zookeeper-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Enrico Olivelli <eolive...@gmail.com>
Subject Re:
Date Mon, 28 Oct 2019 12:46:13 GMT
Il lun 28 ott 2019, 11:25 Andor Molnar <andor@apache.org> ha scritto:

> Nice work Fangmin, great detailed analysis.
>

Yep, great work.
Can you link the jira about jdk11 and SO_LINGER ?

>
> > 2. Using OpenJDK version higher than 10, which has fixed the issue in
> JDK-8207746
>
> Java 9 & 10 are both non-LTS versions which are already end-of-lifetime. I
> believe this is again a strong message against using non-LTS versions in
> production. In the ZooKeeper community probably not worth dealing with
> (developing, testing, validating) non-LTS versions in the future.
>
> Thoughts?
>

I am sorry but it is not clear how  to define 'LTS' because the 'S' part,
support, depends on vendors.

Apart from that issue, I see two sub topics:
- zookeeper client
- zookeeper server

The client runs inside a third party application, and we should give some
support to applications (or libraries) that want to use latest and greatest
java version, mostly because usually this comes with better overall
performances.

For the server the reasoning is different but in my cases it is the same:
with my application I bundle zk server and a few zk clients so if I say to
my customer that jdk13 is supported than it must be supported for each
component.

I think the best tradeoff is currently to support java 8, java 11, latest
stable (13) and try to run nightly tests on current EA



Enrico

>
> Andor
>
>
>
>
> > On 2019. Oct 28., at 2:07, Fangmin Lv <lvfangmin@gmail.com> wrote:
> >
> > 2. Using OpenJDK version higher than 10, which has fixed the issue in
> > JDK-8207746
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message