LIF corrupt or not present

[Rob Yohn]

I have a 10.20 server that I was just notified giving the following error during a make_tape_recovery. Other output follows also. How can I correct this issue? It looks like the ignite tape has been failing for quite a while now.

WARNING: Could not restore LIF volume.
"/opt/ignite/bin/save_config" -f "/var/opt/ignite/recovery/config.recover" "vg00" >> "/var/opt/ignite/logs/makrec.log1" 2
>&1

lvlnboot: LIF information corrupt or not present on "/dev/dsk/c0t5d0".

Use the "mkboot" command to initialize the LIF area.

/usr/sbin/vgcfgbackup "vg00" >> /var/opt/ignite/logs/makrec.log1 2>&1
vgcfgbackup: Unable to read the physical volume.: I/O error
vgcfgbackup: Invalid LVMREC on Physical Volume /dev/rdsk/c0t5d0


lvlnboot -v:
lvlnboot: LIF information corrupt or not present on "/dev/dsk/c0t5d0".
Use the "mkboot" command to initialize the LIF area.
Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
/dev/dsk/c0t5d0 (8/4.5.0) -- Boot Disk
/dev/dsk/c0t8d0 (8/4.8.0)
/dev/dsk/c1t5d0 (8/12.5.0)
/dev/dsk/c1t4d0 (8/12.4.0) -- Boot Disk
Boot: lvol1 on: /dev/dsk/c0t5d0
/dev/dsk/c1t4d0
Root: lvol3 on: /dev/dsk/c0t5d0
/dev/dsk/c1t4d0
Swap: lvol2 on: /dev/dsk/c0t5d0
/dev/dsk/c1t4d0
Dump: lvol2 on: /dev/dsk/c0t5d0, 0


setboot:
Primary bootpath : 8/4.5.0
Alternate bootpath : 8/12.4.0

Autoboot is ON (enabled)
Autosearch is ON (enabled)

lifls /dev/dsk/c0t5d0
read: I/O error
read_blocks(lifvol): error on 1024 byte write
lifls: Can't list /dev/dsk/c0t5d0; not a LIF volume

lifls /dev/dsk/c1t4d0
ISL AUTO HPUX PAD LABEL
ODE MAPFILE SYSLIB CONFIGDATA SLMOD
SLDEV SLDRIVERS SLSCSI MAPPER IOTEST
PERFVER PVCU SSINFO

[Torsten.]

Do as advised an re-initialize the LIF area ...

[Rob Yohn]

I am not familiar with the steps to reinitialize the LIF area. Can it be done while the system is booted from that drive?

It looks like this is the disk I am booted on that is generating this error message.

pg /stand/bootconf
# Boot Device configuration file
# This File contains information regarding the location
# of the boot LIF. It is used by the KERN-RUN fileset to
# update the boot kernel.
l /dev/dsk/c0t5d0

[Torsten.]

Do

# mkboot -l /dev/rdsk/c0t5d0

and check with lifls.

[Rob Yohn]

I issued that command and it gave the following error. I was able to watch the drive and it was being accessed during this command. No error messages in syslog and no red light on the drive.....

mkboot -l /dev/rdsk/c0t5d0
Could not open /dev/rdsk/c0t5d0: I/O error

[Rob Yohn]

diskinfo /dev/rdsk/c0t5d0
SCSI describe of /dev/rdsk/c0t5d0:
vendor: SEAGATE
product id: ST32550W
type: direct access
size: 2082636 Kbytes
bytes per sector: 512

[Tim Nelson · Accepted solution]

My first response would be that the drive is bad..

How about this to prove it or not.

dd if=/dev/rdsk/c0t5d0 of=/dev/null

Let it read through the drive, if no IO error then back to the task at hand. If you get an IO error then you need to replace the drive.

[Bill Hassell]

> mkboot -l /dev/rdsk/c0t5d0
> Could not open /dev/rdsk/c0t5d0: I/O error

Your disk at c0t5d0 is broken. You have probably been running on the mirror copy for quite a wg=hile, especially if you don't have EMS setup to notify you of disk errors and you don't look at syslog regularly.

Ignite/UX is a very useful tool to check the health of your system, especially disks and mountpoints. Ignite/UX errors mean that your system has got some problems.

[Rob Yohn]

Thanks to all that have replied. It looks like I do indeed have a failed/failing disk. Now on to the replacement procedures. Points have been assigned.

[Ganesan R]

Hi,

Replacement procedure:
=======================
Since it is 10.20 I assume that the disks are non hot swapable.

=>Power off the server
=>Replace the faulty disk
=>Boot from alt disk
=>Interupt the boot and boot with qorum override(ISL> hpux -lq
=>vgcfgrestore -n vg00
=>vgchange -a y vg00
=>vgsync
=>lvlnboot -Rv
=>lvlnboot -b
=>lvlnboot -s
=>lvlnboot -d
=>lvlnboot -r
=>lvlnboot -v

[Rob Yohn]

Ganesan, thanks for this info.

Should I do lvreduce/vgreduce of vg00 to remove the bad drive before starting these steps?

Here is output from /etc/lvmtab for vg00. c0t5d0 is the primary drive and also the drive I am having problems with. c1t4d0 is the mirror drive. It looks like I have 2 additional drives as part of vg00.

strings /etc/lvmtab
/dev/vg00
/dev/dsk/c0t5d0
/dev/dsk/c0t8d0
/dev/dsk/c1t5d0
/dev/dsk/c1t4d0

[Ganesan R]

Hi,

No need of reducing the lv's using lvreduce. then obviously vgreduce also not required.

Refer this document if you need more clarifications. it is good document.

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

[Tim Nelson]

Rob,

The clean way would be to reduce the disk but... As it is not responding to writes the lvreduce / vgreduce commands will fail.

The best answer to this is to follow the doc "when good disks go bad". A copy is attached.

[Torsten.]

Because you have hp-ux 10.20 I think your old disk (2GB) is in an old server and most likely not hot swappable - so you need to do an offline replacement anyway. Shut down the server, replace the disk and re-configure it (follow the docs-> vgcfgrestore ...).