a couple of quick questions

[likid0]

first one: is it posible, that if you mirror a disk, and you do the lvextend not in order, for example first lvol3 second lvol8, it wont boot because of these? (after doing all the mkboot, etc, etc)

Second:

When you are at the ISL, i thought you could list files in the /stand, to see what kernels you have to boot from?, can you? how?

[Yang Qin_1]

1. the order used to mirror LV's (lvextend -m1 ..) will not impact "boot". As long as you use pvcreate -B to create mirror PV, go through mkboot stuff and all LV's are mirrored, your system should be able to boot from mirror disk. Even the mirror disk was not properly configured, you should be able to boot from primery boot disk.

2. ISL stand for Initial System Loader. At that point, the kernel is not loaded yet. The current kernel is /stand/vmunix and if you kernel has been modified, you probably have an old kernel /stand/vmunix.prev. Under ISL prompt you can boot to the single user mode:

ISL>hpux -is (;0)/stand/vmunix

If you want to try with old kernel

ISL>hpux -is (;0)/stand/vmunix.prev

[Ivan Krastev]

Daniel,

with "listf" and "ls" you can list available utilities like old kernel.

cheers,
ivan

[James R. Ferguson · Accepted solution]

Hi Daniel:

It is possible to create an un-bootable mirrored disk if the first three logical volumes ('lvol1', 'lvol2' and 'lvol3') of vg00 are not mirrored as the first ones, in order. In part, older IODC firmware could address the swap logical volumes if it lay beyond 2GB.

It is simply good practice to mirror in ascending logical volume order all of vg00's logical volumes.

As for your second question, if you use 'hpux ls' you can list the contents of the '/stand' filesystem to choose the kernel that you want to boot. See the manpages for 'hpux(1M)' for more information and examples.

Regards!

...JRF...

[likid0]

Ok thnx a lot, thats what i was looking for and i couldnt remember isl>hpux ls

[Ralph Grothe]

To avoid inadvertent wrong order mirroring of the boot essential first three LVs (stand, swap, root) I alway use the output of vgdisplay like such (replace Xes with indices of your mirror PV, and remove the echo from xargs)

# vgdisplay -v vg00|awk '/LV Name/{print$NF}'|xargs -n1 -ILV echo lvextend -m 1 LV /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol1 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol2 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol3 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvsnap /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol4 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol5 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol6 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol7 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol8 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol9 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol10 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvol11 /dev/dsk/cXtXdX
lvextend -m 1 /dev/vg00/lvsnap2 /dev/dsk/cXtXdX


A colleague of mine once used a script where he collected LVs from a shell glob on /dev/vg00/l*, which resulted in an unbootable mirror disk.
As you can see the globbing uses a completely misleading collating sequence (with little tunability via LC_COLLATE)

# ls -1 /dev/vg00/l*
/dev/vg00/lvol1
/dev/vg00/lvol10
/dev/vg00/lvol11
/dev/vg00/lvol2
/dev/vg00/lvol3
/dev/vg00/lvol4
/dev/vg00/lvol5
/dev/vg00/lvol6
/dev/vg00/lvol7
/dev/vg00/lvol8
/dev/vg00/lvol9
/dev/vg00/lvsnap
/dev/vg00/lvsnap2