RE: Cutover from 820 to i5/520 -- MIDRANGE-L

> In the past couple of weeks, it went from "yes, we
> can load from the current disks" to "maybe/not
> sure" to "no", then back to yes, and now it's
> "we'll just have to install them & see what happen".

This is from the IBM CE's?  "We'll just have to install them and see what
happens?"  That seems like a rather cavalier attitude.  Has IBM documented
this procedure?  Does IBM provide a roadmap for this scenario?  Doesn't
sound like it.  What it sounds like is that you and your shop are becoming a
convenient guinea pig to see what does and doesn't work.

Yeah, you've got two full backups.  But what if you make the switch like
you're describing and don't find problems until a few weeks out.  Problems
that are hard to pinpoint.

Sounds like it's time for me to buy some stock in Excedrin.

Whatever you end up doing, I'd be interested in knowing how it turns out.

db

> -----Original Message-----
> From: midrange-l-bounces@xxxxxxxxxxxx / Condon, Mike
> Sent: Thursday, October 21, 2004 9:49 PM
>
> We are going forward with a cutover from our current 9406/820
> (p30) to a new
> i5/520 (p10) this weekend, or if logistics don't permit, on the
> 30th. All is
> ready on the 820 side - we just upgraded it from v4r5 to v5r3 and it's all
> stable.
> It's an upgrade moving the serial number over from the 820. We're keeping
> the following, which will be transplanted from the 820 to the i5:
> - Twinax Console card (2843-001)
> - ECS card (2745-001)
> - 10/100 Ethernet (2838-001)
> - RAID Controller (284C-001/2748-001/268C-001/6B02-001)
> - 10 disks from our current ASP/RAID set (Type 6717 Size 7516gb)
>
> So the procedure looks to be:
> 1. Two full system saves.
> 2. Break raid set.
> 3. pwrdwnsys.
> 4. Move the Twinax Console card, ECS card, Ethernet, RAID controller.
> 5. Move the current drives to the new "cans".
> 6. Power up i5 and...........
>    - OS is loaded intact from current ASP devices
>    - Follow standard Restore procedures
> 7. Cleanup work with licensed programs, network, ECS testing, etc.
>
> The issue with step #5/6 is that both our reseller and our IBM
> CE's are not
> certain that we can just move the disks and load the operating system from
> our current load disk (from position 02 on the new 520) and the
> rest of the
> ASP, and then add the new disks (qty 4 15k rpm disks - don't know
> the model
> offhand) & reconfig the raid set. The other option is to move all of the
> gear and then perform the standard restore procedures.
> Regardless, we'll have to deal with Console/ECS/Network.
> So I'm not getting a real warm feeling about how uncertain they are about
> how this will pan out. I'm prepared for either scenario, but
> would of course
> like to load from our disks in place since our users will see the new
> machine much sooner.
> If any of you from experience can let me know how this will more
> than likely
>
> turn out, I'd really appreciate it. In the past couple of weeks, it went
> from "yes, we can load from the current disks" to "maybe/not
> sure" to "no",
> then back to yes, and now it's "we'll just have to install them & see what
> happen".
> On a side note, we will end up with three network ports - (1) 10/100
> (2838-001 from the old machine), (1) 10/100 onboard, and one 10/100/1000
> onboard.
> Is there a way to trunk all three together on the same IP address with
> failover & load balancing? I have all three interfaces going to separate
> switches to cover that point of failure.