Pvlink question for HPUX 11i servers.

[Gino Castoldi_2]

Hi,

Servers: HPUX 11.11 L3000 MC/SG 11.14 (2 node)
EMC DMX 800 PVlinks/Alt links.

We are migrating from a EMC DMX 800 to a DMX 1000. As part of this plan EMC has run reports and they flagged several things that need updating. Here is the items they recommend:

- "Pvlinks timeout=90"
- "BBR set to None"
- "GOLDQPK11i, GOLDBASE11i, GOLDAPPS11i,
B.11.11.0312.4"

I'm not sure what they mean about "BBR" and
setting the "PVLINKS timeout".
The third item must be OS patches.

10 points to any good answer.
TIA, Gino

[Hoefnix · Accepted solution]

Hi,

The PVlink timeout will be the setting LVM will use as timeout before fail-over to the alternate link (second path to the LUN on the SAN) when a link/disk fails.

BBR I have no idea, sorry.

3rd, yes is OS and HW related.

Regards,

Peter

[Hoefnix]

BBR can be Bad Block Relocation (also LVM related)
Just came up in my mind.

Regards,
Peter

[RAC_1]

pvlink timeout is timeout time set for pv. Can be changed with pvchnage -t command. The default is 60. man pvchange for details.

BBR-No idea. (Is it Bad Block Relocation?)

Yes third is patches. Make sure you are up tp date on patches.

Anil

[Rita C Workman]

You got the info on the timeout for your pv and how to change them.

The BBR is bad block re-alocation.
You need to address this...remember if an area of disk goes bad the O/S will re-allocate to another spot and 'mark' that area of space as bad so it can not be used. Now if HP does the re-allocation, it knows where it moved things...but HP BBR is not going to communicate that to EMC.

Generally Storage/Disk Array vendors recommend you use their technology to handle BBR. If you don't set all your logical volumes to NONE than it is possible if something goes wrong that your disk array software will not know where the data went if HP O/S handles marking and moving for BBR instances. See my point...

Hence their suggestion to set BBR to NONE.

Rgrds,
Rita

[Hoefnix]

OK, BBR is correct Bad Block Relocation, see next thread:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=184368
To set it to None:
lvchange -r N

Regards,
Peter

[Todd McDaniel_1]

Here is a slice from an internal note from our PlatSpec...



Recently HP suggests to increase the IO timeout for SAN disk array when HP help tuning some of our systems, also Storage team requests some of our SA to increase this value too for upcoming HDS disk array microcode upgrade.

IO_timeout is used by the device driver to determine how long to wait for disk transactions to complete before concluding that an IO request can not be completed (and the device is offline or unavailable). Pvdisplay command will show the current setting for IO timeout for each device (see pvdisplay example below), most of them will show "default" value, which mostly is 30 seconds. The effect of having 'default' is occasional Patrol pages about "power failure" although the disks are function fine.

By recommendation from HP and storage team, I have changed the IO timeout to 180 seconds on all disks belong to a active volume group on all your systems