ode-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sathwik B P <sathwik...@gmail.com>
Subject Re: Cluster implementation using Hazelcast
Date Mon, 25 May 2015 06:02:27 GMT
Hi Sudharma,

The bridge between IL and Runtime is through interfaces.

Look towards defining the interfaces for clustering component under module
bpel-api.

Initialize your clustering component in ODEServer and pass it on to the
BPELServerImpl through the org.apache.ode.bpel.engine.Contexts


Tammo,

Shouldn't we create a new project module called ode-cluster-hazelcast for
clustering implementation ?

regards,
sathwik

On Fri, May 22, 2015 at 4:56 PM, sudharma subasinghe <suba.11@cse.mrt.ac.lk>
wrote:

> Hi Tammo,
>
> Yes, it is just a configuration option. I thought put clustering code in a
> separate package and then call it from BpelServerImpl class or ODEServer
> class. What would be the appropriate class?
>
> Thank you.
>
> On 21 May 2015 at 11:40, Tammo van Lessen <tvanlessen@gmail.com> wrote:
>
> > Hi Sudharma,
> >
> > I think it is a good starting point to only use WKA, however, why do we
> > need to restrict the approach to that? Isn't this just a configuration
> > option for hazelcast and any ODE user may choose the approach they like
> > once the feature is finished?
> >
> > Best,
> >   Tammo
> >
> > On Thu, May 21, 2015 at 6:54 AM, sudharma subasinghe <
> > suba.11@cse.mrt.ac.lk> wrote:
> >
> >> Hi Tammo,
> >>
> >> I thought to use WKA implement the cluster, not multicast or aws. What
> do
> >> you think?
> >>
> >> On 20 May 2015 at 11:17, sudharma subasinghe <suba.11@cse.mrt.ac.lk>
> >> wrote:
> >>
> >>> Hi Tammo,
> >>>
> >>>
> >>> Yes, I agree with you. So I will use hazelcast.xml to configure.
> >>>
> >>> Thank you
> >>> Sudharma
> >>>
> >>> On 19 May 2015 at 16:25, Tammo van Lessen <tvanlessen@gmail.com>
> wrote:
> >>>
> >>>> Hi Sudharma,
> >>>>
> >>>> is this your own configuration file schema? What about just using the
> >>>> standard hazelcast.xml for the configuration and loading it from
> >>>> WEB-INF/conf? This would enable uses to use all hazelcast features
> (e.g.
> >>>> EC2 auto discovery) and we don't need to replicate to configuration
> >>>> documentation for the cluster config. WDYT?
> >>>>
> >>>> Best,
> >>>>   Tammo
> >>>>
> >>>> On Tue, May 19, 2015 at 8:54 AM, sudharma subasinghe <
> >>>> suba.11@cse.mrt.ac.lk> wrote:
> >>>>
> >>>>> This is the structure of cluster config file. It uses static
> >>>>> configuration
> >>>>>
> >>>>> <clusterconfig enabled=true>
> >>>>>  <membershipScheme>wka</membershipScheme>
> >>>>>
> >>>>> <!-- The host name or IP address of this member -->
> >>>>> <localMemberHost>127.0.0.1</localMemberHost>
> >>>>>
> >>>>> <!-- The TCP port of this member -->
> >>>>> <localMemberPort>4000</localMemberPort>
> >>>>>
> >>>>> <!-- The list of static or well-known members -->
> >>>>> <members>
> >>>>>     <member>
> >>>>>          <hostName>127.0.0.1</hostName>
> >>>>>          <port>4000</port>
> >>>>>     </member>
> >>>>>     <member>
> >>>>>          <hostName>127.0.0.2</hostName>
> >>>>>          <port>4001</port>
> >>>>>     </member>
> >>>>>     <member>
> >>>>>          <hostName>127.0.0.3</hostName>
> >>>>>          <port>4002</port>
> >>>>>     </member>
> >>>>> </members>
> >>>>> </clusterconfig>
> >>>>>
> >>>>> Thank you
> >>>>>
> >>>>>
> >>>>> On 18 May 2015 at 22:17, Nandika Jayawardana <jayawark@gmail.com>
> >>>>> wrote:
> >>>>>
> >>>>>> In axis2, clustering implementation supports both WKA ( well
known
> >>>>>> address
> >>>>>> based clustering ) and multicast. However, in most cases, its
the
> well
> >>>>>> known address based method that is used in production setups.
> Hence, I
> >>>>>> think its better to implement well known address based method.
Can
> >>>>>> you post
> >>>>>> your configuration file and the hazelcast based api design.
> >>>>>>
> >>>>>> Regards
> >>>>>> Nandika
> >>>>>>
> >>>>>> On Thu, May 14, 2015 at 3:16 PM, sudharma subasinghe <
> >>>>>> suba.11@cse.mrt.ac.lk>
> >>>>>> wrote:
> >>>>>>
> >>>>>> > Hi,
> >>>>>> >
> >>>>>> > This is my plan on implementing ODE clustering using hazelcast.
I
> >>>>>> thought
> >>>>>> > to use a static configuration. Each node has a config file
which
> >>>>>> stores
> >>>>>> > hostName and port number of each node in the cluster. So
each node
> >>>>>> is aware
> >>>>>> > of other nodes in the cluster and when a node start up
it will
> join
> >>>>>> to the
> >>>>>> > cluster. Please give me ideas on this.
> >>>>>> >
> >>>>>> > Thanks,
> >>>>>> > Sudharma
> >>>>>> >
> >>>>>>
> >>>>>
> >>>>>
> >>>>
> >>>>
> >>>> --
> >>>> Tammo van Lessen - http://www.taval.de
> >>>>
> >>>
> >>>
> >>
> >
> >
> > --
> > Tammo van Lessen - http://www.taval.de
> >
>

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