Identify internal/external disk on k370 server? is it hot swappable?

[Lucero Perez]

I have noticed that there is one disk failed in my server K370

I dont know if this disk is hot swappable and is there internal or external?


ioscan -funC disk
Class I H/W Path Driver S/W State H/W Type Description
=======================================================================
disk 0 10/0.4.0 sdisk CLAIMED DEVICE SEAGATE ST118273WC
/dev/dsk/c0t4d0 /dev/rdsk/c0t4d0
disk 1 10/0.5.0 sdisk CLAIMED DEVICE SEAGATE ST118273WC
/dev/dsk/c0t5d0 /dev/rdsk/c0t5d0
disk 2 10/0.6.0 sdisk CLAIMED DEVICE SEAGATE ST118273WC
/dev/dsk/c0t6d0 /dev/rdsk/c0t6d0
disk 3 10/0.8.0 sdisk CLAIMED DEVICE SEAGATE ST318275LC
/dev/dsk/c0t8d0 /dev/rdsk/c0t8d0
disk 4 10/0.9.0 sdisk CLAIMED DEVICE SEAGATE ST318275LC
/dev/dsk/c0t9d0 /dev/rdsk/c0t9d0
disk 5 10/0.10.0 sdisk CLAIMED DEVICE SEAGATE ST318436LC
/dev/dsk/c0t10d0 /dev/rdsk/c0t10d0
disk 6 10/0.11.0 sdisk CLAIMED DEVICE SEAGATE ST318275LC
/dev/dsk/c0t11d0 /dev/rdsk/c0t11d0
disk 7 10/0.12.0 sdisk CLAIMED DEVICE SEAGATE ST318275LC
/dev/dsk/c0t12d0 /dev/rdsk/c0t12d0
disk 8 10/0.13.0 sdisk CLAIMED DEVICE SEAGATE ST318275LC
/dev/dsk/c0t13d0 /dev/rdsk/c0t13d0
disk 9 10/0.14.0 sdisk NO_HW DEVICE SEAGATE ST318436LC
/dev/dsk/c0t14d0 /dev/rdsk/c0t14d0

[blah2blah]

from the physical address, 10/*, these drives are on the core i/o card fw diff scsi connector. The scsi address are configurable, there is no way to tell just where a particular drive is located given the scsi address.

but, if they have been addressed according to convention, the scsi addresses 6,5,4,3 are internal. Internal drives are not hot swappable. the rest are external.

cant tell you much from the disk part number, but it does have a sca connection. Most drives with a sca connecton are hot swappable. More then likely, you have a "jamaica" disk cabinet. And, drives in a jamaica are hot swappable

[Khairy · Accepted solution]

hi,

this K class can only supports up to 4 internal disks only.

Usually the numbering will starts from 3 -6. So c0t4d0-c0t6d0 should be internal disks and not hot-plugable.

But for c0t8d0 - c0t14d0, check the server physically if its physically connected to a disk shelf look like the one i attached in this post. If it looks like it, then it is hot-plugable.
It is call HASS storage. (The disk is BLACK in colour).

To know which is the faulty disk, perform this for all disk (c0t8d0-c0t14d0) :
# dd if=/dev/rdsk/c0t13d0 of=/dev/null

When you do this, the targeted disk LED will blink. (Ctrl-C to ends).

Take note all the disk which blinks and proceed to next disk until you reach c0t14d0, the command will show error and abort for c0t14d0.
By this time you will know which c0t14d0.

Execute the following.

# rmsf -H 10/0.14.0

By this time you can take out the disk and replace it with new ones.

Hope this answer your question and really appreciate if u cld assign some points for this. Good luck!

[Lucero Perez]

Thanks so much, I checked physically that disk is external and hot swappable,
Now in order to replace I need to do some preparation?
I know that I can use vgchange and vgcfgrestore

do you have any comments?

[Khairy]

yep.

check if your vg is mirror?

post the following:
# vgdisplay -v

Rgds

[Mridul Shrivastava]

First thing to check if LV/VG is mirrored?

That actually depends for root and non-root vgs even if it is mirrored.

I would recommend you to go through the following doc:

http://docs.hp.com/en/5991-1236/When_Good_Disks_Go_Bad.pdf

[Mridul Shrivastava]

It looks like that you are new to the forum so welcome to the forum and hope you will enjoy the stay ;)

Don't forget to assign the points, to help the other guys know how much their reply helped you in resolving the issue.

[Lucero Perez]

Thanks you all, I have assigned the points and yes I am new in this forum
Also I checked manual when good disks go bad and other posts

my checklist is as follows:

1.- pull out disk from array

2.- As root:
Restore the LVM configuration information for the physical volume
vgcfgrestore -n /dev/vg07 /dev/rdsk/c0t14d0

Activate volume group
vgchange -a y /dev/vg07

Synchronize stale logical volume mirrors
vgsync /dev/vg07


How does it look?

[Patrick Wallek]

Looks fine to me.

[Torsten.]

If you are running an up-to-date 11.11 version, you need to deactivate LVM access to this disk first. Otherwise you cannot vgcfgrestore the config.

Does "pvchange" list the -a option?