doubt from below thread title : replace a bad disk with new disk on the same vg . -

[madhuchakkaravarthy]

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1452068

thread asked by fizan in system administration.

there was a reply from bijessh


Since the disk is mirrored,
You can connect the hard disk and do,
#ioscan -fnC disk
#vgcfgrestore -n vgXX /dev/rdsk/cXtXdx
#vgchange -a y /dev/vgXX
#vgsync /dev/vgXX
For non mirrored disks, you have to restore the data from the backup.
#####################################

disk that was failed is allocated from storage,
so how could we find that particular failed disk is from particular vg and node,

we have eva , what we do is if a disk got failed

1.unplug the failed disk.
2.plug the new disk
3.move it the same dg.
( possiable to identify the dg)
4.automatically sync done while the disk replaced in the same dg.


my doubt is, any commands to find the vg and node for a failed disk

regards


MC

[Matti_Kurkela]

> my doubt is, any commands to find the vg and node for a failed disk

The EVA storage is managed using the "EVA Command View" software. This software can tell you which node would be affected by the disk failure.

With the Command View software package, there is a small utility "evainfo". You should install it on all HP-UX systems connected to EVA: it will allow you to see the EVA disk identifiers in HP-UX.

You can download the evainfo tool here:
(Choose your HP-UX version, then download. The instructions for use are within the Release Notes.)

http://h20000.www2.hp.com/bizsupport/TechSupport/DriverDownload.jsp?prodNameId=3439051〈=en&cc=us&prodTypeId=18964&prodSeriesId=471497&taskId=135

The download link for the evainfo tool at the HP Business Support Center was found in this thread:

http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1177260

MK

[Patrick Wallek]

When replacing a failed disk in an array like an EVA, you should not have to do much on the host side.

When you replace the failed disk the EVA should take care of automatically rebuilding all data that resides on that disk.

The systems that use the EVA should never see any outage. The systems are using LUNs that are set up from the EVA, they do NOT see individual disks.

[madhuchakkaravarthy]

hi matti

aware of evainfo tool,but any idea of using correct options to that,

while creating v disk, we will make a note of wwn no for identifying device name as
evainfo -w wwn no , so r u aware of any correct option to use with evainfo tool to find the failed disk.

since all are in production i cant try out.

regards

MC

[Patrick Wallek]

A 'vdisk' on the EVA is that the HP-UX servers it is presented to use.

This is not the same as the physical disk on the EVA. A vidsk is made up of space carved out of the EVA and can span multiple physical disks. A 'vdisk' should never fail.

I could be wrong, but I don't think you are going to see any information on failed **physical** disks via the 'evainfo' tool.

To see they physical disks in the EVA I think you have to use the CommandViewEVA tool.