Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 24 Jun 2007 09:36:19 +0100
From:      Matthew Seaman <m.seaman@infracaninophile.co.uk>
To:        Matt Juszczak <matt@atopia.net>
Cc:        freebsd-questions@freebsd.org
Subject:   Re: Issue with freebsd 6.2 and disk drives
Message-ID:  <467E2D03.60506@infracaninophile.co.uk>
In-Reply-To: <20070623194759.X41426@saturn.atopia.net>
References:  <20070623194759.X41426@saturn.atopia.net>

next in thread | previous in thread | raw e-mail | index | archive | help
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Matt Juszczak wrote:
> Hi all,
> 
> My dedicated server company is deploying me a new FreeBSD 6.2 server. 
> The box uses SATA drives, and is an AMD Athlon box.
> 
> I was getting the following error flooding my dmesg:
> 
> ad4: TIMEOUT - WRITE_DMA retrying (1 retry left) LBA=*********
> 
> so they replaced all the hardware in the box.  Still, with entirely
> different hardware, I am getting the same error:
> 
> ad4: TIMEOUT - WRITE_DMA retrying (1 retry left) LBA=106848207
> 
> The hard drive in the new box is:
> 
> ad4: 76319MB <Seagate ST3808110AS 3.AAD> at ata2-master SATA150

I recommend installing smartmontools and running the 'long' self
test a few times on your drives -- as the SMART stuff is handled
within the drive firmware the test results won't be confused by
problems with other parts of the system.  Running smartd and
scheduling regular disk self-tests is something I strongly
recommend as good system management practice.

> Before I start telling them that this new hardware is also bad, I
> figured I would see if there are any known problems/bugs with 6.2, or
> possibly a known compatibility issue with SATA (and maybe if I should
> ask them to switch to IDE).

There's no such generic problem with SATA, either with AMD64 or
I386.  What you can run into are: (a) problems with particular
motherboard chipsets and (b) cabling troubles.

(a) For instance the chipset on this motherboard:

http://www.supermicro.com/Aplus/motherboard/Opteron1000/HT1000/H8SSL-i2.cfm

(which uses a ServerWorks HT1000 chipset) always generates
complaints that the disks are connected with 'non ATA133 cables'
(of course they are -- it's a SATA II system...) and that it
is limiting the speed to UDMA66.

Whereas this board:

http://www.supermicro.com/Aplus/motherboard/Opteron1000/MCP55/H8SMi-2.cfm

uses the nVidia MCP55 Pro Chipset which just works, although
you'll need the nfe driver from 7.0 in order to use the on-board
ethernet.

You can usually pick out the motherboard and/or chipsets used on
your hardware from /var/run/dmesg.boot, or failing that try
installing dmidecode from ports.

(b) Cabling troubles are purely down to the build quality achieved
by your supplier.  Generally with SATA cables what seems to work
best is to use cables that are as short as possible given that the
cables should have at least an inch of slack in them, and also the
cables should neither be strained, sharply kinked nor twisted.  Look
out for the practice of gathering a large bight of excess cable into
a loop and securing it with cable ties -- while not intrinsically
bad, this needs to be gone with care, as simply tightening the cable
ties too much can crush the SATA cables and interfere with signal
propagation.  Also double check that the cables are securely plugged
in at each end -- SATA cables should have a positive
click-into-place action.

	Cheers,

	Matthew

- --
Dr Matthew J Seaman MA, D.Phil.                       7 Priory Courtyard
                                                      Flat 3
PGP: http://www.infracaninophile.co.uk/pgpkey         Ramsgate
                                                      Kent, CT11 9PW
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.4 (FreeBSD)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGfi0C8Mjk52CukIwRCCi6AKCVK6q5RZAg0XhfY4toL7Qa9rSp5ACePXTV
KzonsGmq7kLFPzB44rWoamo=
=h6sN
-----END PGP SIGNATURE-----



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?467E2D03.60506>