mdadm / raid

[ron.peterson at yellowbank.com]

On Tue, Aug 19, 2003 at 09:53:00PM -0400, Derek Martin wrote:

> On Tue, Aug 19, 2003 at 09:31:45PM -0400, ron.peterson at yellowbank.com wrote:

> > So here's my stupid question.  I now have two nice new Dell PowerEdge
> > 2650's running software raid using mdadm.  How do I tell, short of
> > popping a disk out, that all is well and good?  I have in fact popped
> > disks out.  That's how I cloned the second machine.  But this is not
> > what I want to do when the machines are in full production.
> 
> In short, you don't.  Whenever you are implementing something new, you
> generally want to do it on a machine that is not already in production,
> precisely because you can't test it thoroughly (e.g. for failures)
> while it is supposed to be serving its customers.  Obvious exceptions
> would be if you have some sort of redundancy (such as multiple DNS
> servers, or SMTP relay machines, etc.), or the service isn't
> particularly critical...
> 
> If the service you are providing is important, then the only way to
> adequately test that this new RAID will perform as expected when a
> disk failure occurs is to simulate an actual failure, i.e. remove a
> drive.  Failing that, you just can't know for sure.  If you don't
> really care if it goes down it doesn't matter much, but then you
> probably wouldn't be putting it on RAID storage, either.

Yes.  I understand all that.  These machines are not in production.  I
can pop disks in and out willy-nilly on a whim, and it just doesn't
matter right now.  And I do.

But it will go into production.  And then I want to be able to verify
the integrity of my RAID array.  How do I do so?  That is the question.

-- 
Ron Peterson                   -o)
87 Taylor Street               /\\
Granby, MA  01033             _\_v
https://www.yellowbank.com/   ---- 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 232 bytes
Desc: not available
URL: <http://lists.blu.org/pipermail/discuss/attachments/20030819/9f177b8f/attachment.sig>