root mirror issue

[Mark McDonald_2]

Hi

I am doing an audit of some old HPUX machines that we have inherited.

lvlnboot shows:
# lvlnboot -v
Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
/dev/dsk/c2t1d0 (0/1/1/0.1.0) -- Boot Disk
/dev/dsk/c2t0d0 (0/1/1/0.0.0)
.
.
.

setboot is correct:
Primary bootpath : 0/1/1/0.1.0
Alternate bootpath : 0/1/1/0.0.0


lvdisplay -v - shows each root lv mirrored from c2t1d0 to c2t0d0


Why does the lvlnboot not show the 2nd disk as --boot?

Also I have noticed the usual boot files missing from lifls on the Alternate disk. I initially tried to copy these across with lifcp but got no space error.

I suspect the -B was missed on the pvcreate or mkboot was missed, but not sure which?

What is my best way to sort this out?

Currently I think the best option is to lv+vgreduce then start from scratch with pvcreate -fB.

[sujit kumar singh]

Hello


U can give the command as following to check if the other disk is alos bootable


#vgcfgrestore -l -n vg00
Volume Group Configuration information in "/etc/lvmconf/vg00.conf"
VG Name /dev/vg00
---- Physical volumes : 1 ----
/dev/rdisk/disk3_p2 (Bootable)

This will tell u all the disks in the VG that are bootable.

however in ur scenario as the O/P shows,

# lvlnboot -v
Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
/dev/dsk/c2t1d0 (0/1/1/0.1.0) -- Boot Disk
/dev/dsk/c2t0d0 (0/1/1/0.0.0)


the boot disk is the dosk that the system has bppted the Kernel or the OS on the last boot.

regards

Sujit

[sujit kumar singh]

hello



if u suspect that the pvcreate -B option was not used for the other disk in the option u can try reading the boot area information of the disk or the autofile content using the following command on a PARISC platform:

1)
#lifls -l /dev/rdsk/c2t0d0
this shows the boot related files present in the LIF area of the disk.

2)
#lifcp /dev/rdsk/c2t0d0:AUTO -

this will copy the contents of the auto file to the stdout that is ur console display and if the AUTO file is present as u can see in the O/P of the last command , u can see the auto strinf stored in the AUTO file in the LIF area of the disk.


So if u find both these things and also as in the previous post the O/P for the command
#vgcfgrestore -l -n /dev/vg00

get tsure that the otehr disk is also bootable and the lvlnboot -v showsthe current boot disk information.

Regards
Sujit

[Mark McDonald_2]

# vgcfgrestore -l -n vg00
Volume Group Configuration information in "/etc/lvmconf/vg00.conf"
VG Name /dev/vg00
---- Physical volumes : 2 ----
/dev/rdsk/c2t1d0 (Bootable)
/dev/rdsk/c2t0d0 (Non-bootable)


The lifls tells me the files are missing.

[likid0]

If that's the case the best thing you can do is completly redo the mirror.

lvextend -m 0 /dev/vg00/lvolx /dev/dsk/c2t0d0
vgreduce /dev/vg00 /dev/dsk/c2t0d0

And then start again, but this time with the pvcreate -B

[Mark McDonald_2]

Thanks, thats what I expected - home time now, I will leave the thread open until tomorrow incase anyone else has any input.

[sujit kumar singh]

Hi


Yes this is suggested that u do whole of the pvcreate -B and do the whole of the miroring of all the LVs in there.
as u also told that LIF area is not having any files in there.

# lifls -l /dev/rdisk/disk3_p2
volume ISL10 data size 7984 directory size 8 08/06/27 23:56:28
filename type start size implement created
===============================================================
ISL -12800 584 242 0 08/06/27 23:56:28
AUTO -12289 832 1 0 08/06/27 23:56:28
HPUX -12928 840 1024 0 08/06/27 23:56:28
PAD -12290 1864 1468 0 08/06/27 23:56:28
LABEL BIN 3336 8 0 98/01/01 00:17:20

can u see any of the files as listed in there in the lifls -l display?

If u have not an LIF area in there only option is to do a pvcreate -B and do mirroring all over.




and yes to remove the mirrors of the LV from that disk make sure that u use the command for all the LVs on that disk as

#lvreduce -m 0 /dev/vg00/lvol1 /dev/dsk/c2t0d0


and so on.

after that make sure that u do
#vgreduce /dev/vg00 /dev/dsk/c2t0d0
#lvlnboot -R
#lvlnboot -v

then again

#pvcreate -f /dev/rdsk/c2t0d0
#pvcreate -B /dev/rdsk/c2t0d0
#vgextend /dev/vg00 /dev/dsk/c2t0d0
#mkboot -l /dev/rdsk/c2t0d0
#mkboot -a "hpux /stand/vmunix -lq" /dev/rdsk/c2t0d0


Then follow the mirroring for all the LVs in the VG as

#lvextend -m 1 /dev/vg00/lvol1 /dev/dsk/c2t0d0

Repeat this for all the LVs, This can take take a lot of time.

#lvlnboot -R /dev/vg00
#lvlnboot -v

to check that u have the Second disk also included again in there.

#vi /stand/bootconf to put the entry of that disk as


l /dev/rdsk/c2t0d0


this is l that is SMALL L and NOT 1


then u have done the necessary things done


also make sure that u have the syatem setboot alternate path pointing to this disk


#setboot -a
#setboot

to see that the boot path is set properly


Regards

Sujit


check that this disk is set

[Mark McDonald_2]

Thanks - as I stated at the top setboot is correct.

Also:
lifls is only showing the LABEL
and lifcp returns with outofspace.

Does this mean that pvcreate did have the -B option?

[sujit kumar singh]

Hi


Can u try giving the following command on the disk as:


#mkboot -a "hpux /stand/vmunix -lq" /dev/rdsk/c2t0d0

and after that can u give the O/P of the command
#lifls -l /dev/rdsk/c2t0d0
#lifcp /dev/rdsk/c2t0d0:AUTO -
#vgcfgrestore -l -n vg00

the effort to do an mkboot is not going to harm the existing LVM structre on the disk


Regards

Sujit

[Mark McDonald_2]

# mkboot -a "hpux /stand/vmunix -lq" /dev/rdsk/c3t15d0
There appear to be non-boot logical volumes on this device.
Overwriting them could destroy all the data on this device
Should the logical volumes be overwritten [y/n]?


Seems a bit risky without a change in place!!!

[sujit kumar singh]

hi


Better to go with an all over pvcreate -B

that is a wise option



regards


Sujit

[sujit kumar singh]

Hi


Please go ahead with the Change Managemnet in place and with the reqd backups only in this case.

But will like to point out that u tried the mkboot on a Disk that already belongs to a DATA VG which is not bootable which is likely to throw errors like that.

Can u just try on the same Disk that u tried to do an mkboot the command #lifls -l for reading the LIF area and see if u can see a LABEL file there?


That will make clearer that a disk belonging to a DATA VG contains any LABEL in the LIF !!!!! Actually that is not supposed to have an LIF area altogether!!!!

I reiterate that better in this czase to follow the Change Management Process.

regards
Sujit

[Mark McDonald_2]

sujit - read my post a few above regarding lifls and the LABEL

[Torsten. · Accepted solution]

The lifls command must show at least the following entries:

# lifls -l /dev/rdsk/c2t2d0
volume ISL10 data size 7984 directory size 8 99/10/28 15:23:53
filename type start size implement created
===============================================================
ISL -12800 584 240 0 99/10/28 15:23:53
AUTO -12289 824 1 0 99/10/28 15:23:53
HPUX -12928 832 800 0 99/10/28 15:23:53
PAD -12290 1632 1700 0 99/10/28 15:23:54
LABEL BIN 3336 8 0 06/01/12 12:49:52

All the files were created by

#mkboot -l ...

command.

An empty LIF area looks like this:

#lifls -l /dev/rdsk/c0t6d0
volume c0t6d0 data size 29 directory size 1 06/03/06 20:28:11
filename type start size implement created
===============================================================
LABEL BIN 8 8 0 06/03/06 20:28:11


# mkboot -l /dev/rdsk/c0t6d0

#lifls -l /dev/rdsk/c0t6d0
volume c0t6d0 data size 7984 directory size 8 06/03/06 20:28:11
filename type start size implement created
===============================================================
ISL -12800 584 306 0 02/05/10 16:25:44
AUTO -12289 896 1 0 02/05/10 16:25:44
HPUX -12928 904 848 0 02/05/10 16:25:44
PAD -12290 1752 1580 0 02/05/10 16:25:45
LABEL BIN 3336 8 0 06/03/06 20:44:07

If you get errors or nothing while doing this commands, the PV was created without "-B" switch - hence no boot area.

In order to fix this, lvreduve the mirror, vgreduce the PV and start over beginning with "pvcreate -B ..."

For reference see the appendix here:

http://docs.hp.com/en/5991-1236/When_Good_Disks_Go_Bad_WP.pdf

[sujit kumar singh]

thanks Torsten,


For clarifying the thing


Regards
Sujit

[Mark McDonald_2]

Thanks Torsten. ok, pvcreate was definitely ran as I do have a lif area.

# lifls -l /dev/dsk/c2t0d0
volume c2t0d0 data size 29 directory size 1 07/03/28 11:56:48
filename type start size implement created
===============================================================
LABEL BIN 8 8 0 07/03/28 11:56:48

So I can confirm that it was the mkboot that was missed.

Mark

[sujit kumar singh]

hello


yes u r correct that seems that pvcreate -B was done but some how mkboot -l and mkboot -a "hpux /stand/vmunix -l2" was not done/


So my understanding says that u can do a mkboot with the above stated command, that will work.


If u feel not confident doing that u better should go ahead with Change Management and do an overall mirroring startting with the pvcreate -B

Regards

Sujit

[Mark McDonald_2]

This is a billing box so I will do it the safe way. Thanks

[Mark McDonald_2]

.