You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by Andrew Purtell <ap...@apache.org> on 2022/10/13 21:37:01 UTC
Status of the Compaction Server WIP and next steps
There are two (competing? complimentary?) umbrella issues for the
Compaction Server work-in progress:
HBASE-24666: Offload the replication source/sink job to independent
Replication Server
HBASE-25714: Offload the compaction job to independent Compaction Server
First question: What is the relationship, if any, between these umbrellas
and their respective subtasks?
Next, if I understand correctly, the HBASE-24666 changes were committed
into master branch; while the HBASE-25714 changes were committed into a
feature branch named after the JIRA issue. Is that correct?
My understanding also is that the HBASE-24666 changes are ports of
something that is running in production somewhere. Does anyone know the
status of the HBASE-25714 changes in that regard?
Finally, what should be the path forward here? From my quick look at the
state of things, HBASE-24666 is committed to master but needs polish before
it could be considered feature complete. There are open subtasks under the
umbrella that capture some of the missing things. Some comments on the
umbrella itself also indicate the design of coprocessor support in this
compaction server implementation has not yet been considered, which would
need to be added. It would make the most sense to me to continue the
HBASE-24666 work in master branch until completed. I reserve the right to
change this depending on answers to above questions or subsequent
discussion.
--
Best regards,
Andrew
Re: Status of the Compaction Server WIP and next steps
Posted by Andrew Purtell <ap...@apache.org>.
Yeah, ok, It is obvious in retrospect, but thank you for pointing that out.
I am not sure how I got them confused. Let me make sure there are no cross
linked issues.
My interest is HBASE-25714 so will follow up there.
On Thu, Oct 13, 2022 at 11:16 PM 张铎(Duo Zhang) <pa...@gmail.com>
wrote:
> One is 'Compaction' Server, the other is 'Replication' Server, not the
> same :)
>
> Andrew Purtell <an...@gmail.com> 于2022年10月14日周五 14:07写道：
> >
> > Thanks. Will need to update the fix versions on those issues then.
> >
> > So then the question is why do we have two competing branches for the
> same feature? What is the path forward? How do we choose one over the
> other? Or is it complimentary work? (doubtful)
> >
> >
> > > On Oct 13, 2022, at 10:19 PM, 张铎 <pa...@gmail.com> wrote:
> > >
> > > I think the fix versions for the sub tasks of HBASE-24666 are
> > > incorrect. You can check the PRs, the target branch is HBASE-24666,
> > > not master. I'm not sure why some of the issues have fix versions set
> > > to 3.0.0-alpha-x...
> > >
> > > Andrew Purtell <ap...@apache.org> 于2022年10月14日周五 05:37写道：
> > >>
> > >> There are two (competing? complimentary?) umbrella issues for the
> > >> Compaction Server work-in progress:
> > >>
> > >> HBASE-24666: Offload the replication source/sink job to independent
> > >> Replication Server
> > >>
> > >> HBASE-25714: Offload the compaction job to independent Compaction
> Server
> > >>
> > >> First question: What is the relationship, if any, between these
> umbrellas
> > >> and their respective subtasks?
> > >>
> > >> Next, if I understand correctly, the HBASE-24666 changes were
> committed
> > >> into master branch; while the HBASE-25714 changes were committed into
> a
> > >> feature branch named after the JIRA issue. Is that correct?
> > >>
> > >> My understanding also is that the HBASE-24666 changes are ports of
> > >> something that is running in production somewhere. Does anyone know
> the
> > >> status of the HBASE-25714 changes in that regard?
> > >>
> > >> Finally, what should be the path forward here? From my quick look at
> the
> > >> state of things, HBASE-24666 is committed to master but needs polish
> before
> > >> it could be considered feature complete. There are open subtasks
> under the
> > >> umbrella that capture some of the missing things. Some comments on the
> > >> umbrella itself also indicate the design of coprocessor support in
> this
> > >> compaction server implementation has not yet been considered, which
> would
> > >> need to be added. It would make the most sense to me to continue the
> > >> HBASE-24666 work in master branch until completed. I reserve the
> right to
> > >> change this depending on answers to above questions or subsequent
> > >> discussion.
> > >>
> > >> --
> > >> Best regards,
> > >> Andrew
>
Re: Status of the Compaction Server WIP and next steps
Posted by "张铎(Duo Zhang)" <pa...@gmail.com>.
One is 'Compaction' Server, the other is 'Replication' Server, not the same :)
Andrew Purtell <an...@gmail.com> 于2022年10月14日周五 14:07写道：
>
> Thanks. Will need to update the fix versions on those issues then.
>
> So then the question is why do we have two competing branches for the same feature? What is the path forward? How do we choose one over the other? Or is it complimentary work? (doubtful)
>
>
> > On Oct 13, 2022, at 10:19 PM, 张铎 <pa...@gmail.com> wrote:
> >
> > I think the fix versions for the sub tasks of HBASE-24666 are
> > incorrect. You can check the PRs, the target branch is HBASE-24666,
> > not master. I'm not sure why some of the issues have fix versions set
> > to 3.0.0-alpha-x...
> >
> > Andrew Purtell <ap...@apache.org> 于2022年10月14日周五 05:37写道：
> >>
> >> There are two (competing? complimentary?) umbrella issues for the
> >> Compaction Server work-in progress:
> >>
> >> HBASE-24666: Offload the replication source/sink job to independent
> >> Replication Server
> >>
> >> HBASE-25714: Offload the compaction job to independent Compaction Server
> >>
> >> First question: What is the relationship, if any, between these umbrellas
> >> and their respective subtasks?
> >>
> >> Next, if I understand correctly, the HBASE-24666 changes were committed
> >> into master branch; while the HBASE-25714 changes were committed into a
> >> feature branch named after the JIRA issue. Is that correct?
> >>
> >> My understanding also is that the HBASE-24666 changes are ports of
> >> something that is running in production somewhere. Does anyone know the
> >> status of the HBASE-25714 changes in that regard?
> >>
> >> Finally, what should be the path forward here? From my quick look at the
> >> state of things, HBASE-24666 is committed to master but needs polish before
> >> it could be considered feature complete. There are open subtasks under the
> >> umbrella that capture some of the missing things. Some comments on the
> >> umbrella itself also indicate the design of coprocessor support in this
> >> compaction server implementation has not yet been considered, which would
> >> need to be added. It would make the most sense to me to continue the
> >> HBASE-24666 work in master branch until completed. I reserve the right to
> >> change this depending on answers to above questions or subsequent
> >> discussion.
> >>
> >> --
> >> Best regards,
> >> Andrew
Re: Status of the Compaction Server WIP and next steps
Posted by Andrew Purtell <an...@gmail.com>.
Thanks. Will need to update the fix versions on those issues then.
So then the question is why do we have two competing branches for the same feature? What is the path forward? How do we choose one over the other? Or is it complimentary work? (doubtful)
> On Oct 13, 2022, at 10:19 PM, 张铎 <pa...@gmail.com> wrote:
>
> I think the fix versions for the sub tasks of HBASE-24666 are
> incorrect. You can check the PRs, the target branch is HBASE-24666,
> not master. I'm not sure why some of the issues have fix versions set
> to 3.0.0-alpha-x...
>
> Andrew Purtell <ap...@apache.org> 于2022年10月14日周五 05:37写道：
>>
>> There are two (competing? complimentary?) umbrella issues for the
>> Compaction Server work-in progress:
>>
>> HBASE-24666: Offload the replication source/sink job to independent
>> Replication Server
>>
>> HBASE-25714: Offload the compaction job to independent Compaction Server
>>
>> First question: What is the relationship, if any, between these umbrellas
>> and their respective subtasks?
>>
>> Next, if I understand correctly, the HBASE-24666 changes were committed
>> into master branch; while the HBASE-25714 changes were committed into a
>> feature branch named after the JIRA issue. Is that correct?
>>
>> My understanding also is that the HBASE-24666 changes are ports of
>> something that is running in production somewhere. Does anyone know the
>> status of the HBASE-25714 changes in that regard?
>>
>> Finally, what should be the path forward here? From my quick look at the
>> state of things, HBASE-24666 is committed to master but needs polish before
>> it could be considered feature complete. There are open subtasks under the
>> umbrella that capture some of the missing things. Some comments on the
>> umbrella itself also indicate the design of coprocessor support in this
>> compaction server implementation has not yet been considered, which would
>> need to be added. It would make the most sense to me to continue the
>> HBASE-24666 work in master branch until completed. I reserve the right to
>> change this depending on answers to above questions or subsequent
>> discussion.
>>
>> --
>> Best regards,
>> Andrew
Re: Status of the Compaction Server WIP and next steps
Posted by "张铎(Duo Zhang)" <pa...@gmail.com>.
I think the fix versions for the sub tasks of HBASE-24666 are
incorrect. You can check the PRs, the target branch is HBASE-24666,
not master. I'm not sure why some of the issues have fix versions set
to 3.0.0-alpha-x...
Andrew Purtell <ap...@apache.org> 于2022年10月14日周五 05:37写道：
>
> There are two (competing? complimentary?) umbrella issues for the
> Compaction Server work-in progress:
>
> HBASE-24666: Offload the replication source/sink job to independent
> Replication Server
>
> HBASE-25714: Offload the compaction job to independent Compaction Server
>
> First question: What is the relationship, if any, between these umbrellas
> and their respective subtasks?
>
> Next, if I understand correctly, the HBASE-24666 changes were committed
> into master branch; while the HBASE-25714 changes were committed into a
> feature branch named after the JIRA issue. Is that correct?
>
> My understanding also is that the HBASE-24666 changes are ports of
> something that is running in production somewhere. Does anyone know the
> status of the HBASE-25714 changes in that regard?
>
> Finally, what should be the path forward here? From my quick look at the
> state of things, HBASE-24666 is committed to master but needs polish before
> it could be considered feature complete. There are open subtasks under the
> umbrella that capture some of the missing things. Some comments on the
> umbrella itself also indicate the design of coprocessor support in this
> compaction server implementation has not yet been considered, which would
> need to be added. It would make the most sense to me to continue the
> HBASE-24666 work in master branch until completed. I reserve the right to
> change this depending on answers to above questions or subsequent
> discussion.
>
> --
> Best regards,
> Andrew