You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@kafka.apache.org by Muralidhar Basani via dev <de...@kafka.apache.org> on 2026/08/06 21:23:20 UTC

Re: [DISCUSS] KIP-1362: Add a rollback-compatible storage format option for KRaft migration

Hi Anton,
Thanks for the KIP. It closes a real gap for 3.9 migrations.

I have a couple of minor comments.

mb0 - My understanding is this can only ship on 3.9.x since ZK migration is
removed on 4.0+. Maybe worth adding it to the KIP in the scope/target
section.

mb1 - Is there a possibility of misusing the flag
`--zk-rollback-compatible` (for ex, applying it on a non-migration cluster
or after finalization, then we see V0 meta props.) It may be better to add
some guidance in the KIP.

Thanks,
Murali

On Thu, Jul 9, 2026 at 12:18 PM Anton Agestam via dev <de...@kafka.apache.org>
wrote:

> Hi everyone,
>
> I would like to start the discussion for KIP-1362: Add a
> rollback-compatible storage format option for KRaft migration.
>
> KIP:
>
> https://cwiki.apache.org/confluence/spaces/KAFKA/pages/440303727/KIP-1362+Add+a+rollback-compatible+storage+format+option+for+KRaft+migration
> Jira: https://issues.apache.org/jira/browse/KAFKA-20762
> PR/reference implementation: https://github.com/apache/kafka/pull/22751
>
> In summary: this KIP proposes closing a current gap, where the ability to
> do rollback of an ongoing KRaft migration breaks when cluster member nodes
> are replaced. The proposed solution is to add a CLI flag to the storage
> formatting tool that causes the resulting storage to not break when rolled
> back.
>
> I'm looking forward to feedback on this proposal.
>
> Thanks,
> Anton
>
> --
> [image: Aiven] <https://www.aiven.io/>
> *Anton Agestam* (he/him or they/them)
> Software Engineer, *Aiven*
> anton.agestam@aiven.io   |   +46 704 486 289
> aiven.io <https://www.aiven.io/>   |
> <https://www.facebook.com/aivencloud>
> <https://www.linkedin.com/company/aiven/>    <https://twitter.com/aiven_io
> >
>

Re: [DISCUSS] KIP-1362: Add a rollback-compatible storage format option for KRaft migration

Posted by Anton Agestam via dev <de...@kafka.apache.org>.
Hello Murali, thanks for your thoughtful feedback.

mb0: Thanks for pointing this out, the observation is correct. I added a
sentence under Proposed Changes to clarify that this KIP only targets the
3.9 development branch.

mb1: There is some potential, yes, as it can be incorrectly provided when
provisioning a node with version 4.0+, i.e. that is not meant to go through
KRaft migration. However, I think the documentation as proposed already is
sufficient to steer users right with regards to this. See also the
reference PR, that adds a clear statement that the flag shall not be used
beyond KRaft migration.

BR,
Anton

Den tors 6 aug. 2026 kl 23:23 skrev Muralidhar Basani via dev <
dev@kafka.apache.org>:

> Hi Anton,
> Thanks for the KIP. It closes a real gap for 3.9 migrations.
>
> I have a couple of minor comments.
>
> mb0 - My understanding is this can only ship on 3.9.x since ZK migration is
> removed on 4.0+. Maybe worth adding it to the KIP in the scope/target
> section.
>
> mb1 - Is there a possibility of misusing the flag
> `--zk-rollback-compatible` (for ex, applying it on a non-migration cluster
> or after finalization, then we see V0 meta props.) It may be better to add
> some guidance in the KIP.
>
> Thanks,
> Murali
>
> On Thu, Jul 9, 2026 at 12:18 PM Anton Agestam via dev <
> dev@kafka.apache.org>
> wrote:
>
> > Hi everyone,
> >
> > I would like to start the discussion for KIP-1362: Add a
> > rollback-compatible storage format option for KRaft migration.
> >
> > KIP:
> >
> >
> https://cwiki.apache.org/confluence/spaces/KAFKA/pages/440303727/KIP-1362+Add+a+rollback-compatible+storage+format+option+for+KRaft+migration
> > Jira: https://issues.apache.org/jira/browse/KAFKA-20762
> > PR/reference implementation: https://github.com/apache/kafka/pull/22751
> >
> > In summary: this KIP proposes closing a current gap, where the ability to
> > do rollback of an ongoing KRaft migration breaks when cluster member
> nodes
> > are replaced. The proposed solution is to add a CLI flag to the storage
> > formatting tool that causes the resulting storage to not break when
> rolled
> > back.
> >
> > I'm looking forward to feedback on this proposal.
> >
> > Thanks,
> > Anton
> >
> > --
> > [image: Aiven] <https://www.aiven.io/>
> > *Anton Agestam* (he/him or they/them)
> > Software Engineer, *Aiven*
> > anton.agestam@aiven.io   |   +46 704 486 289
> > aiven.io <https://www.aiven.io/>   |
> > <https://www.facebook.com/aivencloud>
> > <https://www.linkedin.com/company/aiven/>    <
> https://twitter.com/aiven_io
> > >
> >
>