[Spread-users] Help with configuration file

[Aswin Almeida]

Folks,

One follow up question to this:

Is there anything wrong with specifying a localhost entry for a Spread 
Segment that defines a.b.c.255?

For example:

Spread_Segment x.2.2.255 {
        localhost           127.0.0.1    <---
        other1               128.2.2.10
        other2               128.2.2.11
}

-Aswin

At 03:50 PM 03/31/2003 -0500, Kevin Carlson wrote:
>Thanks to all who responded with assistance!
>
>Jonathan Stanton wrote:
>>
>>On Mon, Mar 31, 2003 at 11:40:18AM -0500, Kevin Carlson wrote:
>>
>>>
>>>Is there anyone that can point me to a resource that would explain the
>>>"Spread Segment" section of the configuration file other than the User's
>>>guide?  I have read and re-read the section on configuration and can't
>>>quite understand why each machine must be listed and what the different
>>>segments represent.  Yes, I am a spread newbie. ;-)  I am very familiar
>>>with other message bus technologies such as MQSeries, SwiftMQ, and other
>>>JMS implementations but this seems to be a little different.
>>>
>>>What I am attempting to do is set up a spread daemon on a single machine
>>>that other machines both on and off the local network can connect to.  I
>>>
>>
>>
>>In this case, all your Spread_Segment needs to contain is the one
>>machine that will run a daemon. For example:
>>
>>Spread_Segment  1.2.3..255:4803 {
>>
>>         myhost          1.2.3.4
>>}
>>
>>Then any spread client can connect to the daemon running on the machine
>>myhost.mydomain.com and messages it sends to other clients will be
>>forwarded to those other clients who are also connected to
>>myhost.mydomain.com.
>>
>>
>>>
>>>gather that each segment is informing the daemon where it must broadcast
>>>data to, but why the need for each machine is confusing.  Is this to
>>>
>>
>>
>>Ah. I see the confusion. In Spread the Segment defines a set of daemons
>>who will communicate together, it says nothing about where the actual
>>clients are who connect to the daemons. Commonly you will run a number
>>of daemons (maybe one on each machine that is participating with a
>>client, maybe on fewer) and so the Segment section lists all of the
>>machines that will be running a daemon process. The list of machines is
>>needed for both protecting the daemons and for the protocols to work, as
>>they need to know the entire set of "potential" daemons who may run,
>>even if not all of them are executing all the time.
>>
>>The clients actually connect individually over TCP or local IPC sockets
>>to the daemon, the daemon does not "broadcast" at the UDP level to all
>>of the clients.
>>
>>
>>>
>>>control access to the spread daemon?  Am I even on the right track?
>>>
>>
>>
>>The spread.conf doesn't itself specify any access control limits on
>>clients (i.e. restricting a client from connecting to a daemon) it only
>>specifies which daemons will listen to each other (if a daemon starts
>>running on a machine who is NOT in the Segment section, then the other
>>daemons will ignore it.) You can implement access control for the
>>clients by using the built in Access Control API. You can find
>>documentation on it in the paper:
>>
>><http://www.cnds.jhu.edu/pub/papers/cnds-2001-2.pdf>http://www.cnds.jhu.edu/pub/papers/cnds-2001-2.pdf
>>
>>Jonathan
>>
>>
>_______________________________________________ Spread-users mailing list 
>Spread-users at lists.spread.org 
>http://lists.spread.org/mailman/listinfo/spread-users