Re: Ports for redundant rings

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



12.08.2014 16:33, Jan Friesse wrote:
> Vladislav,
> 
>> 12.08.2014 15:39, Jan Friesse wrote:
>>> Vladislav,
>>>
>>>> Hi,
>>>>
>>>> I've seen several times that it is recommended to shift mcast port
>>>> number setting for the second ring. But from the code (git master) I
>>>> see
>>>> that only for broadcast mode default port number is calculated
>>>> differently from others (I mean from UDPM and UDPU).
>>>> I'm in process of writing a framework which would generate
>>>> corosync.conf
>>>> for my clusters, so want to clarify, do I really need a port shift for
>>>> UDPM (I expect that for UDPU this is not needed, as sockets are
>>>> bound to
>>>> a specific IP).
>>>
>>> you are right. Port shift is strictly needed only for broadcast. For
>>> multicast you can use different multicast addresses.
>>
>> Thanks!
>> So, for UDPM pairs maddress:mport should be unique, not just mport.
>>
>> And default mcast group address is computed from a cluster name and ring
>> number, correct?
>>
> 
> right for corosync 2.x. For corosync 1.x you have to use cman.

Yep, I meant 2.x.
Thank you!

_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss




[Index of Archives]     [Linux Clusters]     [Corosync Project]     [Linux USB Devel]     [Linux Audio Users]     [Photo]     [Yosemite News]    [Yosemite Photos]    [Linux Kernel]     [Linux SCSI]     [X.Org]

  Powered by Linux