LVM

[Indrajit Bhagat]

how can we judgement that particular hardisk in a system is not working and what is steps to remiove the hardisk without data loss.

[A. Clay Stephenson · Accepted solution]

You will probably see error messages in syslog.log; ioscan -fn will display NO_HW; and commands like dd if=/dev/rdsk/c1t2d0 bs=256k of=/dev/null will fail before completing.

How do you expect to not have data loss if you have a failed disk and are not mirrored?
If are the LVOL's/Volumes on the disk are mirrored then you follow the standard replacement procedure which varies with LVM/VxVM and whether or not the failed disk is a root disk. If you aren't mirrored then you better have backup.

[skt_skt]

diskinfo

if there is lvm , try pvdisplay too.

i will vague to give a generic anwer with respect to disk replace. We can give the exact stesp if we can clarify what exactly you have..

[skt_skt]

you a r a bad boy @ looking at your profile....

[Ninad_1]

Hi,

Please refer to this doc
http://docs.hp.com/en/5991-1236/When_Good_Disks_Go_Bad.pdf

This is a very nice doc, which will tell you what you are asking here.

Regards,
Ninad

[Aashique]

Hi,
Use this command:
ioscan -fnC disk

check the all HD is okk or not.

If your hard disk is not mirrored then when you will remove the bad HD all data will be lost.

Thanks & Regards

A. Aashique

[tkc]

run 'ioscan -fnC disk' to verify if the disk is responding. run 'diskinfo' command on the disk device file to confirm if the system can read info from the disk. run 'dd' command to verify all blocks on the disk are readable.

[Aneesh Mohan]

Do
ioscan -fnH Disk h/w path ---check it out whether the s.w state is NO_HW
diskinfo -v /dev/rdsk/cxtyd0 -----give exact value for x , y
dd if=/dev/rdsk/cxtyd0 of=/dev/null bs=1024k count=1000

The above steps give you the clear information about the disk which you are suspecting .

If the disk is bad as per the above analyse and if you have mirror copy of the data do replace the disk .

vgcfgrestore vg /dev/dsk/cxtyd0
vgsync vgname