Re: Discussion on Default Implementation for isDriverMetadataServiceAvailable

2024-05-16 Thread Yong Zhang
I think it's an improvement of the error handle in the metadata failure case. Not all cases need to implement it. +1 to Andrey. That depends on which releases this feature wants to go. Yong On Fri, 17 May 2024 at 07:56, Andrey Yegorov wrote: > I am ok either way. > > if this doesn't have

Re: Discussion on Default Implementation for isDriverMetadataServiceAvailable

2024-05-16 Thread Andrey Yegorov
I am ok either way. if this doesn't have default implementation it can't be included into the 4.17.x releases and will have to wait for 4.18 (breaking change, e.g. pulsar implements MetadataClientDriver interface.) OTOH if this is used for anything beyond tests I'd suggest default

Re: Discussion on Default Implementation for isDriverMetadataServiceAvailable

2024-05-16 Thread Enrico Olivelli
ZhangJian, Il giorno mer 8 mag 2024 alle ore 01:34 ZhangJian He ha scritto: > Hi, bookKeepers > > I've reviewed the PR that introduces `isDriverMetadataServiceAvailable`. I > have concerns about providing a default implementation that returns a > constant value like true, it's not a default

Discussion on Default Implementation for isDriverMetadataServiceAvailable

2024-05-07 Thread ZhangJian He
Hi, bookKeepers I've reviewed the PR that introduces `isDriverMetadataServiceAvailable`. I have concerns about providing a default implementation that returns a constant value like true, it's not a default interface like this. https://github.com/apache/bookkeeper/pull/4342#discussion_r1591761669