Re: [DISCUSS] PIP-378 Add ServiceUnitStateTableView abstraction in ExtensibleLoadMangerImpl

2024-09-16 Thread Heesung Sohn
Hi all, In the PIP, I missed how to migrate from system topic to metadata store for the ExtensibleLoadMangerImpl service unit(bundle) ownership storage. I updated this part in the pip, and please check if this direction is fine too. https://github.com/apache/pulsar/pull/23300/commits/5d25ef8535bd

Re: [VOTE] PIP-378 Add ServiceUnitStateTableView abstraction (ExtensibleLoadMangerImpl only)

2024-09-16 Thread PengHui Li
+1 (binding) Regards, Penghui On Mon, Sep 16, 2024 at 8:17 AM Yunze Xu wrote: > +1 (binding) > > Thanks, > Yunze > > On Mon, Sep 16, 2024 at 3:06 AM Lari Hotari wrote: > > > > +1 (binding) > > > > -Lari > > > > On Sun, 15 Sept 2024 at 21:34, Heesung Sohn wrote: > > > > > > Hi all, > > > > > >

Re: [VOTE] PIP-378 Add ServiceUnitStateTableView abstraction (ExtensibleLoadMangerImpl only)

2024-09-16 Thread Yunze Xu
+1 (binding) Thanks, Yunze On Mon, Sep 16, 2024 at 3:06 AM Lari Hotari wrote: > > +1 (binding) > > -Lari > > On Sun, 15 Sept 2024 at 21:34, Heesung Sohn wrote: > > > > Hi all, > > > > I would like to start a vote on PIP-378 to add > > ServiceUnitStateTableView abstraction and make it pluggable.

Re: [DISCUSS] PIP-379: Key_Shared Draining Hashes for Improved Message Ordering

2024-09-16 Thread Lari Hotari
Hi Girish, Thank you for your feedback and for raising an important concern about the potential memory impact of the draining hashes hashmap, especially for large numbers of hashes. I'm glad to inform you that the memory impact related to the draining hashes state is not significant. On the br

Re: [DISCUSS] PIP-379: Key_Shared Draining Hashes for Improved Message Ordering

2024-09-16 Thread Girish Sharma
This is a good proposal and solves a big blocker that we have in using key_shared for a few certain use cases. My concern is on the memory impact of maintaining the draining hashes hashmap, especially when the hashes are in the order of 10,000s or more. Will there be checks or constraints to limit

Re: [DISCUSS] PIP-379: Key_Shared Draining Hashes for Improved Message Ordering

2024-09-16 Thread Lari Hotari
Hi PengHui, Thank you for your suggestion. I really appreciate your input on this, and I understand that without context, your proposed approach would indeed be the way forward. In this particular case, though, there are compelling reasons why PIP-379 suggests replacing the "recently joined co