[Top] [All Lists]

Re: Recover a XFS on raid -1 (linear) when one disk is broken

To: Jan Banan <b@xxxxxxxxxxxx>
Subject: Re: Recover a XFS on raid -1 (linear) when one disk is broken
From: Chris Wedgwood <cw@xxxxxxxx>
Date: Sat, 31 Jul 2004 02:12:20 -0700
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <410B4BC3.8000404@xxxxxxxxxxxx>
References: <40F6DBC1.6050909@xxxxxxxxxxxx> <20040715205910.GA9948@xxxxxxxxxxxxxxxxxxxxx> <40F9321C.7060403@xxxxxxxxxxxx> <20040717203943.GL20260@xxxxxxxxxxxxxxx> <410ADC0A.6060100@xxxxxxxxxxxx> <20040731054924.GA4748@xxxxxxxxxxxxxxxxxxxxx> <410B4BC3.8000404@xxxxxxxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
On Sat, Jul 31, 2004 at 09:35:31AM +0200, Jan Banan wrote:

> Where do I find those SMART values? (sorry for being stupid)

'smartctl -a /dev/hda' or similar.  smartctl is usually part of of the
smart-suite package.  Looking at a craptop drive here which has been
abused heavily I see:

pain:~# smartctl -a /dev/hda
smartctl version 5.32 Copyright (C) 2002-4 Bruce Allen


  1 Raw_Read_Error_Rate     0x000d   100   100   050    Pre-fail  Offline      
-       12884901912
  2 Throughput_Performance  0x0005   099   096   050    Pre-fail  Offline      
-       3210
  3 Spin_Up_Time            0x0007   100   100   050    Pre-fail  Always       
-       0
  4 Start_Stop_Count        0x0032   094   094   000    Old_age   Always       
-       6591
  5 Reallocated_Sector_Ct   0x0033   100   100   010    Pre-fail  Always       
-       7


The last line show says 7 sectors have been reallocated.  I did this
deliberately when I got some read errors by noting the location of the
errors (so I could determine which file(s) were affected) and then
writing data over those blocks to force the drive to reallocate.

It's not a perfect solution but it was quick and easy to do and I
don't have to get another craptop drive (assuming things don't get
significantly worse).


P.S. I really prefer it if you can cc' me on replies as well as the

<Prev in Thread] Current Thread [Next in Thread>