Subject: kern/14007: uncorrectable data error reading fsbn -- problems with IDE hard disk
To: None <gnats-bugs@gnats.netbsd.org>
From: None <sen@eccosys.com>
List: netbsd-bugs
Date: 09/18/2001 23:10:24
>Number: 14007
>Category: kern
>Synopsis: uncorrectable data error reading fsbn -- problems with IDE hard disk
>Confidential: no
>Severity: serious
>Priority: high
>Responsible: kern-bug-people
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Tue Sep 18 23:11:00 PDT 2001
>Closed-Date:
>Last-Modified:
>Originator: Sen Nagata
>Release: 1.5.1
>Organization:
>Environment:
NetBSD 1.5.1 (GENERIC_LAPTOP) #33: Mon Jul 2 15:56:09 CEST 2001 he@nsa.uninett.no:/usr/src/sys/arch/i386/compile/GENERIC_LAPTOP i386
>Description:
While using 1.5.2, when trying to read some files, I heard my hard
disk make some unhealthy sounding noises. Switching to a console, I
see something like:
wd0: transfer error, downgrading to Ultra-DMA mode 1
wd0(pciide0:0:0): using PIO mode 4, Ultra-DMA mode 1 (using DMA data transfers)
wd0e: uncorrectable data error reading fsbn 5489344 of 5489344-5489359 (wd0 bn 7894369; cn 8353 tn 12 sn 28), retrying
...
I've also noticed that sometimes when the disk problems occur, there
are several mode downgrade attempts -- e.g. starting from some mode ->
Ultra-DMA mode 1 PIO mode 4 -> DMA mode 2 PIO mode 4 -> PIO mode 4.
This happened to me a day or so after upgrading to 1.5.2, so I did a fresh install of 1.5.2 on to a different hard disk and a different machine, followed by transferring data (cp -pR). Bulk transfers of
data would fail occasionally, but transferring files individually
appeared to work (I interlaced copying with syncing the disk).
Today, the same thing started to occur on the new machine so I booted
a 1.5.1 kernel hoping the problem would go away, but no help there
either.
I've experienced this on a custom compiled kernel (uses the default
pciide and wd settings) as well as the GENERIC_LAPTOP kernel.
BTW, both hard disk are IBM hard disks and the machines I tried
this on are ThinkPads (600E and X20).
I hope someone else can reproduce this but searching the archives
and pr forms didn't turn up anything for me.
>How-To-Repeat:
>Fix:
>Release-Note:
>Audit-Trail:
>Unformatted: