[plug] Hdd probs

[Craig Ringer]

On Thu, 2004-02-19 at 22:52, Michael Collard wrote:
> Hiya Pluggers, I write this email on my barely working system. My 80Gb
> HDD (seagate) seems to not like reading/writing properly any more, I
> suspect it has bad sectors somehow. Thought I'd take this time while my
> complete system backs up, while it is still possible, to ask if anyone
> else has had problems with these 80Gb Seagate drivers?

I have a couple of 80GB SATA Seagates, as well as 120GB models. One of
my 80GB drives may be in the process of failing:

[root at bucket root]# smartctl -d 3ware,0 -H /dev/sda
smartctl version 5.27 Copyright (C) 2002-4 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED
Please note the following marginal Attributes:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE     
UPDATED  WHEN_FAILED RAW_VALUE
  7 Seek_Error_Rate         0x000f   046   029   030    Pre-fail 
Always   In_the_past 12210694997685

... not that it really means much. After all, the other one appears to
be doing fine:
[root at bucket root]# smartctl -d 3ware,1 -a /dev/sda 2>&1 | grep Seek
  7 Seek_Error_Rate         0x000f   079   060   030    Pre-fail 
Always       -       87332504

as are the 3 120GB models:
[root at bucket root]# smartctl -d 3ware,2 -a /dev/sda 2>&1 | grep Seek
  7 Seek_Error_Rate         0x000f   085   064   030    Pre-fail 
Always       -       383946881
[root at bucket root]# smartctl -d 3ware,3 -a /dev/sda 2>&1 | grep Seek
  7 Seek_Error_Rate         0x000f   085   064   030    Pre-fail 
Always       -       383161113
[root at bucket root]# smartctl -d 3ware,4 -a /dev/sda 2>&1 | grep Seek
  7 Seek_Error_Rate         0x000f   085   064   030    Pre-fail 
Always       -       380610482

... and I haven't heard anything about high failure rates in modern
Seagate disks. When you say it has bad sectors - do you get errors in
the kernel log (`dmesg`) about it failing to read/write certain sectors?
What errors /do/ you see?

Can you send the output of:
# smartctl -a /dev/hdx 
(you may need to install smartmontools)?

I've answered a couple of posts recently about disk testing, too - check
out the archives.

Craig Ringer