On Wed, May 2, 2012 at 10:53 AM, Srinath Perera <[email protected]> wrote:
> Hi Azeez, > > I am not clear how this fits within the big picture? Did you mean to > use this only for LXC case or for general MT/LB case as well? If yes, > where we do the MT specific bit? > This can be used in any case where you receive messages on one socket, while internally Tribes is bound to another socket. There is nothing related to MT here. > > --Srinath > > On Wed, May 2, 2012 at 8:30 AM, Sanjiva Weerawarana <[email protected]> > wrote: > > This is awesome Azeez - much better than adding routes to the LB node! > > > > I've been thinking about the network address problem .. what set of IPs > to > > use for the LXCs. Using 10.* or 192.168.* requires manual intervention > > because anything we pick may be an already used network the organization > .. > > and those will then become inaccessible to any app running inside an LXC. > > > > I looked at 169.254 and it looks like that's an option. See [1][2][3] .. > its > > called a "link local" address and it seems like the right thing for us to > > use. So we'd give say 169.254.99.* for the internal boxes and have > iptables > > forward the ports to those. Looks like we can simply bring up the LXC, > have > > it pick up whatever IP in the 169.254 range and then add a default route > to > > the host IP and register itself with the host (host IP would have to be a > > boot param). > > > > [1] http://en.wikipedia.org/wiki/Private_network > > [2] http://tools.ietf.org/html/rfc5735 > > [3] http://tools.ietf.org/html/rfc3927 > > > > Sanjiva. > > > > On Tue, May 1, 2012 at 11:05 PM, Afkham Azeez <[email protected]> wrote: > >> > >> I have got this working for the well-known membership scheme. Let's test > >> it with the LXC setup tomorrow. The route addition to the LB machine is > not > >> a scalable approach since we will need to do this in all worker node > >> machines as well since the members in a cluster should be able to > >> communicate with each other. > >> > >> > >> On Tue, May 1, 2012 at 9:25 PM, Afkham Azeez <[email protected]> wrote: > >>> > >>> I just went through the Tribes code and figured out that the local > member > >>> host and bind address can be configured to be different values. It is > the > >>> localMemberHost that is advertised by Tribes while > >>> the localMemberBindAddress is the IP address used for binding the > server > >>> socket. I will make the necessary changes to Axis2 and inform this > list. My > >>> plan is to add a new optional parameter called localMemberBindAddress > to the > >>> axis2.xml's clustering section, in addition to the already existing > >>> localMemberHost. If localMemberBindAddress is not specified, we assume > that > >>> it is same as the localMemberHost. > >>> > >>> e.g. > >>> <parameter name="localMemberHost">10.100.3.67</parameter> > >>> > >>> <parameter > >>> name="localMemberBindAddress">192.168.1.121</parameter> > >>> > >>> Then we have to setup iptables on the physical machine to do port > >>> forwarding. In this model, we have to make sure that all Carbon > processes > >>> started within LXCs have unique ports. Otherwise the port forwarding > will > >>> not work. > >>> > >>> Thoughts welcome. > >>> > >>> -- > >>> Afkham Azeez > >>> Director of Architecture; WSO2, Inc.; http://wso2.com > >>> Member; Apache Software Foundation; http://www.apache.org/ > >>> > >>> email: [email protected] cell: +94 77 3320919 > >>> blog: http://blog.afkham.org > >>> twitter: http://twitter.com/afkham_azeez > >>> linked-in: http://lk.linkedin.com/in/afkhamazeez > >>> > >>> Lean . Enterprise . Middleware > >>> > >> > >> > >> > >> -- > >> Afkham Azeez > >> Director of Architecture; WSO2, Inc.; http://wso2.com > >> Member; Apache Software Foundation; http://www.apache.org/ > >> > >> email: [email protected] cell: +94 77 3320919 > >> blog: http://blog.afkham.org > >> twitter: http://twitter.com/afkham_azeez > >> linked-in: http://lk.linkedin.com/in/afkhamazeez > >> > >> Lean . Enterprise . Middleware > >> > >> > >> _______________________________________________ > >> Dev mailing list > >> [email protected] > >> http://wso2.org/cgi-bin/mailman/listinfo/dev > >> > > > > > > > > -- > > Sanjiva Weerawarana, Ph.D. > > Founder, Chairman & CEO; WSO2, Inc.; http://wso2.com/ > > email: [email protected]; phone: +94 11 763 9614; cell: +94 77 787 6880| +1 > > 650 265 8311 > > blog: http://sanjiva.weerawarana.org/ > > > > > > Lean . Enterprise . Middleware > > > > _______________________________________________ > > Dev mailing list > > [email protected] > > http://wso2.org/cgi-bin/mailman/listinfo/dev > > > > > > -- > ============================ > Srinath Perera, Ph.D. > http://www.cs.indiana.edu/~hperera/ > http://srinathsview.blogspot.com/ > -- *Afkham Azeez* Director of Architecture; WSO2, Inc.; http://wso2.com Member; Apache Software Foundation; http://www.apache.org/ * <http://www.apache.org/>** email: **[email protected]* <[email protected]>* cell: +94 77 3320919 blog: **http://blog.afkham.org* <http://blog.afkham.org>* twitter: **http://twitter.com/afkham_azeez*<http://twitter.com/afkham_azeez> * linked-in: **http://lk.linkedin.com/in/afkhamazeez* * * *Lean . Enterprise . Middleware*
_______________________________________________ Dev mailing list [email protected] http://wso2.org/cgi-bin/mailman/listinfo/dev
