Re: [Discussion] Breaking change of Pulsar standalone's start

2022-12-13 Thread Yunze Xu
FYI, I pushed a PR: https://github.com/apache/pulsar/pull/18909 Thanks, Yunze On Tue, Dec 13, 2022 at 7:55 PM Yunze Xu wrote: > > In addition to the setup process, Go client assumes the cluster was > not created [1] while C++ and Python clients don't care if it's > created because it checks if t

Re: [Discussion] Breaking change of Pulsar standalone's start

2022-12-13 Thread Yunze Xu
In addition to the setup process, Go client assumes the cluster was not created [1] while C++ and Python clients don't care if it's created because it checks if the cluster exists before. [1] https://github.com/apache/pulsar-client-go/blob/1fe9b624ca673e3a647780b7c6c4b2567089285a/scripts/pulsar-t

[Discussion] Breaking change of Pulsar standalone's start

2022-12-13 Thread Yunze Xu
Hi all, I found a breaking change introduced from #15186 [1] that changed the behavior of Pulsar standalone. Before #15186, it used Pulsar's built-in admin client to create tenant (public) and namespace (public/default). After #15186, it created the metadata nodes directly to the metadata store.