Re: Catastrophic disk failure

According to DST, all eight drives are about 58%, but that's not the Host LPAR. Last we documented it (in July) we were at 68% on the Host LPAR itself.
As far as I can tell, there are no more disk slots available. This box has 8 disks, totaling 1.116 TB, which was SUPPOSEDLY RAID enabled, making 976.5 GB available (one array).

So, I think you're right that I may not be able to configure RAID5 before restoring.
Sadly, I'm pretty sure I can't shrink the hosted LPAR which uses a NWSSTG device. Everything I've seen only allows me to make it bigger.

So, now it looks like I have to
1.    install the new drive
2.    restore the v7r3 system
3.    determine the minimum size that I can make the hosted LPAR (how much free space is unused.This LPAR is only used to move objects down from v7r3 to v7r1, then from v7r1 down to v5r4; no licensed program installed)
4.    do a separate SAVSYS on the hosted LPAR (which is v7r1)
5.    Install the v7r3 system again, this time configuring RAID5
6.    Create a new, smaller hosted LPAR for v7r1
7.    Restore the v7r1 OS to the hosted LPAR

Thanks, everyone, for the advice.

Best Regards,

Thomas Garvey


On 1/8/2021 10:54 AM, Mark Waterbury wrote:

Thomas,

On the primary LPAR that hosts the other LPARs, what percentage of total DASD was in use, prior to the failure?  And, how many DASD drives are in that system?  Is it "maxed out" with no available slots to add additional DASD?

If the total DASD usage was > 75% you will likely not have enough space to restore everything on the same DASD after configuring RAID-5 or RAID-6.  As others have mentioned, you will lose the equivalent of one drive for RAID-5 and two drives for RAID-6 "striping" etc.
So, you might need to restore the entire system as non-RAID protected, and then figure out how much space is used, and back up each LPAR separately, and decide what data can be moved "elsewhere" (e.g. for example move IFS data to a SAN/NAS device accessed via NFS etc., to reduce the total storage consumption to allow for RAID protection.

Hope that helps,
Mark S. Waterbury

On Friday, January 8, 2021, 10:35:47 AM EST, Thomas Garvey <tgarvey@xxxxxxxxxx> wrote:
Thanks, everyone.

I think I'm hosed on this. RAID is not enabled and the drive is failed.
Replacing the drive means whatever was on the failed drive is gone,
including whatever part of the OS it's still trying to execute in the
IPL. So, my system is corrupted.
So, we'll have to de-configure and remove the bad drive, insert the new
drive, configure it, enable RAID and load the entire system and
libraries from our system save tape.
One last question: if my system save is from a server which did NOT have
RAID enabled, will my system be restored as RAID enabled?  I mean,
what's on my tape does not have the extra parity bit, so when the system
is restored will the restore put it back with a parity bit because the
server itself is now RAID enabled? Am I making sense?

Best Regards,

Thomas Garvey


On 1/8/2021 9:12 AM, Patrik Schindler wrote:

Hello Thomas,

Am 08.01.2021 um 15:39 schrieb Thomas Garvey <tgarvey@xxxxxxxxxx>:

Why would IBM deliver a box with RAID capability and not configure it?

Misunderstanding? Or maybe there was a RAID5-Setup-Specify not being satisfied with the order? Or it’s really a customer thing, but I doubt it.

Is it possible RAID5 was our responsibility?

Honestly: Clarify with your dealer. He should be able to tell.

Can it be set up AFTER the OS and Licensed software is installed by IBM?

Yes, as long as there’s enough free space: You’ll lose one disk worth of capacity to the checksums of RAID5.

:wq! PoC