Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

The KIP is accepted with 3 binding votes (David, Jun, Jason) and 1 non-binding vote (Alexandre). Thanks! On Tue, Feb 28, 2023 at 3:06 PM Jason Gustafson wrote: > Thanks Calvin, +1 from me. > > On Mon, Feb 27, 2023 at 9:41 AM Calvin Liu > wrote: > > > Hi Jason, > > Updated the fields

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

Thanks Calvin, +1 from me. On Mon, Feb 27, 2023 at 9:41 AM Calvin Liu wrote: > Hi Jason, > Updated the fields accordingly. Also, rename the BrokerState to > ReplicaState. > Thanks. > > On Wed, Feb 22, 2023 at 4:38 PM Jason Gustafson > > wrote: > > > Hi Calvin, > > > > The `BrokerState` struct

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

Hi Jason, Updated the fields accordingly. Also, rename the BrokerState to ReplicaState. Thanks. On Wed, Feb 22, 2023 at 4:38 PM Jason Gustafson wrote: > Hi Calvin, > > The `BrokerState` struct I suggested would replace the `BrokerId` field in > older versions. > > { "name": "ReplicaId",

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

Hi Calvin, The `BrokerState` struct I suggested would replace the `BrokerId` field in older versions. { "name": "ReplicaId", "type": "int32", "versions": "0-13", "entityType": "brokerId", "about": "The broker ID of the follower, of -1 if this request is from a consumer." }, {

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

To Jose: 1. Actually I have a second thoughts about the naming ReplicaEpoch. The BrokerEpoch only applies to the replication protocol between the brokers. For others like the KRaft controller, this field can be ignored. So can we change the name to ReplicaEpoch when we really use it in other

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

To Jason: 1. Related to the Fetch Request fields change, previously you suggested deprecating the ReplicaId and moving it into a BrokerState field. How about we just make the BrokerEpoch a tag field? - The ReplicaId is currently in use and is filled every time. So that we can keep the way simple.

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

To Jason: 1. We can make the BrokerEpoch a tagged field. But I am not sure about your proposed metadata structure. In the BrokerState, do we need to store the BrokerId again? It would duplicate with ReplicaId. 2. Considering that the broker reboot data loss case is rare and Kraft is coming soon.

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

Hi Calvin, Thanks for the improvement. 1. In the KIP, you suggest changing the Fetch request to "Rename the ReplicaId to BrokerId" and "Add a new Field BrokerEpoch". The Fetch RPC is used by replicas that are not brokers, for example controllers in KRaft. Can we keep the name "ReplicaId" and use

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

Hi Calvin, thanks for the KIP! A couple questions below: - Would it make sense to make the broker epoch a tagged field inside the Fetch request? It is a little annoying to increase the size of consumer fetch requests for a field that they will not use. Perhaps we could factor out broker id and

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

+1 (binding). Thanks for the KIP, Calvin! On Tue, Feb 14, 2023 at 1:10 AM Jun Rao wrote: > Hi, Calvin, > > Thanks for the KIP. +1 > > Jun > > On Fri, Feb 10, 2023 at 11:03 AM Alexandre Dupriez < > alexandre.dupr...@gmail.com> wrote: > > > +1 (non-binding). > > > > Le ven. 10 févr. 2023 à 19:02,

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

Hi, Calvin, Thanks for the KIP. +1 Jun On Fri, Feb 10, 2023 at 11:03 AM Alexandre Dupriez < alexandre.dupr...@gmail.com> wrote: > +1 (non-binding). > > Le ven. 10 févr. 2023 à 19:02, Calvin Liu a > écrit : > > > > Hi all, > > > > I'd like to call for a vote on KIP-903, which proposes a fix to

Re: [VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

+1 (non-binding). Le ven. 10 févr. 2023 à 19:02, Calvin Liu a écrit : > > Hi all, > > I'd like to call for a vote on KIP-903, which proposes a fix to the broker > reboot data loss KAFKA-14139 > > It changes the Fetch and AlterPartition requests

[VOTE] KIP-903: Replicas with stale broker epoch should not be allowed to join the ISR

Hi all, I'd like to call for a vote on KIP-903, which proposes a fix to the broker reboot data loss KAFKA-14139 It changes the Fetch and AlterPartition requests to include the broker epochs. Then the controller can use these epochs to help