Mirroring on rx-integrity and Hp-UX versions

[Alessandro Guarda]

Dear All,
I'm very new to HP-UX, so excuse me for my dumb questions...
I'm speaking of an RX3600 with HP-UX 11.31 fully patched, with 4 internal physical SAS 146GB disks.
I've 2 questions:
1- I'm trying to understand how to mirror the boot disk (it's a single physical disk)in a way that if it fails, the mirrored one can automatically take the ownership to mantain the system and eventually boot. In the LVM pdf guide I see that 1st I have to make the 2nd disk to partecipate at the VG00 (the only one in the system), but I also see that if I'm operating on an Integrity (like my Rx-3600), in order to make the mirror bootable, I've to choose the 2nd partition of the mirror disk (/dev/rdisk/disk10_p2). But... I've 2nd partition only on my boot disk: the others haven't partitions at all, they're still raw...

Attached there is a text file with the output of some *display command to better understand.

So, what have I to do in order to have a real working and bootable mirror?


2- let's speak about HP-UX versions. I don't have clear the differences between let's say the 11.31 released in Feb 2007 _with all patches applied_ and a 11.31 released in March 2008. Have the patching (HW, quality etc etc) the faculty to bring an old installation to the current one (staying within the same HP-UX version of course)?

Many many thanks!
Alessandro

[Devender Khatana]

Hello,

You simply need to follow the mirror procedure for integrity systems. That itself include creating three partitions on the mirror disk. The procedure itself will create additional partitions and there device files.

The links mentions mirroring for all version including 11iv3.

http://docs.hp.com/en/5992-1910/5992-1910.pdf

I have not gone through the new releases so can not comment on your second question.

HTH,
Devender

[James R. Ferguson]

HI Alessandro:

With regard to your second question about the differences between the February 2007 (initial) release of 11.31 and the March 2008 update, see:

http://www.hp.com/softwarereleases/releases-media2/latest/08_03/OEUR_0803.pdf

Regards!

...JRF...

[Torsten.]

I would strongly suggest to create a hardware raid on 2 internal disks (data will be deleted) and restore or cold-install the OS onto.

If you have the default built-in sas controller, here is the procedure:

http://docs.hp.com/en/J6369-90071/index.html

[Vivek Bhatia]

Hve a look at the simplest way of creating a mirror in integrety machine.
This is my own created document..which is short and explainatory...

Thanks
-Vivek Bhatia

[Alessandro Guarda]

Thanks guys!
It took me all the afternoon, reading and understanding the documents, but finally I got the mirror working (I've tryed also to boot on the alternate disk) fine, both for Boot and also for another data disks (mounted on /var/opt/omni and /opt/omni, where I'll place the Dataprotector files and DB).

Many many thanks to all!

Just one note: in the Devender Khatana's PDF there is something not perfect:
it's pretty all referred to legacy path system (/dev/dsk/cxtydz.... instead of /dev/disk/diskxx)
And where it says:
6. Use the insf command to create device files for all the newly-created partitions:
# /usr/sbin/insf â e â H hardware_path_to_mirror_disk
if you use the "old" HW path (obtained from ioscan â fnkC disk) instead of the "new" one (obtained -if I remember correctly- from ioscan â fNkC disk) you got only the legacy special files, not the persistent paths special files.

Best regards,
Alex

[Torsten.]

Did you consider to use the hardware mirror instead of the LVM mirror?

HDD replacement could be a pain with SAS drives - with H/W mirroring it is just plug'n'play ...

[Alessandro Guarda]

Hi Torsten,
yes, I started with it but:
- I lost the system because that particular SAS controller mix the drive IDs when the FW is upgraded, and - this is REALLY bad!- there aren't references to this problem in the FW update instruction. I found them only AFTER, when the system was lost and after much time searching into Hp website.
- I want to have more control of the system and also I want to learn the OS, so "loosing time" troubleshooting LVM is not time lost but things learned :-)

Best regards,
Alessandro

[Torsten.]

Lost the system after FW update?

I cannot imagine - more details?

[Devender Khatana]

Hello Alex,

With reference to the doubt you have got on the PDF, actually we are defining the hardware path not the block device file.
The device naming convention has changed in 11iV3 from /dev/dsk/cxtydz to /dev/disk/hdiskx but the hardware path will still remain same and can be got from the ioscan output.

Like in the following output 0/0/1/1.2.0 is hardware path for disk at instance 1 and /dev/dsk/c1t2d0 is its block device file.

myprod1:/root > ioscan -fnC disk
Class I H/W Path Driver S/W State H/W Type Description
==========================================================================
disk 0 0/0/1/1.2.0 sdisk CLAIMED DEVICE HP 73.4GATLAS10K3_73_SCA
/dev/dsk/c1t2d0 /dev/rdsk/c1t2d0
disk 1 0/0/2/0.2.0 sdisk CLAIMED DEVICE HP 73.4GATLAS10K3_73_SCA
/dev/dsk/c2t2d0 /dev/rdsk/c2t2d0
disk 2 0/0/2/1.2.0 sdisk CLAIMED DEVICE HP DVD-ROM 305
/dev/dsk/c3t2d0 /dev/rdsk/c3t2d0
disk 249 0/3/0/0.1.8.0.0.0.4 sdisk CLAIMED DEVICE HP HSV200
/dev/dsk/c19t0d4 /dev/rdsk/c19t0d4

Running insf on the hardware path will create device files for all the partitions on this disk which includes your three partiotion created for making it a mirror of root.

HTH,
Devender

[Alessandro Guarda]

Hi Torsten,
here there are the details:
http://docs.hp.com/en/13875/sas_fw_update.pdf (again 30 minutes to find this white paper...). The bad thing is that this WP should be linked and in evidence in the "Installation informations" in the new FW download page, not (let's say) "hidden in the windings of HP website".

Hi Devender,
what I'm saying is that there are also 2 types of HW paths.
In my case, for example, I've:

ioscan -fNCdisk
Class I H/W Path Driver S/W State H/W Type Description
===========================================================================
disk 8 64000/0xfa00/0x0 esdisk CLAIMED DEVICE HP DG146ABAB4
disk 9 64000/0xfa00/0x1 esdisk CLAIMED DEVICE HP DG146ABAB4
disk 10 64000/0xfa00/0x2 esdisk CLAIMED DEVICE HP DG146ABAB4
disk 11 64000/0xfa00/0x3 esdisk CLAIMED DEVICE HP DG146ABAB4

OR I can have:

ioscan -fnCdisk
Class I H/W Path Driver S/W State H/W Type Description
===============================================================================
disk 4 0/4/1/0.0.0.0.0 sdisk CLAIMED DEVICE HP DG146ABAB4
/dev/dsk/c0t0d0 /dev/rdsk/c0t0d0
/dev/dsk/c0t0d0s1 /dev/rdsk/c0t0d0s1
/dev/dsk/c0t0d0s2 /dev/rdsk/c0t0d0s2
/dev/dsk/c0t0d0s3 /dev/rdsk/c0t0d0s3
disk 5 0/4/1/0.0.0.1.0 sdisk CLAIMED DEVICE HP DG146ABAB4
/dev/dsk/c0t1d0 /dev/rdsk/c0t1d0
disk 7 0/4/1/0.0.0.2.0 sdisk CLAIMED DEVICE HP DG146ABAB4
/dev/dsk/c0t2d0 /dev/rdsk/c0t2d0
/dev/dsk/c0t2d0s1 /dev/rdsk/c0t2d0s1
/dev/dsk/c0t2d0s2 /dev/rdsk/c0t2d0s2
/dev/dsk/c0t2d0s3 /dev/rdsk/c0t2d0s3
disk 6 0/4/1/0.0.0.3.0 sdisk CLAIMED DEVICE HP DG146ABAB4
/dev/dsk/c0t3d0 /dev/rdsk/c0t3d0


The difference is that if I use insf -e -H command with HW path in new format (for example 64000/0xfa00/0x0), I obtain block devices with new persistent names like "/dev/disk/disk8_p2", instead if I use insf -e -H command with HW path in legacy format (for example 0/4/1/0.0.0.0.0), I obtain only block devices with legacy names like "/dev/dsk/c0t0d0s2".

Best regards, Alessandro.

[Torsten.]

OK, it was *SAS* firmware.

Funny, you should move the disks from

A_B_C
D___E
F_G_H

to

D_C_B
A___H
G_F_E


OK, I agree - you would be a bit lost without the doc. Anyway, a vgexport/import for vg00 may help too.



There is no hint about this in the firmware package - IMHO is is bad.