Faster session timeout for crashed sessions

["Fournier, Camille F. [Tech]" <Ca...@gs.com>]

Hi all,

I haven't heard anything back from my questions on https://issues.apache.org/jira/browse/ZOOKEEPER-922 in a while. We are very keen to have this functionality here because our users really do not understand why data sticks around for a while after their connections are killed. I'd like to move forward in a direction that people are comfortable with so I don't throw away a bunch of work. Complicating this is the major changes we potentially have in the pipeline for 3.4, including 702, which will make merging any changes I write to trunk now require a serious rewrite.

So, first of all, could I get some thoughts on 922?
Secondly, I know we've been discussing a 3.4 release in mid-April. Is that still on track? Does anyone have a vague notion of when the next feature-containing release would be after 3.4? It would be very painful to have to wait for several months to get this feature in.

Any comments would be greatly appreciated.

Thanks,
Camille

RE: Faster session timeout for crashed sessions

["Fournier, Camille F. [Tech]" <Ca...@gs.com>]

OK, thanks. Let me know what you think of 922. If there's any help I can provide for 3.4 let me know.

C

-----Original Message-----
From: Mahadev Konar [mailto:mahadev@apache.org] 
Sent: Tuesday, March 22, 2011 2:35 AM
To: dev@zookeeper.apache.org
Cc: Fournier, Camille F. [Tech]
Subject: Re: Faster session timeout for crashed sessions

Hi Camille,
 As for ZOOKEEPER-922, I can take a shot at responding on it. As for
the release, I had planned for a 3.4 release for mid of april but due
to some hadoop mapreduce next work, I havent been able to focus on the
3.4 release. I think we might be delayed by a week or so. Hopefully we
can still meet a mid april timeline.
I am very keen on using 3.4 release in Hadoop MapReduce next, so would
really like to push for it.

As for 3.5 release, there havent been any discussions on it. Please
feel free to start one on the dev list in case you want to.

thanks
mahadev

On Mon, Mar 21, 2011 at 12:26 PM, Fournier, Camille F. [Tech]
<Ca...@gs.com> wrote:
> Hi all,
>
> I haven't heard anything back from my questions on https://issues.apache.org/jira/browse/ZOOKEEPER-922 in a while. We are very keen to have this functionality here because our users really do not understand why data sticks around for a while after their connections are killed. I'd like to move forward in a direction that people are comfortable with so I don't throw away a bunch of work. Complicating this is the major changes we potentially have in the pipeline for 3.4, including 702, which will make merging any changes I write to trunk now require a serious rewrite.
>
> So, first of all, could I get some thoughts on 922?
> Secondly, I know we've been discussing a 3.4 release in mid-April. Is that still on track? Does anyone have a vague notion of when the next feature-containing release would be after 3.4? It would be very painful to have to wait for several months to get this feature in.
>
> Any comments would be greatly appreciated.
>
> Thanks,
> Camille
>
>

Re: Faster session timeout for crashed sessions

[Mahadev Konar <ma...@apache.org>]

Hi Camille,
 As for ZOOKEEPER-922, I can take a shot at responding on it. As for
the release, I had planned for a 3.4 release for mid of april but due
to some hadoop mapreduce next work, I havent been able to focus on the
3.4 release. I think we might be delayed by a week or so. Hopefully we
can still meet a mid april timeline.
I am very keen on using 3.4 release in Hadoop MapReduce next, so would
really like to push for it.

As for 3.5 release, there havent been any discussions on it. Please
feel free to start one on the dev list in case you want to.

thanks
mahadev

On Mon, Mar 21, 2011 at 12:26 PM, Fournier, Camille F. [Tech]
<Ca...@gs.com> wrote:
> Hi all,
>
> I haven't heard anything back from my questions on https://issues.apache.org/jira/browse/ZOOKEEPER-922 in a while. We are very keen to have this functionality here because our users really do not understand why data sticks around for a while after their connections are killed. I'd like to move forward in a direction that people are comfortable with so I don't throw away a bunch of work. Complicating this is the major changes we potentially have in the pipeline for 3.4, including 702, which will make merging any changes I write to trunk now require a serious rewrite.
>
> So, first of all, could I get some thoughts on 922?
> Secondly, I know we've been discussing a 3.4 release in mid-April. Is that still on track? Does anyone have a vague notion of when the next feature-containing release would be after 3.4? It would be very painful to have to wait for several months to get this feature in.
>
> Any comments would be greatly appreciated.
>
> Thanks,
> Camille
>
>