Re: Configuring disk for self maintenance.

[DrFranken <midrange@xxxxxxxxxxxx>]

Here's what *I* would do in this scenario.

Once all the data is off the old drives, do indeed 'End RAID' making all 
those disks 'Non-Configured'.

Now start hot spare. You may think this is out of sequence but it's not. 
In your scenario you want 'significant' hot spares. So pick as many as 
you'd like. 3? 5? 9? IBM i doesn't care, you can make half of 'em hot 
spare if you'd like.

Now Restart RAID. As you suggest RAID stripe will now be even across 
active units. You don't need to say 'With Hot Spare' here because you 
already set those up. ALL of your hot spares will be excepted from RAID.

Now as  SSDs kick the digital bucket it will simply draw from the pool 
of hot spares.

Note that if you have these drives spread across adapter pairs you'll 
want your hot spares spread evenly as well. If failures over time are 
too high on one adapter pair you can later move hot spares as needed. 
Since they are not in use you can easily hot remove and hot add them.


        - Larry "DrFranken" Bolhuis

www.Frankeni.com
www.iDevCloud.com - Personal Development IBM i timeshare service.
www.iInTheCloud.com - Commercial IBM i Cloud Hosting.

On 7/24/2017 9:15 AM, Rob Berendt wrote:
> One advantage to stopping/starting RAID 5 is that it keeps your RAID strip
> balanced.  Unlike the crappy RAID stripe we had when money was really
> tight and we kept adding a disk at a time.  Nothing like having an 8 drive
> RAID set with striping only over 2 drives...
>
> One disadvantage is that you'll need to keep an eye on the ASP 2 threshold
> as the size will change as you remove failed drives over time.  As each
> drive is 0.75TB that can add up quickly.
>
>
> Rob Berendt