[dm-devel] question about output of dmsetup status

[Sandra Escandor]

I've been experimenting a bit with dmraid eventing, and I notice that
when I do the following, the result is not what I expect - but perhaps
someone can enlighten me on what is truly expected:

 

1.	Setup a RAID0+1 with dmraid
2.	Register the RAID using dm_dso_reg_tool. I also ensure that it
is properly registerd using "dm_dso_reg_tool -m".
3.	I note the major and minor number of the devices that comprise
the RAID, using "dmsetup status".
4.	I start a write onto the RAID volume using a large file (3GB).
5.	During the write process, I pull one of the RAID member disks
out so that I may simulate a RAID failure.
6.	I notice that the "dmsetup status" command shows one of the
disks as being marked disabled - which is good. I take note of the
disabled disk's major:minor numbers. 
7.	I reconnect the RAID member disk that I pulled out in step 5.
When I look for that disk's major:minor numbers using "ls -l /dev/sd*",
I don't see that major:minor number pair. 
8.	When I run "dmsetup status" again, the major:minor numbers for
the pulled disk is still the same as the one I saw in Step 3.

 

What I expected was that the major:minor number pair for the pulled disk
would correspond to the one that I saw in Step 7. Is my assumption
incorrect? The reason why I was expecting it this way was because I was
thinking that if I use dmsetup to determine the RAID status and see a
disabled drive, I can follow it back to a /dev/sd* and then get the hard
drive's serial number - then I would know which member disk to replace.
Is there a better way?

 

Thanks for clarifying,

Sandra

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/dm-devel/attachments/20110603/e37f8254/attachment.htm>