You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ratis.apache.org by Attila Doroszlai <ad...@apache.org> on 2024/09/01 06:39:49 UTC
Re: Release Ratis 3.1.1
> > 1. Do we need to release the ratis-thirdparty 1.0.7 version first?
> RATIS-2119. Enable reproducible builds (#52)
> RATIS-2110. Publish SBOM artifacts (#51)
Both of these are only improvements for the build, so I don't think
it's necessary to delay 3.1.1.
It would be nice to prepare the release of 1.0.7 in parallel, and
include it in 3.1.1 only if that one needs to wait for anything else.
-Attila
Re: Release Ratis 3.1.1
Posted by Xinyu Tan <ta...@apache.org>.
Hi, Tsz Wo
Sure, I will include them.
I am currently having some issues with gpg signing, I will fix it soon to release rc0 version.
Best
------------
Xinyu Tan
On 2024/09/03 03:51:58 Tsz Wo Sze wrote:
> Hi Xinyu,
>
> Sure, let's skip Ratis-2129 in Ratis 3.1.1. How about RATIS-2145 and
> RATIS-2146? Let's include them? They can be cherry-picked cleanly.
>
> RATIS-2145. Follower hangs until the next trigger to take a snapshot.
> (#1139)
> RATIS-2146. Fixed possible issues caused by concurrent deletion and
> election when member changes (#1140)
>
> Tsz-Wo
>
>
> On Sun, Sep 1, 2024 at 9:20 PM Xinyu Tan <ta...@apache.org> wrote:
>
> > Hi,
> >
> > Based on Attila's suggestion, I reviewed the three newly added commits in
> > ratis-thirdparty and found that they have almost no impact on those who
> > depend on it, so it may not be necessary to release a new version for this.
> > Therefore, I will proceed directly with the release of ratis 3.1.1. If
> > anyone thinks it is necessary to release a new version of ratis-thirdparty,
> > please reply in the email, and I will release it concurrently.
> >
> > Additionally, regarding the commits included in Ratis 3.1.1, since
> > Ratis-2129 is a performance optimization issue, and performance
> > optimizations often come with potential risks, I do not recommend including
> > this commit in a version that Ozone urgently needs to be stable. If there
> > are no further suggestions, I will proceed with the release of the rc0
> > version based on the branch-3.1.1_review branch prepared by Tsz-Wo (many
> > thanks to Tsz-Wo for organizing the commits).
> >
> > Best
> > -------------
> > Xinyu Tan
> >
> > On 2024/09/01 06:39:49 Attila Doroszlai wrote:
> > > > > 1. Do we need to release the ratis-thirdparty 1.0.7 version first?
> > >
> > > > RATIS-2119. Enable reproducible builds (#52)
> > > > RATIS-2110. Publish SBOM artifacts (#51)
> > >
> > > Both of these are only improvements for the build, so I don't think
> > > it's necessary to delay 3.1.1.
> > >
> > > It would be nice to prepare the release of 1.0.7 in parallel, and
> > > include it in 3.1.1 only if that one needs to wait for anything else.
> > >
> > > -Attila
> > >
> >
>
Re: Release Ratis 3.1.1
Posted by Xinyu Tan <ta...@apache.org>.
Hi, Chuang
As Duong said in jira1983[1]:
The revert in release-3.1.1 branch is a workaround for 3.1 only. The root issue (RATIS-2141) only happens in 3.1 because 3.1 only contains the change from this issue which is halfway of the zero-copy implementation. The master branch contains the complete zero-copy implementation.
[1] https://issues.apache.org/jira/browse/RATIS-1983
Best
--------------
Xinyu Tan
On 2024/09/10 22:02:23 Wei-Chiu Chuang wrote:
> I updated jira fixed versions according to the current release-3.1.1 branch.
>
> One thing that stands out: RATIS-1983
> <https://issues.apache.org/jira/browse/RATIS-1983> is reverted from
> release-3.1.1 but not from master. Is that intentional?
>
> On Mon, Sep 2, 2024 at 8:52 PM Tsz Wo Sze <sz...@gmail.com> wrote:
>
> > Hi Xinyu,
> >
> > Sure, let's skip Ratis-2129 in Ratis 3.1.1. How about RATIS-2145 and
> > RATIS-2146? Let's include them? They can be cherry-picked cleanly.
> >
> > RATIS-2145. Follower hangs until the next trigger to take a snapshot.
> > (#1139)
> > RATIS-2146. Fixed possible issues caused by concurrent deletion and
> > election when member changes (#1140)
> >
> > Tsz-Wo
> >
> >
> > On Sun, Sep 1, 2024 at 9:20 PM Xinyu Tan <ta...@apache.org> wrote:
> >
> > > Hi,
> > >
> > > Based on Attila's suggestion, I reviewed the three newly added commits in
> > > ratis-thirdparty and found that they have almost no impact on those who
> > > depend on it, so it may not be necessary to release a new version for
> > this.
> > > Therefore, I will proceed directly with the release of ratis 3.1.1. If
> > > anyone thinks it is necessary to release a new version of
> > ratis-thirdparty,
> > > please reply in the email, and I will release it concurrently.
> > >
> > > Additionally, regarding the commits included in Ratis 3.1.1, since
> > > Ratis-2129 is a performance optimization issue, and performance
> > > optimizations often come with potential risks, I do not recommend
> > including
> > > this commit in a version that Ozone urgently needs to be stable. If there
> > > are no further suggestions, I will proceed with the release of the rc0
> > > version based on the branch-3.1.1_review branch prepared by Tsz-Wo (many
> > > thanks to Tsz-Wo for organizing the commits).
> > >
> > > Best
> > > -------------
> > > Xinyu Tan
> > >
> > > On 2024/09/01 06:39:49 Attila Doroszlai wrote:
> > > > > > 1. Do we need to release the ratis-thirdparty 1.0.7 version first?
> > > >
> > > > > RATIS-2119. Enable reproducible builds (#52)
> > > > > RATIS-2110. Publish SBOM artifacts (#51)
> > > >
> > > > Both of these are only improvements for the build, so I don't think
> > > > it's necessary to delay 3.1.1.
> > > >
> > > > It would be nice to prepare the release of 1.0.7 in parallel, and
> > > > include it in 3.1.1 only if that one needs to wait for anything else.
> > > >
> > > > -Attila
> > > >
> > >
> >
>
Re: Release Ratis 3.1.1
Posted by Wei-Chiu Chuang <we...@apache.org>.
I updated jira fixed versions according to the current release-3.1.1 branch.
One thing that stands out: RATIS-1983
<https://issues.apache.org/jira/browse/RATIS-1983> is reverted from
release-3.1.1 but not from master. Is that intentional?
On Mon, Sep 2, 2024 at 8:52 PM Tsz Wo Sze <sz...@gmail.com> wrote:
> Hi Xinyu,
>
> Sure, let's skip Ratis-2129 in Ratis 3.1.1. How about RATIS-2145 and
> RATIS-2146? Let's include them? They can be cherry-picked cleanly.
>
> RATIS-2145. Follower hangs until the next trigger to take a snapshot.
> (#1139)
> RATIS-2146. Fixed possible issues caused by concurrent deletion and
> election when member changes (#1140)
>
> Tsz-Wo
>
>
> On Sun, Sep 1, 2024 at 9:20 PM Xinyu Tan <ta...@apache.org> wrote:
>
> > Hi,
> >
> > Based on Attila's suggestion, I reviewed the three newly added commits in
> > ratis-thirdparty and found that they have almost no impact on those who
> > depend on it, so it may not be necessary to release a new version for
> this.
> > Therefore, I will proceed directly with the release of ratis 3.1.1. If
> > anyone thinks it is necessary to release a new version of
> ratis-thirdparty,
> > please reply in the email, and I will release it concurrently.
> >
> > Additionally, regarding the commits included in Ratis 3.1.1, since
> > Ratis-2129 is a performance optimization issue, and performance
> > optimizations often come with potential risks, I do not recommend
> including
> > this commit in a version that Ozone urgently needs to be stable. If there
> > are no further suggestions, I will proceed with the release of the rc0
> > version based on the branch-3.1.1_review branch prepared by Tsz-Wo (many
> > thanks to Tsz-Wo for organizing the commits).
> >
> > Best
> > -------------
> > Xinyu Tan
> >
> > On 2024/09/01 06:39:49 Attila Doroszlai wrote:
> > > > > 1. Do we need to release the ratis-thirdparty 1.0.7 version first?
> > >
> > > > RATIS-2119. Enable reproducible builds (#52)
> > > > RATIS-2110. Publish SBOM artifacts (#51)
> > >
> > > Both of these are only improvements for the build, so I don't think
> > > it's necessary to delay 3.1.1.
> > >
> > > It would be nice to prepare the release of 1.0.7 in parallel, and
> > > include it in 3.1.1 only if that one needs to wait for anything else.
> > >
> > > -Attila
> > >
> >
>
Re: Release Ratis 3.1.1
Posted by Tsz Wo Sze <sz...@gmail.com>.
Hi Xinyu,
Sure, let's skip Ratis-2129 in Ratis 3.1.1. How about RATIS-2145 and
RATIS-2146? Let's include them? They can be cherry-picked cleanly.
RATIS-2145. Follower hangs until the next trigger to take a snapshot.
(#1139)
RATIS-2146. Fixed possible issues caused by concurrent deletion and
election when member changes (#1140)
Tsz-Wo
On Sun, Sep 1, 2024 at 9:20 PM Xinyu Tan <ta...@apache.org> wrote:
> Hi,
>
> Based on Attila's suggestion, I reviewed the three newly added commits in
> ratis-thirdparty and found that they have almost no impact on those who
> depend on it, so it may not be necessary to release a new version for this.
> Therefore, I will proceed directly with the release of ratis 3.1.1. If
> anyone thinks it is necessary to release a new version of ratis-thirdparty,
> please reply in the email, and I will release it concurrently.
>
> Additionally, regarding the commits included in Ratis 3.1.1, since
> Ratis-2129 is a performance optimization issue, and performance
> optimizations often come with potential risks, I do not recommend including
> this commit in a version that Ozone urgently needs to be stable. If there
> are no further suggestions, I will proceed with the release of the rc0
> version based on the branch-3.1.1_review branch prepared by Tsz-Wo (many
> thanks to Tsz-Wo for organizing the commits).
>
> Best
> -------------
> Xinyu Tan
>
> On 2024/09/01 06:39:49 Attila Doroszlai wrote:
> > > > 1. Do we need to release the ratis-thirdparty 1.0.7 version first?
> >
> > > RATIS-2119. Enable reproducible builds (#52)
> > > RATIS-2110. Publish SBOM artifacts (#51)
> >
> > Both of these are only improvements for the build, so I don't think
> > it's necessary to delay 3.1.1.
> >
> > It would be nice to prepare the release of 1.0.7 in parallel, and
> > include it in 3.1.1 only if that one needs to wait for anything else.
> >
> > -Attila
> >
>
Re: Release Ratis 3.1.1
Posted by Xinyu Tan <ta...@apache.org>.
Hi,
Based on Attila's suggestion, I reviewed the three newly added commits in ratis-thirdparty and found that they have almost no impact on those who depend on it, so it may not be necessary to release a new version for this. Therefore, I will proceed directly with the release of ratis 3.1.1. If anyone thinks it is necessary to release a new version of ratis-thirdparty, please reply in the email, and I will release it concurrently.
Additionally, regarding the commits included in Ratis 3.1.1, since Ratis-2129 is a performance optimization issue, and performance optimizations often come with potential risks, I do not recommend including this commit in a version that Ozone urgently needs to be stable. If there are no further suggestions, I will proceed with the release of the rc0 version based on the branch-3.1.1_review branch prepared by Tsz-Wo (many thanks to Tsz-Wo for organizing the commits).
Best
-------------
Xinyu Tan
On 2024/09/01 06:39:49 Attila Doroszlai wrote:
> > > 1. Do we need to release the ratis-thirdparty 1.0.7 version first?
>
> > RATIS-2119. Enable reproducible builds (#52)
> > RATIS-2110. Publish SBOM artifacts (#51)
>
> Both of these are only improvements for the build, so I don't think
> it's necessary to delay 3.1.1.
>
> It would be nice to prepare the release of 1.0.7 in parallel, and
> include it in 3.1.1 only if that one needs to wait for anything else.
>
> -Attila
>