File resource SPI , the current DS file resource storage supports HDFS, S3 and 
NONE, making plug-in allows users to customize storage methods to meet various 
storage needs


And the task spi should be the highest priority
===============================================================================================================
????????????????????DS????????????????????HDFS??S3 and 
NONE??????????????????????????????????????????????????????????


Task SPI ??????????????????????


------------------ ???????? ------------------
??????:                                                                         
                                               "dev"                            
                                                        
<[email protected]&gt;;
????????:&nbsp;2020??11??4??(??????) ????10:20
??????:&nbsp;"dev"<[email protected]&gt;;

????:&nbsp;Re:Re: [DISCUSS]Which part of DolphinScheduler can be more scalable?



log SPI +1















?? 2020-11-04 10:07:58??"JUN GAO" <[email protected]&gt; ??????
&gt;log SPI +1
&gt;
&gt;But I think this is a low priority task, we should give priority to
&gt;complete the registration center SPI and task SPI.
&gt;
&gt;
&gt;
&gt;CalvinKirs <[email protected]&gt; ??2020??11??3?????? ????11:04??????
&gt;
&gt;&gt; Thank you very much for your answer, I misunderstood what you meant.
&gt;&gt;
&gt;&gt;
&gt;&gt; log SPI+1
&gt;&gt;
&gt;&gt;
&gt;&gt; Best&nbsp; wishes??
&gt;&gt; CalvinKirs
&gt;&gt;
&gt;&gt;
&gt;&gt; On 11/3/2020 22:55??leon bao<[email protected]&gt; wrote??
&gt;&gt; Maybe it didn't express my meaning well.
&gt;&gt; Yes, DolphinScheduler doesn't need to do this storage. We just need to 
do
&gt;&gt; the log plug-in function to support users to implement their own log
&gt;&gt; storage mode, just like we do alert plug-ins.
&gt;&gt;
&gt;&gt;
&gt;&gt; ????????????????????????????
&gt;&gt; 
??????DolphinScheduler??????????????????????????????????????????????????????????????????????????????????????????????????????
&gt;&gt;
&gt;&gt;
&gt;&gt; boyi <[email protected]&gt; ??2020??11??3?????? ????10:37??????
&gt;&gt;
&gt;&gt; hi:
&gt;&gt;
&gt;&gt;
&gt;&gt; Agree with Calvin KIRs
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt; Generally, there will be a unified log collection for such files, such 
as
&gt;&gt; ELK. Especially when deployed in the docker environment, we can 
consider
&gt;&gt; focusing on more important things
&gt;&gt;
&gt;&gt;
&gt;&gt; --------------------------------------
&gt;&gt; BoYi ZhangE-mail : [email protected]
&gt;&gt; On 11/3/2020 22:23??CalvinKirs<[email protected]&gt; wrote??
&gt;&gt; I understand what you mean, but what I want to express is whether it 
would
&gt;&gt; be better for users to do this part of the work. Generally, companies 
have
&gt;&gt; their own log collection system. If we were to do this part of the 
storage,
&gt;&gt; the workload would be great, and the benefits would not be great.
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt; 
??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????
&gt;&gt;
&gt;&gt;
&gt;&gt; Best&nbsp; wishes??
&gt;&gt; CalvinKirs
&gt;&gt;
&gt;&gt;
&gt;&gt; On 11/3/2020 22:09??leon bao<[email protected]&gt; wrote??
&gt;&gt; @CalvinKirs
&gt;&gt;
&gt;&gt; about log spi , now have a requirement in scalable
&gt;&gt; services(master/worker), this kind of application scenario requires 
that
&gt;&gt; the task log cannot be stored in the worker / master, but needs to be
&gt;&gt; stored in a third-party place, which maybe database or other
&gt;&gt; storage. Therefore, if the dolphin scheduler can provide this plug-in
&gt;&gt; function, users can read and write logs according to their own needs.
&gt;&gt;
&gt;&gt;
&gt;&gt; 
??????????????????????????????????????????????(master/worker)????????????????????????????????????????????worker/master????
&gt;&gt;
&gt;&gt;
&gt;&gt; 
????????????????????????????????????????????????????????????????DolphinScheduler??????????????????????????????????????????????????????????.
&gt;&gt;
&gt;&gt; CalvinKirs <[email protected]&gt; ??2020??11??3?????? ????6:40??????
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt; Great planning!
&gt;&gt; But I have a little question, what are the specific requirements of the
&gt;&gt; log SPI? I am not very clear at present, are we only implementing SPI 
for
&gt;&gt; data storage? If this is the case, is it necessary? I think this user 
can
&gt;&gt; use logagent (or other technologies) for related implementations. 
Different
&gt;&gt; users have different needs. Some users may also involve aggregation,
&gt;&gt; calculation, and even different magnitudes, and may use additional
&gt;&gt; components. Therefore, if we store the original data in this piece, a 
lot
&gt;&gt; of redundant data may be generated.
&gt;&gt;
&gt;&gt;
&gt;&gt; ??????????????
&gt;&gt;
&gt;&gt;
&gt;&gt; 
??????????????????????SPI????????????????????????????????????????????????????????????SPI????????????????????????????????????????????????????????????????logagent????????????????
&gt;&gt;
&gt;&gt;
&gt;&gt; 
????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????
&gt;&gt;
&gt;&gt;
&gt;&gt; Best&nbsp; wishes??
&gt;&gt; CalvinKirs
&gt;&gt;
&gt;&gt;
&gt;&gt; On 11/3/2020 11:47??leon bao<[email protected]&gt; wrote??
&gt;&gt; Hello Everyone:
&gt;&gt;
&gt;&gt; DS has good horizontal scalability with its non central design
&gt;&gt; architecture, which attracts many developers. With more and more 
users, the
&gt;&gt; demand for scheduling is becoming more and more complex.
&gt;&gt; At the same time, the functional design of DS is required to be more
&gt;&gt; scalable??for example: the plug-in function of alarm mode.
&gt;&gt; So&nbsp; we can discuss what parts of plug-ins DS can do at present. 
We can
&gt;&gt; reconstruct DolphinScheduler according to the results of this 
discussion.
&gt;&gt; At present, there are several parts of demand:
&gt;&gt;
&gt;&gt; - alert model:
&gt;&gt; refer to:
&gt;&gt; https://github.com/apache/incubator-dolphinscheduler/issues/3049
&gt;&gt;
&gt;&gt; - task plugin:
&gt;&gt; refer to:
&gt;&gt; https://github.com/apache/incubator-dolphinscheduler/issues/2869
&gt;&gt;
&gt;&gt; - register center:
&gt;&gt; refer to:
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt; 
https://lists.apache.org/thread.html/r755a57e3b859563de2dddf8aa2f336fcf28934e7bbb2c3f97fe5fe3d%40%3Cdev.dolphinscheduler.apache.org%3E
&gt;&gt; https://github.com/apache/incubator-dolphinscheduler/issues/3961
&gt;&gt;
&gt;&gt; - log model:
&gt;&gt; The current log is recorded by writing local files of the server.
&gt;&gt; Can we make this plug-in type, which can facilitate users to extend the
&gt;&gt; log reading and writing types, such as writing to the database or other
&gt;&gt; third-party systems.
&gt;&gt;
&gt;&gt; - global task queue
&gt;&gt; At present, tasks are stored in the memory queue of the master, which
&gt;&gt; results in the priority of a task can only work within the scope of a
&gt;&gt; master.
&gt;&gt; In order to make the priority of a task effective globally, we need a
&gt;&gt; global queue to make the global priority work.
&gt;&gt; (in version 1.2, we used zookeeper as the global queue, which was 
removed
&gt;&gt; because of the delay of ZK operation)
&gt;&gt;
&gt;&gt; Implementation details can be discussed within each topic. Here, we 
only
&gt;&gt; discuss the requirements.
&gt;&gt; Very appreciate you can put forward more opinions.
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt; 
==================================================================================================================================================================
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt; 
DS??????????????????????????????????????????????????????????????????????????????DS????????????????????????????????????????????????DS????????????????????????????
&gt;&gt; 
??????????????????????????????????????????????????DS????????????????????????????????????????????????????????DS????????????????????
&gt;&gt; ??????????????????????????
&gt;&gt;
&gt;&gt; - ????????(running)
&gt;&gt; ??????????
&gt;&gt; https://github.com/apache/incubator-dolphinscheduler/issues/3049
&gt;&gt;
&gt;&gt; - ????????
&gt;&gt; https://github.com/apache/incubator-dolphinscheduler/issues/2869
&gt;&gt;
&gt;&gt; - ????????
&gt;&gt; ??????????
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt; 
https://lists.apache.org/thread.html/r755a57e3b859563de2dddf8aa2f336fcf28934e7bbb2c3f97fe5fe3d%40%3Cdev.dolphinscheduler.apache.org%3E
&gt;&gt; https://github.com/apache/incubator-dolphinscheduler/issues/3961
&gt;&gt;
&gt;&gt; - ????????
&gt;&gt; 
??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????
&gt;&gt;
&gt;&gt; - ????????????
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt; 
??????????????????master??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????(????1.2??????????????zookeeper????????????,????????zk??????????????????????????)??
&gt;&gt;
&gt;&gt; 
????????????????????????????????????????????????????????????????????????????????????
&gt;&gt;
&gt;&gt; --
&gt;&gt; DolphinScheduler(Incubator)&nbsp; PPMC
&gt;&gt; BaoLiang ????
&gt;&gt; [email protected]
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt; --
&gt;&gt; DolphinScheduler(Incubator)&nbsp; PPMC
&gt;&gt; BaoLiang ????
&gt;&gt; [email protected]
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt; --
&gt;&gt; DolphinScheduler(Incubator)&nbsp; PPMC
&gt;&gt; BaoLiang ????
&gt;&gt; [email protected]
&gt;&gt;
&gt;
&gt;
&gt;-- 
&gt;
&gt;DolphinScheduler(Incubator)&nbsp; PPMC
&gt;Jun Gao ????
&gt;[email protected]

Reply via email to