Spare disk - hot standby?

[Ratzie]

I am testing out my MSA30 with the 6420 Smart Array controller.
Easy peasy setup!

I wanted to test fail over, and pulled a disk.
I do get logs, and it recognizes the disk as pulled.

My question is:
Why not the automatic failover to the spare disk?
Do I have to initiate a command?
I can not find anything in the docs.
Is the spare disk a "hot" hot standby?

**************

# of Logical Drives................. 7
# of Physical Disks Configured...... 14
# of Physical Disks Detected........ 13

# RAID Size Status

0 1+0 69460 MB OK
1 1+0 69460 MB OK
2 1+0 69460 MB OK
3 1+0 69460 MB OK
4 1+0 69460 MB USING INTERIM RECOVERY MODE
5 1+0 69460 MB OK
6 0 69460 MB OK

******************
Location Ch ID Type Capacity Status

external 1 0 DISK 72.8 GB OK
external 1 1 DISK 72.8 GB OK
external 1 2 DISK 72.8 GB OK
external 1 3 DISK 72.8 GB OK
external 1 4 DISK 72.8 GB OK
external 1 5 DISK 72.8 GB OK
external 1 8 DISK 72.8 GB SPARE
external 1 15 PROCESSOR
external 2 0 DISK 72.8 GB OK
external 2 1 DISK 72.8 GB OK
external 2 2 DISK 72.8 GB OK
external 2 3 DISK 72.8 GB OK
N/A 2 4 N/A N/A FAILED
external 2 5 DISK 72.8 GB OK
external 2 7 PROCESSOR
external 2 8 DISK 72.8 GB OK

[Torsten.]

Did you assign the spare disk to *this* logical drive?

[Ratzie]

So is that the step that is missing.
I have just assigned a disk as being spare.

So when a disk fails, I need to assign it to the logical? Then it will pick it up?

[Uwe Zessin]

Yes, my thoughts, too. The SmartArray controllers do not have "global" hot spares - you need to assign them to the desired logical disks.

[Torsten.]

You should assign the spare to a logical drive *before* a disk fails ;-)

from saconfig

To add spare drive to existing logical drive

saconfig /dev/cissX -A -s

[Ratzie]

That makes it difficult when you have 7 RAID 1 logical drives, and only 1 spare disk.
But I get it...

[Torsten.]

From the manual:

...
For any fault-tolerant configuration, you can create further protection against data loss by
assigning a physical disk as an online spare (or â hot spareâ ).
...
When multiple logical drives are defined on a controller, spare disks must be assigned
to each logical drive.
...


RAID Technology Overview
HP Smart Array RAID Controllers
HP Part Number: J6369-90050

[Ratzie]

So now I am in a pickle...
I assigned the disk as a spare:
saconfig /dev/ciss4 -A 0 -s 1:8

And when I try to assign it to a logical it complains...

# saconfig /dev/ciss4 -A 4 -s 1:8


saconfig (/dev/ciss4): Error - FW command bad status (IDENTIFY_PHYSICAL_DRIVE)
saconfig (/dev/ciss4): Command completed, but error has occurred, cmd_status=1

Physical drive 148 not responding

[Torsten. · Accepted solution]

Try to assign it to another logical drive (you are testing ..) - I think you cannot assign it to an already failed drive, because the controller want to write the information to all related disks.

[Ratzie]

Nope, same thing for any logical.

[Torsten.]

IMHO the controller won't allow modifications as long as there is a failed drive. Put back the drive, assign the spare and try again.

BTW, what driver version are you currently using?

swlist |grep -i raid

[Ratzie]

RAID-00 B.11.11.01 PCI RAID; Supptd HW=A5856A
RAID-01 B.11.11.0709 RAID SA; Supptd HW=A7143A/A9890A/A9891A


That makes absolutely no sense.

If I can 6 logical drives that are RAID 1.
the chassis only holds 14 drives.
I have to have a spare pre-assigned to the logical disk. Well that is impossible since I do not have enough disks.
Can i pre-assign a spare disk to more then one logical?

[Ratzie]

And my next question, is how do I delete that spare, that I do not have assigned to a logical...

The only command I see is to delete spare from logical.

[Torsten.]

I don't know how did you assign the spare, because you did not share this information.

IMHO it makes sense to share the spare disk between several logical drives - would you agree?

[Ratzie]

In a previous reply:
"So now I am in a pickle...
I assigned the disk as a spare:
saconfig /dev/ciss4 -A 0 -s 1:8"

I guess I should have been clearer.

I am very lucky that this is a test box.

[Torsten.]

So the spare disk is already assigned to logical drive 0.

If you put back the "failed" drive you can assign this disk to your other logical drives.

Do this first and try again to "fail" a disk or "fail" a disk in logical drive 0.

[Torsten.]

Any results?

[Ratzie]

I put the failed drive in, deleted the spare.
I was under the impression, that when I assigned the spare -A 0 -s 1:8
That I was assigning it as RAID 0. Me bad.

Took how many correspondents to figure this out!
I have deleted ALL configuration and will try to break again.

I am going to try assigning to multiple logicals. From the notes I do not know if this is possible.

Will update.

[Ratzie]

So unless I am missing something, you can not assign a spare disk to multiple logicals :(

Which does not make sense.

# saconfig /dev/ciss4 -A 0 1 -s 1:8
Improper syntax

[Uwe Zessin]

It says "improper syntax", not that you cannot do that. I'm not familiar with the 'saconfig' command (I assume that -A specifies the logical disk), but I would try it with two separate commands:
# saconfig /dev/ciss4 -A 0 -s 1:8
# saconfig /dev/ciss4 -A 1 -s 1:8

[Torsten.]

One command line for each logical drive (or some very basic scrpting):

# saconfig /dev/ciss4 -A 1 -s 1:8
# saconfig /dev/ciss4 -A 2 -s 1:8
...

[Ratzie]

Ahhhhh, I have sorted it out.
As previous post, the commands need to be on separate lines.

NOTE!!!
You can not assign a spare to an already degraded logical.

So assign you spares before the system decides to take a nose dive!

[Torsten.]

Did you do the test again?

What is sautil/saconfig is telling you?