Dump-definition in lvlnboot

[Maarten van Maanen]

I've just mirroring the system-disk on our K370 HPUX 11.00 with the intent to be able to use that disk for booting the system as well. Now, when I run the lvlnboot -v command I get the following output.

# lvlnboot -v
Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
/dev/dsk/c2t4d0 (10/0.4.0) -- Boot Disk
/dev/dsk/c2t5d0 (10/0.5.0) -- Boot Disk
Boot: lvol1 on: /dev/dsk/c2t4d0
/dev/dsk/c2t5d0
Root: lvol3 on: /dev/dsk/c2t4d0
/dev/dsk/c2t5d0
Swap: lvol2 on: /dev/dsk/c2t4d0
/dev/dsk/c2t5d0
No Dump Logical Volume configured

Current path "/dev/dsk/c1t0d0" is an alternate link, skip.
Current path "/dev/dsk/c1t0d1" is an alternate link, skip.
Current path "/dev/dsk/c0t1d2" is an alternate link, skip.
Current path "/dev/dsk/c0t1d3" is an alternate link, skip.

There are two things I don't understand here. Obviously, there is no dump device configured. The question is, should I do that and if so, should it be configured for both mirrored disks.

Second, I get these Current path "/dev/dsk/c1t0d0" is an alternate link, skip-messages. These disk are a part of our AutoRAID system and are not used as system disks at all. So, why do they appear in my lvlnboot -v output ???

[Victor BERRIDGE]

1)
I suppose so, because all my machine that I have installed with a mirrored root volume group shows:
Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
/dev/dsk/c0t5d0 (8/12.5.0) -- Boot Disk
/dev/dsk/c0t8d0 (8/12.8.0) -- Boot Disk
Boot: lvol1 on: /dev/dsk/c0t5d0
/dev/dsk/c0t8d0
Root: lvol3 on: /dev/dsk/c0t5d0
/dev/dsk/c0t8d0
Swap: lvol2 on: /dev/dsk/c0t5d0
/dev/dsk/c0t8d0
Dump: lvol2 on: /dev/dsk/c0t5d0, 0

Now just a question, im curious, why mirror vg00 if you are on RAID5?

[John Palmer]

Yes you should have a dump volume configured
lvlnboot -d /dev/vg00/lvol2
will do it.

You will notice that after doing that lvlnboot -v will not show that the dump volume is mirrored, this is normal.

The alternate link messages are being generated because lvlnboot analyses all your volume groups (not just vg00) and outputs a warning for any alternate link found. You can safely ignore these.

I would query your strategy of mirroring vg00 within the same Autoraid unit though. You will not be increasing the resilience of the system but you will be affecting performance!

[Kofi ARTHIABAH]

A dump volume is required if you want to be able to capture the state of the system in a crash. depending on your environment, you might not need to turn it on.

Second question - from the lvlnboot man pages:
"The lvlnboot command updates all physical volumes in the volume group."
the devices on the AutoRAID have alternate links configured and because of that, lvlnboot needs not check for the presence of logical volumes that are root boot, dump etc., so it skips those devices. Presumeably, if they were physical links, lvlnboot will scan them for the presence of root, boot, or dump logical volumes.

[CHRIS_ANORUO]

There is no problem with the alternate link, skip message. You can configure a dump device with this command:
lvlnboot -d /dev/vg00/lvol2 same volume as swap.
lvlnboot -s /dev/vg00/lvol2.

[Maarten van Maanen]

For general information:
De 12H is used for Oracle DBMS only, not for system. That's why I'm mirroring the system disk c2t4d0 to c2t5d0. These disks are not part of the 12H.
The other info is excellent. Just wonder why HP didn't include dump when the set up the system in the first place...

[John Palmer]

Sorry Maarten,

I should have looked more closely at the disk names.

They are on the same controller so it's not an ideal situation but they tend to be very reliable so it shouldn't be a problem.

Regards,
John

[Tim Malnati]

Using the command lvlnboot -d /dev/vg00/lvol2 is correct where placing the dump on the swap volume is the preferred configuration.

I wonder about the comments that suggest you are mirroring raid 5. Your original post said you are not and the device addresses (10/0.4.0 and 10/0.5.0) suggest that they are internal drives on probably something like a K box. But I do note that both mirrors are using the same controller which will work but is not the best in both terms of performance and failure resistance.

I also question why you are receiving these alternate link messages. What I know about an autoraid unit is limited at best where I don't have one and probably never will. But I have used EMC drives all over the place and they have alternate links normally defined. I have never seen any error messages in any of these configurations mentioning anything about alternate link definitions. As a policy we don't use the EMC drives for vg00 (we usually have a mirrored Jamaica using two SCSI channels), so they are not part of the boot picture, but then again, neither is your autoraid.