On 07/06/2012 05:16 PM, Robert Middleswarth wrote:
On 07/06/2012 08:19 AM, Itamar Heim wrote:
> On 07/05/2012 10:08 PM, Robert Middleswarth wrote:
>> On 07/05/2012 02:19 PM, Itamar Heim wrote:
>>> On 07/05/2012 10:06 AM, Robert Middleswarth wrote:
>>>>> > 2) You can not change what interface / IP gluster uses, it will
>>>>>> always use the ovirtmgmt network. This is a weakness as many
people
>>>>>> have
>>>>>> an independent network just for storage and they can't use
them with
>>>>>> 3.1.
>>>>>
>>>>> sorry, i don't understand this one - only the management of
>>>>> gluster is
>>>>> done via this interface. you can define a different logical
>>>>> network in
>>>>> the cluster for storage, configure ip addresses for them, and define
>>>>> the mount point that way.
>>>>> (well, apart from potential bugs on network definitions in 3.1 which
>>>>> may still exist)
>>>>>
>>>>
>>>> Unless I am missing something if I setup Gluster to use anything other
>>>> then the ovirtmgmt network then I can't create volumes because the
>>>> engine uses the ovirtmgmt networks IP's and Gluster doesn't
recognize
>>>> it.
>>>
>>> engine uses ovirtmgmt to create them.
>>> you can define another network to consume them (assuming you have more
>>> than a single nic. otherwise, i'm not sure why it matters).
>>>
>> Are you suggesting that I add both Network IP into Gluster. That would
>> work but how would I know what network Gluster would use to sync up
>> with?
>
> I can think of two items:
> 1. network for hosts running VMs to communicate over - I assume will
> be based on the url of the export you provide (well, for nfs. for
> native/posix, not sure how redirection will work).
>
> maybe worth adding something similar to the 'display network' -
> telling gluster which interface should be provided for clients to
> communicate over
I use localhost:/volumename so they talk to the localhost no network
traffic between the oVirt node and the gluster systems all the traffice
is between the gluster nodes.
>
I don't think so. it only means the host uses the local gluster service
to fetch the data, then uses the network to directly access the node
which has the needed volume to run the VM.
> 2. for communication between gluster nodes for replication, etc.
>
> maybe worth defining something like 'live migration network' or
> 'storage network'[1] - telling nodes which interface they should use
> for replication between nodes.
I think in the end that is the proper fix. Defining a storage network.
Then when you go to create gluster volumes it would create them using
the IP of the storage network. Also once you get the gluster peer setup
cleaned up it would allow the nodes to be joined that way as well.
I think there are two things here:
- client to gluster communication (for both metadata and data)
- replication/sync between gluster nodes
>
> [1] both storage and live migration network are still not available
> for ovirt today, just a concept.
> display network is available today.
>
>>
>> Thanks
>> Robert
>>
>
>