Disk Mirroring on a HP UNIX Cluster

[Vassilios]

Hello,
I'm interested in knowing whether a cluster has disk mirroring or not. We had to repartition the /stand directory on each node of a cluster, and the way of doing this was to use ignite. However, with ignite, when you restore using ignite, it doesnt restore the mirrored disks. How can we be sure that we have mirrored disks in the cluster, and if so, what can we do to activate them??

Thanks

[Arunvijai_4]

Some good docs which might be useful,

http://docs.hp.com/en/B3936-90079/ch02s03.html
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=610320
http://docs.hp.com/en/B3936-90073/ch02s03.html
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=774529

-Arun

[Borislav Perkov]

Hi Vassilios,

You can see it if you execute:
lvdisplay command on logical volume fo stand
# lvdisplay /dev/vg00/lvol1
--- Logical volumes ---
LV Name /dev/vg00/lvol1
VG Name /dev/vg00
LV Permission read/write
LV Status available/syncd
Mirror copies 1
Consistency Recovery MWC
Schedule parallel
LV Size (Mbytes) 300
Current LE 75
Allocated PE 150
Stripes 0
Stripe Size (Kbytes) 0
Bad block off
Allocation strict/contiguous
IO Timeout (Seconds) default

The Mirror copies parameter gives you that the volume is mirrored once.
After cloning with ignite you need to restore manualy the mirrors with lvextend command.
regards,
Borislav

[Muthukumar_5]

Use lvdisplay -v to get mirror information's.

for lvol in `bdf | awk '!/FILESYSTEM/ { print $1; }`
do
grep -i mirror $lvol
done

hth.

[Rita C Workman]

SG is software that allows for failing software within a package to continue to run on a different node. A main requirement to do this is shared disks. The RAID setting of these disks is strictly a decision made by how much mgmt is willing to spend. Whether it's mirrored RAID or RAID5.

So SG does not make disks mirrored, it simply requires that the nodes within the cluster can 'see' them.

Now...to the basics of your statement. You discuss /stand, which is a mountpoint on vg00. And VG00 is NEVER part of the SG failover disks. You can check if your disks here are mirrored by simply lvdisplay -v /dev/vgname/lvolname it will show you how many mirror copies are attached.
So use the same commands in your vg's that are part of your cluster.

Rita

[Muthukumar_5]

You can good documents,

http://docs.hp.com/ with your keyword(s).

hth.

[Devender Khatana]

Hi,

The root of a cluster node should be always mirrored and you should also find it same.

Confirm by

#vgdisplay -v /dev/vg00

The best way to remirror after restoring from ignite is to use the complete mirroring procedure. But do that in the end when you are sure that everything else was successful and only for this reason only you should not automate mirroring with ignite restore which infact you can do.

HTH,
Devender

[Vassilios]

Hi. What is the "complete mirroring procedure".

[Devender Khatana]

Hi,


Procedure for Root Disk Mirroring

The HP MirrorDisk-UX product is a prerequisite for the following procedure. You can check that MirrorDisk-UX is installed by the following command
# swlist -l fileset | grep -i mir
which should return something similar to this:
LVM.LVM-MIRROR-RUN
In this example the disk to be used as an additional mirror copy will be /dev/dsk/c0t5d0 on which as well the root as the swap logical volume will be mirrored. The primary mirror copy will be on /dev/dsk/c0t6d0.
To mirror the root file system, add a bootable LVM disk. Create a physical volume using pvcreate with the -B option.
# pvcreate -B /dev/rdsk/c0t5d0
Add the physical volume to the existing root volume group with vgextend:
# vgextend /dev/vg00 /dev/dsk/c0t5d0
Use mkboot to place boot utilities in the boot area:
# mkboot -l /dev/rdsk/c0t5d0
Use mkboot to add an AUTO file in boot LIF area:
# mkboot -a "hpux" /dev/rdsk/c0t5d0
Or, use the -lq option to allow the system to boot in the event that one of the disks is unavailable, resulting in a loss of quorum.
# mkboot -a "hpux -lq" /dev/rdsk/c0t5d0
NOTE: Use the -lq option when the root volume group contains only 2 disks. If more than 2 disks are in the root volume group, the -lq option is not needed as quorum will be maintained even if 1 disk fails.
Check the contents of the AUTO file with
# lifcp /dev/rdsk/c0t5d0:AUTO -
Alternatively, one could set the alternate boot path variable according to e.g.
# setboot -a 8/0/19/0.5.0
The hardware address of the disk that will become the mirror disk can be determined with 'ioscan -fnC' disk. To check use setboot without any arguments.
Use pvdisplay -v to the order of the logical volumes on the primary disk.
# pvdisplay -v /dev/dsk/c0t6d0
NOTE: In the following, the standard sequence lvol1, lvol2, lvol3, and so forth, is assumed.
Mirror the boot logical volume, primary swap logical volume, root logical volume and other logical volumes as needed, to the mirror disk in the correct sequence
# lvextend -m 1 /dev/vg00/lvol1 /dev/dsk/c0t5d0
# lvextend -m 1 /dev/vg00/lvol2 /dev/dsk/c0t5d0
# lvextend -m 1 /dev/vg00/lvol3 /dev/dsk/c0t5d0
# lvextend -m 1 (additional LV)
Use pvdisplay -v to the order of the logical volumes on the mirror disk to insure they are in the correct order.
# pvdisplay -v /dev/dsk/c0t5d0
Update all physical volumes in the volume group so that the logical volume become the root, boot, primary swap, or a dump volume when the system is next booted:
For HP-UX 10.20 (not updated to) and 11.X only:
# lvlnboot -r /dev/vg00/lvol3 /dev/vg00
# lvlnboot -b /dev/vg00/lvol1 /dev/vg00
# lvlnboot -s /dev/vg00/lvol2 /dev/vg00
# lvlnboot -d /dev/vg00/lvol2 /dev/vg00
For HP-UX 10.01, 10.10, and updates to 10.20 only:
# lvlnboot -r /dev/vg00/lvol1 /dev/vg00
# lvlnboot -s /dev/vg00/lvol2 /dev/vg00
# lvlnboot -d /dev/vg00/lvol2 /dev/vg00
Check with
# lvlnboot -v
Both disks must be listed for root, boot and swap definitions.
NOTE: If forgotten, it often occurs that the system has to be booted in LVM maintenance mode. If you need to know how to boot systems in LVM maintenance mode or other mode, please refer to document OZBEKBRC00000607.
In order to test the mirror configuration, shut down the system and boot from the mirror disk.
# shutdown -ry 0
The system will shut down and reboot. As the system starts to come back up, look for the message (system's output may vary):
TO INTERRUPT THE BOOT SEQUENCE, PRESS ...
Interrupt the boot sequence. This will display the Boot Admin display. At the boot_admin> prompt type:
bo Mirror_disk_hardware_address or
bo alt
if the alternate boot device has been configured. The system will prompt to interact with IPL, answer 'no', and the system will boot from the mirror root disk.
NOTE: This example includes creating a mirror copy of the primary swap logical volume. The primary swap mirror does not need to be on a specific disk or at a specific location, but it does need to be allocated on contiguous disk space. The recommended mirror policy for primary swap is to have the Mirror Write Cache and the Mirror Consistency Recovery mechanisms disabled.
When primary swap is mirrored and your primary swap device also serves as a dump area, you must make sure that Mirror Write Cache and Mirror Consistency Recovery is set to off at boot time to avoid loss of your dump. To reset these options, you will need to reboot your system in maintenance mode. Then use the lvchange(1M) command with the -M n and -c n options.
Installation of Offline Diagnostics (ODE) on Mirrored Boot Disks
The following hardware support tools are distributed on the Support Plus Media in the OnlineDiag swinstall bundle:
â ¢ Support Tools Manager (STM)
â ¢ EMS Hardware Monitors
â ¢ HP Predictive Support tools (Series 800 only)
â ¢ LIF-resident offline diagnostics (ODE)
There is an easy way to install the ODE to the mirror if you have already installed the OnlineDiag bundle.
Check the existence of the updatediaglif file:
# ll /usr/sbin/diag/lif/updatediaglif
-r--r--r-- 1 bin bin 820224 Jul 26 21:56 /usr/sbin/diag/lif/updatediaglif
List the contents of the LIF directory of the mirror:
# lifls -l /dev/rdsk/c0t5d0
volume ISL10 data size 7984 directory size 8 96/05/28 04:10:13
filename type start size implement created
===============================================================
ISL -12800 584 240 0 96/05/28 04:10:13
AUTO -12289 824 1 0 96/05/28 04:10:13
HPUX -12928 832 848 0 96/05/28 04:10:13
PAD -12290 1680 1652 0 96/05/28 04:10:14
LABEL BIN 3336 8 0 99/01/25 06:00:41
Install the ODE LIF files with mkboot and protect (-p) the existing LIF files which you detected with lifls as described above, e.g.:
# mkboot -b /usr/sbin/diag/lif/updatediaglif \
-p ISL -p AUTO -p HPUX -p PAD -p LABEL /dev/rdsk/c0t5d0
For 11.00 64-bit use the updatediaglif2 file:
# mkboot -b /usr/sbin/diag/lif/updatediaglif2 \
-p ISL -p AUTO -p HPUX -p PAD -p LABEL /dev/rdsk/c0t5d0
Check with
# lifls /dev/rdsk/c0t5d0
One will find additional LIF files, e.g.

ODE MAPFILE SYSLIB CONFIGDATA SLMOD
SLDEV SLDRIVERS SLSCSI MAPPER IOTEST
PERFVER PVCU SSINFO ISL HPUX
AUTO PAD LABEL


HTH,
Devender