[prev in list] [next in list] [prev in thread] [next in thread] 

List:       aix-l
Subject:    Re: LVM lesson
From:       Jonathan Fosburgh <syjef () MAIL ! MDANDERSON ! ORG>
Date:       2005-12-08 17:29:39
Message-ID: 200512081129.39351.syjef () mail ! mdacc ! tmc ! edu
[Download RAW message or body]

On Thursday 08 December 2005 10:05 am, SUBSCRIBE aix-l Tina N wrote:

>
> We do have multiple sharks (4 of them).  We have two FC adapters on each
> partition - each FC is connected to separate sharks.
> I have never looked at which shark the LUNS are being presented from.  I
> assumed with the way they are connected, that the redundancy
> was built in with Shark configurations.
>

You have direct connections to the Sharks?  From what you describe you have 
redundancy within the Sharks (failure of a drive will not cause data loss, 
and there are multiple internal paths to the disks) however you do not appear 
to have redundant connections.  Since you are not doing LVM mirroring, does 
leave you with a vulnerability:  Loss of a card in the AIX server or on the 
Shark, or loss of a cable will cause you to lose access to the filesystems 
contained on that Shark, which does lead to possible corruption of filesystem 
metadata.  Of course, you also can't do concurrent code loads.  

Since you are not the storage admin ... is it possible you have two Sharks 
with expansion frames.  In that scenario, what I just described may very well 
be irrelevent.  
[prev in list] [next in list] [prev in thread] [next in thread]