Mirror of Raidz for data reliability

[Adam Nowacki]

My understanding of this setup is:
- there are 2 chasis housing disks, each with a 2 port SAS expander,
- there are 2 servers, each with a 2 port SAS HBA (1st port connected to 
1st SAS expander, 2nd port connected to 2nd SAS expander).

On 2012-05-17 13:17, George Kontostanos wrote:
> On Thu, May 17, 2012 at 1:52 PM, George Kontostanos
> <gkontos.mail at gmail.com>  wrote:
>> On Thu, May 17, 2012 at 3:29 AM, Marcelo Araujo<araujobsdport at gmail.com>  wrote:
>>> George,
>>>
>>>>
>>>> If you have 2 machines then your best bet would be HAST.
>>>
>>>
>>> So, why use ETHERNET if my Machine can see both JBOD?
>>
>> I was under the impression that you are talking about 2 different
>> physical machines.
>>
>>>> If you are worried about data replication you can always use lagg with
>>>> 2 or more interfaces.
>>>
>>>
>>> It is different than data replication. It is data protection.
>>
>> Storing the same data in 2 different locations is data protection.
>>
>> Regards
>>
>
> Ok, after reading more carefully your first post I realized what you
> are trying to do.
>
> 2 Machines, 2 different controllers. Yet interconnected. So, in a way
> both machines would be able to see both controllers.
>
> This is very interesting but there are some implications.
>
> 1) Suppose you manage to create a mirror consisted by drives on those
> different controllers. If you reboot machine #1 machine#2 might panic.
> It is not like loosing a drive, here we are loosing a controller.
>
> 2) Both machines have to be online and the pool has to be mounted
> readonly on the standby! You don't want both of them to accidentally
> write at the same pool.
>
> 3) HAST requires tcp to work therefore it is a no go. HAST also works
> in the vdev level. Therefore the resources should not be online on the
> standby server.
>
> Good luck, this is certainly very interesting.
>