Hi all,

  I run a Barrier example from ZooKeeper Tutorial(
http://zookeeper.apache.org/doc/r3.3.1/zookeeperTutorial.html#sc_barriers),
but I have a question about that does the example just enables a group of
processes to synchronize the beginning of a computation? does it not ensure
that the end of a computation? or is it a bug?

Thanks.

Shen

Hi Shen, I'm not sure why you say that it does not synchronize the end  
of the computation. The leave() method serves that purpose. Are we  
looking at the same page?

https://cwiki.apache.org/confluence/display/ZOOKEEPER/Tutorial

-Flavio

On Feb 22, 2011, at 7:53 AM, ChingShen wrote:

> Hi all,
>
>  I run a Barrier example from ZooKeeper Tutorial(
> http://zookeeper.apache.org/doc/r3.3.1/zookeeperTutorial.html#sc_barriers)
> ,
> but I have a question about that does the example just enables a  
> group of
> processes to synchronize the beginning of a computation? does it not  
> ensure
> that the end of a computation? or is it a bug?
>
> Thanks.
>
> Shen

flavio
junqueira

research scientist

[EMAIL PROTECTED]
direct +34 93-183-8828

avinguda diagonal 177, 8th floor, barcelona, 08018, es
phone (408) 349 3300    fax (408) 349 3301

Hi Flavio,

  I think I got it, I get the Barrier code from (
http://zookeeper.apache.org/doc/r3.3.1/zookeeperTutorial.html#sc_barriers)that
create a EPHEMERAL_SEQUENTIAL znode rather than EPHEMERAL znode, it's
maybe a documentation bug.

Shen

On Tue, Feb 22, 2011 at 4:24 PM, Flavio Junqueira <[EMAIL PROTECTED]> wrote:

> Hi Shen, I'm not sure why you say that it does not synchronize the end of
> the computation. The leave() method serves that purpose. Are we looking at
> the same page?
>
> https://cwiki.apache.org/confluence/display/ZOOKEEPER/Tutorial
>
> -Flavio
>
> On Feb 22, 2011, at 7:53 AM, ChingShen wrote:
>
> Hi all,
>
>  I run a Barrier example from ZooKeeper Tutorial(
> http://zookeeper.apache.org/doc/r3.3.1/zookeeperTutorial.html#sc_barriers)
> ,
> but I have a question about that does the example just enables a group of
> processes to synchronize the beginning of a computation? does it not ensure
> that the end of a computation? or is it a bug?
>
> Thanks.
>
> Shen
>
>
> *flavio*
> *junqueira*
>
> research scientist
>
> [EMAIL PROTECTED]
> direct +34 93-183-8828
>
> avinguda diagonal 177, 8th floor, barcelona, 08018, es
> phone (408) 349 3300    fax (408) 349 3301
>
>
>

The one in the tutorial is not optimized, so the way to differentiate  
the nodes is by appending the node name, and not sequential identifiers.

-Flavio

On Feb 22, 2011, at 10:07 AM, ChingShen wrote:

> Hi Flavio,
>
>   I think I got it, I get the Barrier code from (http://zookeeper.apache.org/doc/r3.3.1/zookeeperTutorial.html#sc_barriers
> ) that create a EPHEMERAL_SEQUENTIAL znode rather than EPHEMERAL  
> znode, it's maybe a documentation bug.
>
> Shen
>
> On Tue, Feb 22, 2011 at 4:24 PM, Flavio Junqueira <fpj@yahoo-
> inc.com> wrote:
> Hi Shen, I'm not sure why you say that it does not synchronize the  
> end of the computation. The leave() method serves that purpose. Are  
> we looking at the same page?
>
> https://cwiki.apache.org/confluence/display/ZOOKEEPER/Tutorial
>
> -Flavio
>
> On Feb 22, 2011, at 7:53 AM, ChingShen wrote:
>
>> Hi all,
>>
>>  I run a Barrier example from ZooKeeper Tutorial(
>> http://zookeeper.apache.org/doc/r3.3.1/zookeeperTutorial.html#sc_barriers)
>> ,
>> but I have a question about that does the example just enables a  
>> group of
>> processes to synchronize the beginning of a computation? does it  
>> not ensure
>> that the end of a computation? or is it a bug?
>>
>> Thanks.
>>
>> Shen
>
> flavio
> junqueira
>
> research scientist
>
> [EMAIL PROTECTED]
> direct +34 93-183-8828
>
> avinguda diagonal 177, 8th floor, barcelona, 08018, es
> phone (408) 349 3300    fax (408) 349 3301
>
>
>
>

flavio
junqueira

research scientist

[EMAIL PROTECTED]
direct +34 93-183-8828

avinguda diagonal 177, 8th floor, barcelona, 08018, es
phone (408) 349 3300    fax (408) 349 3301