar file in some specific
place, such as
Hi Gao,
May be I didn't express clearly,I think the dolphinscheduler-plugin module is a
common module for all plugin design,
It's not neccessarily to put the implement of a plugin in the
dolphinscheduler-plugin module.
I also agree with the following way your said,I see the pr you submited didn
I create a new
PR, https://github.com/apache/incubator-dolphinscheduler/pull/2679
Because I'm not sure which methods will use the Retryer, so I just create
RetryerUtils which main function is retryCall
that will retry 3 times if exceptions throw or return result is false and wait
1 second betwe
yes, your description is proper, thanks
Best Regards
---
DolphinScheduler(Incubator) PPMC
Lidong Dai 代立冬
dailidon...@gmail.com
---
leon bao 于2020年5月12日周二 下午5:06写道:
> i think you want to show the DAG using list mode.
> DS show DAG using graphic mode, that will have some
Hello!
Abort dependent task check rules, I think the current implementation is
problematic.
E.g. check logic: this week
The current implementation:
1.get this Monday
2.the date list of this Monday to current date(asc order)
3.Iterate through the list of dates in sequence to check the status of a
i think you want to show the DAG using list mode.
DS show DAG using graphic mode, that will have some problems:
1. once the number of tasks is large, graphic mode will be confusing.
2. it is difficult to find the specified task in a complex DAG.
lidong dai 于2020年5月12日周二 下午4:31写道:
> I want to s
OK??I will test it
tanks
-- --
??: "qiaozhan...@outlook.com"https://github.com/apache/incubator-dolphinscheduler/pull/2693??commited
dev-1.3.0 ??please verify
??
DolphinScheduler(Incubator) PPMC
Zhanwei Qiao ??
qia
This PR https://github.com/apache/incubator-dolphinscheduler/pull/2693,commited
dev-1.3.0 ,please verify
―
DolphinScheduler(Incubator) PPMC
Zhanwei Qiao 乔占卫
qiaozhan...@outlook.com
发件人: qiaozhan...@outlook.com
发送时间: 2020-05-12 16:55
收件人: dev
主题: 回复: 回复: 【DISCUSS】Deleting the pro
Hi xinchun
I have commited,please test
Thx
―
DolphinScheduler(Incubator) PPMC
Zhanwei Qiao 乔占卫
qiaozhan...@outlook.com
发件人: qiaozhan...@outlook.com
发送时间: 2020-05-11 17:39
收件人: dev
主题: 回复: 【DISCUSS】Deleting the process instance does not delete the log file of
the task instance
I want to say do we need to implement List dependency?
Best Regards
---
DolphinScheduler(Incubator) PPMC
Lidong Dai 代立冬
dailidon...@gmail.com
---
JUN GAO 于2020年5月12日周二 下午12:26写道:
> Hi , @lidong dai
> Sorry , I don`t know your mean about this .
> You want to discuss ho
21 matches
Mail list logo