zfs discuss - Mar 2010 - Couple Questions about replacing a drive in a zpool

First a little background, I''m running b130, I have a zpool with two
Raidz1(each 4 drives, all WD RE4-GPs) "arrays" (vdev?). 
They''re in a Norco-4220 case ("home" server), which just
consists of SAS backplanes (aoc-usas-l8i ->8087->backplane->SATA
drives).  A couple of the drives are showing a number of hard/transport/media
errors (weekly scrubs are fine) and I''m guessing it can explain some of
the slower throughput I''ve been seeing lately (errors are increasing
more rapidly).

I do have a Hotspare for the zpool and I''m thinking of doing an
advanced replacement (RMA) one at a time to minimize risk/downtime.

Here''s the first question, since the current drive is working, when I
choose to replace with the Hotspare, does the current drive cease to have data
written to it?  Eg, if the hotspare fails, does a resliver need to occur on the
current drive?

Second question, is there a way to power-down the current drive while the system
is running?  I mean, by command-line, which I would think would be more graceful
than my current plan of pulling the drive.  The reason I ask is I''m
only semi-confident I know the physical layout of the (logical) drive ordering.

Last question, is there another way to see HD serial#s?  Iostat doesn''t
show the serial#s.  I''m willing to power down the system to pull the
slots I think the drives are in to retrieve the serials, it would just be nice
not to since my home network depends on a lot of the services (AD/DHCP/DNS) that
run on the server. Though, if I do have to power it down, its not the end of the
world.

Example Iostat -En output:
c8t6d0           Soft Errors: 0 Hard Errors: 160 Transport Errors: 213
Vendor: ATA      Product: WDC WD2002FYPS-0 Revision: 5G04 Serial No:
Size: 2000.40GB <2000398934016 bytes>
Media Error: 50 Device Not Ready: 0 No Device: 0 Recoverable: 0
Illegal Request: 0 Predictive Failure Analysis: 0
-- 
This message posted from opensolaris.org

> First a little background, I''m running b130, I have a
> zpool with two Raidz1(each 4 drives, all WD RE4-GPs)
> "arrays" (vdev?).  They''re in a Norco-4220 case
> ("home" server), which just consists of SAS
> backplanes (aoc-usas-l8i ->8087->backplane->SATA
> drives).  A couple of the drives are showing a number
> of hard/transport/media errors (weekly scrubs are
> fine) and I''m guessing it can explain some of the
> slower throughput I''ve been seeing lately (errors are
> increasing more rapidly).
> 
> I do have a Hotspare for the zpool and I''m thinking
> of doing an advanced replacement (RMA) one at a time
> to minimize risk/downtime.  
> 
> Here''s the first question, since the current drive is
> working, when I choose to replace with the Hotspare,
> does the current drive cease to have data written to
> it?  Eg, if the hotspare fails, does a resliver need
> to occur on the current drive?
> 
> Second question, is there a way to power-down the
> current drive while the system is running?  I mean,
> by command-line, which I would think would be more
> graceful than my current plan of pulling the drive.
> The reason I ask is I''m only semi-confident I know
> the physical layout of the (logical) drive
>  ordering.
> 
> Last question, is there another way to see HD
> serial#s?  Iostat doesn''t show the serial#s.  I''m
> willing to power down the system to pull the slots I
> think the drives are in to retrieve the serials, it
> would just be nice not to since my home network
> depends on a lot of the services (AD/DHCP/DNS) that
> run on the server. Though, if I do have to power it
> down, its not the end of the world.
> 
> Example Iostat -En output:
> c8t6d0           Soft Errors: 0 Hard Errors: 160
> Transport Errors: 213
> Vendor: ATA      Product: WDC WD2002FYPS-0 Revision:
> 5G04 Serial No:
> Size: 2000.40GB <2000398934016 bytes>
> Media Error: 50 Device Not Ready: 0 No Device: 0
> Recoverable: 0
> Illegal Request: 0 Predictive Failure Analysis: 0
Someone pointed me to a thread and to try cfgadm -v to get the serial#, but it
doesn''t work for me.  Sounds like it would be for SATA only?

However, through that thread i found smartmontools  (smartctl) and I was able to
use that to determine the serial numbers.  Fantastic!
-- 
This message posted from opensolaris.org