PVLINK

[Steven E. Protter]

I am though certified, clueless on this issue.

I have two fiber cards on my machines, connected to a gadzooks fabric switch and a drive array, not HP.

The drive array people indicate that they support PVLINK for failover capability.

I would like to configure the system so that if one fibre card fails there is an automatic cutover to the second and no systems admin intervention is required.

Is this possible?

If so how?

Steve

[harry d brown jr]

Steve,

Add the ALTERNATE PVLINKS to the VG's using vgextend. And to load-balance across both paths alternate the primary between the two paths.

Can you post your current primary and your secondary paths?

live free or die
harry

[Steven E. Protter]

Right now Harry, I really can't because they haven't plugged the second fibre cards into the switch.

One hardware path looks like its going to be 0/4
The other 0/7

Here is an ioscan fromo the current setup. The 0/7 was at one point plugged in.

[1740#] 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 73.4GST373405LC
/dev/dsk/c1t2d0 /dev/rdsk/c1t2d0
disk 1 0/0/2/0.2.0 sdisk CLAIMED DEVICE HP 73.4GST373405LC
/dev/dsk/c2t2d0 /dev/rdsk/c2t2d0
disk 2 0/0/2/1.2.0 sdisk CLAIMED DEVICE HP DVD-ROM 304
/dev/dsk/c3t2d0 /dev/rdsk/c3t2d0
disk 4 0/4/0/0.8.0.255.6.8.0 sdisk CLAIMED DEVICE XIOTECH STOR
AGE
/dev/dsk/c6t8d0 /dev/rdsk/c6t8d0
disk 3 0/7/0/0.8.0.255.6.8.0 sdisk NO_HW DEVICE XIOTECH STOR
AGE
/dev/dsk/c5t8d0 /dev/rdsk/c5t8d0


Does this help? I am sligtly clueless in this issue as well.

[Sridhar Bhaskarla]

Hi Steve,

There is not disk at the path
0/7/0/0.8.0.255.6.8.0 right now. It was there but it has been taken out. Let them configure the Alternate Link on the switch/drive array. Once it is done, you can get it back after scanning with ioscan -f.

Once you get it back, you can do a vgextend and get the PVLINK. It's automatic and will failover and revert back after the LINK is fixed.

-Sri

[Steven E. Protter]

can someone post an example?

[Sridhar Bhaskarla]

Steve,

Let me say I created a volume group with c0t0d0. strings /etc/lvmtab would show

...
/dev/vg01
/dev/dsk/c0t0d0
...
..

#vgdisplay -v vg01
...
..
--- Physical volumes ---
PV Name /dev/dsk/c0t0d0
PV Status available
..



Later I got alternate LINK established through another controller and say the corresponding disk is c1t0d0 (only c# changes as the target (t) and the LUN (d) would remain the same)

Then I would extend the volume group with this link

vgextend vg01 /dev/dsk/c1t0d0

This will appear as alternate link in vgdisplay -v command.


#vgdisplay -v vg01
...
..
--- Physical volumes ---
PV Name /dev/dsk/c0t0d0
PV Status available
PV Name /dev/dsk/c1t0d0 Alternate Link
..


-Sri

[Byron Myers]

Steve, One way to find out which disks are laternate paths to each other, check out the following thread - look at Terry's answer. You can determine which disks are alternate paths without having to pvcreate them - good for problem analysis.
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xa2aae7613948d5118fef0090279cd0f9,00.html

[Steven E. Protter]

My pea brain molecules are beginning to understand. I'm going to take a shot at this as soon as they hook up and zone the second fibre card.

Steve

[Eugeny Brychkov]

You will be able to find all links to FC devices if you'll know how to decode hw path and which FC address device has at both loops.
disk 4 0/4/0/0.8.0.255.6.8.0 sdisk CLAIMED DEVICE XIOTECH STORAGE
disk 3 0/7/0/0.8.0.255.6.8.0 sdisk NO_HW DEVICE XIOTECH STORAGE
Cards' hw addresses are 0/4/0/0 and 0/7/0/0, 8 stands for private loop, 0.255 stands for target addressing mode, 6.8.0 stands for device alpa=104 and LUN 0 (physical disk always have LUN 0 and no subluns). For me this config looks like either both FC HBAs are in the same loop or this disk is configured to have same alpa in 2 loops (like HP Surestore FC10 enclosure)
Eugeny