[Bug][Connector-v2][RedisSink]Fix failures caused by different cluster mode operations.#5300
[Bug][Connector-v2][RedisSink]Fix failures caused by different cluster mode operations.#5300lightzhao wants to merge 9 commits intoapache:devfrom
Conversation
Hisoka-X
left a comment
There was a problem hiding this comment.
Thanks @lightzhao , over all fine. @TyrantLucifer maybe you should take a look, please.
|
The code style should be fixed. |
|
hello @lightzhao Can you provide v2 connector documentation for this expressing thanks |
TyrantLucifer
left a comment
There was a problem hiding this comment.
Expire feature has some problems, I think this feature #4975 should revert and redesign it. For expire key, we should only operation once not every time after set operation.
|
BTW, whether in cluster mode or standalone mode, all operations should be abstracted into upper level interfaces |
What does it mean to operate only once, and how to implement only one operation? |
This looks good, abstract class + interface can implement different client operations. |
ok, i try it. |
|
I will redesign. |
Purpose of this pull request
If the redis mode is a cluster type, it will fail when operating the expire method.
So different redis cluster types should declare and use corresponding objects.
Check list
New License Guide
release-note.