Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 23 Jan 2009 11:21:48 -0700
From:      Chris Jones <chris@cjones.org>
To:        freebsd-questions@freebsd.org
Subject:   JMicron JMB363 lossage?
Message-ID:  <497A0ABC.9010600@cjones.org>
In-Reply-To: <44ljt5olwm.fsf@be-well.ilk.org>
References:  <4974BE49.2080508@cjones.org> <44ljt5olwm.fsf@be-well.ilk.org>

Next in thread | Previous in thread | Raw E-Mail | Index | Archive | Help
Now it appears that this machine fails under heavy I/O load. I was 
trying again to copy data across from one drive to the other, and I got 
a bunch of write errors. Unfortunately the kernel was in single-user 
mode, so I don't have a good record of the errors. Before I investigate 
more deeply, has anybody seen this before? Here are some (hopefully) 
relevant excerpts from dmesg:

FreeBSD 7.0-RELEASE #0: Sun Nov 16 18:08:04 MST 2008
    chris@evilmax.cjones.org:/usr/src/sys/amd64/compile/EVILMAX
...
ACPI APIC Table: <DELL   PE_SC3  >
...
acpi0: <DELL PE_SC3> on motherboard
acpi0: [ITHREAD]
acpi0: Power Button (fixed)
...
pcib0: <ACPI Host-PCI bridge> port 0xcf8-0xcff on acpi0
pci0: <ACPI PCI bus> on pcib0
...
pcib6: <ACPI PCI-PCI bridge> irq 35 at device 10.0 on pci0
pci6: <ACPI PCI bus> on pcib6
atapci1: <JMicron JMB363 SATA300 controller> port 
0xdce0-0xdce7,0xdcd8-0xdcdb,0xdce8-0xdcef,0xdcdc-0xdcdf,0xdcf0-0xdcff 
mem 0xefffe000-0xefffffff irq 35 at device 0.0 on pci6
atapci1: [ITHREAD]
atapci1: AHCI called from vendor specific driver
atapci1: AHCI Version 01.00 controller with 2 ports detected
ata4: <ATA channel 0> on atapci1
ata4: [ITHREAD]
ata5: <ATA channel 1> on atapci1
ata5: [ITHREAD]
ata6: <ATA channel 2> on atapci1
ata6: [ITHREAD]
...
ad8: 238418MB <WDC WD2500YS-18SHB1 20.06C06> at ata4-master SATA300
ad10: 238418MB <WDC WD2500YS-18SHB1 20.06C06> at ata5-master SATA300
WARNING: Expected rawoffset 0, found 63
WARNING: Expected rawoffset 0, found 63
...
Trying to mount root from ufs:/dev/ad8s1a

I was copying from ad8 to gvinum volumes on ad10, and I was alternately 
getting errors from ad8 or ad10, depending on whether I was using dump 
or tar to do the copying. It appears to only happen under heavy disk 
load, after 5-10 minutes.

Chris

Lowell Gilbert wrote:
> Chris Jones <chris@cjones.org> writes:
>
>   
>> I have an amd64 machine with two drives. I've got my system set up on
>> ad8, and I'm building a RAID0 array using gvinum that will span ad8
>> and ad10. So for now, I have partitions on ad8 plus a set of volumes
>> on ad10 that I want to move all my data to.
>>
>> Today I was doing a test move of all the data, using dump | restore
>> like this:
>>
>> # dump -0aL -f - /var | ( cd /mnt/var && restore -rf - )
>>
>> I got about a dozen messages like this:
>>
>>  DUMP: read error from /dev/ad8s1d: Bad address: [block 10992192]:
>> count=5120
>>
>> What is causing this message, and is it a cause for alarm? Here's my
>> bsdlabel for ad8s1:
>>
>> # /dev/ad8s1:
>> 8 partitions:
>> #        size   offset    fstype   [fsize bsize bps/cpg]
>>  a:  1048576        0    4.2BSD     2048 16384     8
>>  b:  4123872  1048576      swap
>>  c: 488279547        0    unused        0     0         # "raw" part,
>> don't edit
>>  d: 104857600  5172448    4.2BSD     2048 16384 28528
>>  e:  2097152 110030048    4.2BSD     2048 16384 28528
>>  f: 376152347 112127200    4.2BSD     2048 16384 28528
>>
>> Thanks in advance for any help.
>>     
>
> It could be a serious problem, but is not necessarily such.
> Can you use smarttools to query the disk firmware for *its*
> opinion?
>
>   



Want to link to this message? Use this URL: <http://docs.FreeBSD.org/cgi/mid.cgi?497A0ABC.9010600>