Debian Bug report logs - #598862
hdparm: Nasty messages from the kernel on boot

version graph

Package: hdparm; Maintainer for hdparm is Stephen Gran <sgran@debian.org>; Source for hdparm is src:hdparm.

Reported by: Cesare Leonardi <celeonar@gmail.com>

Date: Sat, 2 Oct 2010 17:09:04 UTC

Severity: normal

Found in version hdparm/9.32-1

Done: Michael Meskes <meskes@debian.org>

Bug is archived. No further changes may be made.

Toggle useless messages

View this report as an mbox folder, status mbox, maintainer mbox


Report forwarded to debian-bugs-dist@lists.debian.org, Stephen Gran <sgran@debian.org>:
Bug#598862; Package hdparm. (Sat, 02 Oct 2010 17:09:07 GMT) Full text and rfc822 format available.

Acknowledgement sent to Cesare Leonardi <celeonar@gmail.com>:
New Bug report received and forwarded. Copy sent to Stephen Gran <sgran@debian.org>. (Sat, 02 Oct 2010 17:09:07 GMT) Full text and rfc822 format available.

Message #5 received at submit@bugs.debian.org (full text, mbox):

From: Cesare Leonardi <celeonar@gmail.com>
To: Debian Bug Tracking System <submit@bugs.debian.org>
Cc: debian-kernel@lists.debian.org
Subject: hdparm: Nasty messages from the kernel on boot
Date: Sat, 2 Oct 2010 19:01:23 +0200
[Message part 1 (text/plain, inline)]
Package: hdparm
Version: 9.32-1
Severity: normal

The latest update brought kernel 2.6.32-24 and hdparm 9.32-1.
Rebooting with this kernel, during the initial udev activity (when it
change to KMS mode and everything is black) there was a long pause with
the hd light on but no noise from it.
The system seemed locked but after about 30 seconds an error message
where printed and the system keep on booting normally.

The error message is attached as "hd-boot-errors-1.txt" (one of the
first i've seen) and "hd-boot-errors-2-dmesg.txt" (an entire boot log).
And in fact they talk about frozen disk.
Also, from dmesg, please note the timing between the last normal message
and the errors.

First i have to say that my Samsung HM160 has less than a year and is
highly improbable faulty.
Then i have to say that the freeze is not reproducible 100% on every
boot. But in normal mode it's highly probable. Today i've tried many
reboot in recovery mode without problem, then i've reboot normally
and... PAF!

The reason i've CCed the kernel team is that this problem happen only
with 2.6.32-24 but i wasn't able to reproduce it newer kernel, in
particular with 2.6.35-rc6 i use everyday. Maybe they have a clue of
something that has changed at the ATA level.
But the fact that it's not 100% reproducible make me think a race
condition somewhere or something related to #546813. Or something
related to APM, since this release changelog talks about changes to -M
and i use this (see below).

As you noted i've customized hdparm to make APM almost disabled: this is
necessary to avoid a popular problem of frequent load/unload cycle that
with default settings would shorten my hard disk life too much:
http://www.thinkwiki.org/wiki/Problem_with_hard_drive_clicking
Please, look at the latest attachment ("hd-smartctl-errors.txt"): as you
can see the freezes are logged in the S.M.A.R.T. database and you can
see that they happened always during the APM state change.
If smart log this, i suppose it's a rather serious event for him.

If i run "hdparm -B 254 /dev/sda" on the command line it works ok.

I've fsck the disk but i haven't see data corruption. Probably because
the error occur when the drive is mounted read-only.

Hope to have forgotten nothing.
Ciao.

Cesare.


-- System Information:
Debian Release: squeeze/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (1, 'experimental')
Architecture: i386 (i686)

Kernel: Linux 2.6.32-5-686 (SMP w/1 CPU core)
Locale: LANG=en_US.utf8, LC_CTYPE=en_US.utf8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash

Versions of packages hdparm depends on:
ii  libc6                         2.11.2-6   Embedded GNU C Library: Shared lib
ii  lsb-base                      3.2-25     Linux Standard Base 3.2 init scrip

hdparm recommends no packages.

Versions of packages hdparm suggests:
pn  apmd                          <none>     (no description available)

-- Configuration Files:
/etc/default/hdparm changed:
RAID_WORKAROUND=no

/etc/hdparm.conf changed:
quiet 
/dev/disk/by-id/ata-SAMSUNG_HM160HC_S12TJF0S982076 {
    apm = 254
    #acoustic_management = fast
}


-- no debconf information
[hd-boot-errors-1.txt (text/plain, attachment)]
[hd-boot-errors-2-dmesg.txt (text/plain, attachment)]
[hd-smartctl-errors.txt (text/plain, attachment)]

Information forwarded to debian-bugs-dist@lists.debian.org, Stephen Gran <sgran@debian.org>:
Bug#598862; Package hdparm. (Fri, 08 Oct 2010 21:24:03 GMT) Full text and rfc822 format available.

Acknowledgement sent to Cesare Leonardi <celeonar@gmail.com>:
Extra info received and forwarded to list. Copy sent to Stephen Gran <sgran@debian.org>. (Fri, 08 Oct 2010 21:24:03 GMT) Full text and rfc822 format available.

Message #10 received at 598862@bugs.debian.org (full text, mbox):

From: Cesare Leonardi <celeonar@gmail.com>
To: 598862@bugs.debian.org
Subject: Re: hdparm: Nasty messages from the kernel on boot
Date: Fri, 08 Oct 2010 23:21:06 +0200
[Message part 1 (text/plain, inline)]
Some updates on this problem. I cannot understand if it's a kernel or an 
hdparm bug (excluding an improbable failing drive).

This week i've always booted the Squeeze kernel (2.6.32-24) to observe 
this bug: no errors except today, when i needed to reboot in single mode 
for some administration works and the drive freezed for some tenth of 
seconds.
The dmesg is attached (hd-boot-errors-3-dmesg.txt) together with the 
messages as i see it on the display (hd-boot-errors-3.txt).
I've also attached the "smartctl -a /dev/sda" output.

Note that:

- in dmesg the kernel messages stops at about 8.5 sec and restart with 
the ata errors at about 36.8 sec. Thus about 30 seconds during which the 
system was substantially frozen.

- The "done." after the ata1.00 messages is from udev and is the 
termination of the "Waiting for /dev to be fully populated" phase, where 
the init script do "udevadm settle".

- In the smartctl output there is a detail that the last time i've 
forgotten to mention: the "UDMA_CRC_Error_Count" raw value corresponds 
to the number of ata freeze i've seen so far. And every time it happen, 
it increase. In fact, the kernel error talks about a DMA error.

- In the smartctl output, the details about the ata errors shows that 
they are always near the "SET FEATURES [Enable APM]". In fact i use 
"apm=254" in /etc/hdparm.conf.

About a week ago, i've tried to extract the hard disk to see if there 
were problems on the connector contacts but i haven't seen nothing 
strange. I've also tried to purge hdparm and reinstalling it.

So far i was unable to find find others with the very same problem but 
it could be to the fact that i was able to trigger only with 2.6.32 and 
hparm 9.32-1. I was unable to reproduce with newer kernel.
But given that 2.6.32 will be the Squeeze kernel, i'm a bit worried.

Ciao.

Cesare.
[hd-boot-errors-3-dmesg.txt (text/plain, attachment)]
[hd-boot-errors-3.txt (text/plain, attachment)]
[hd-smartctl-errors-3.txt (text/plain, attachment)]

Information forwarded to debian-bugs-dist@lists.debian.org, Stephen Gran <sgran@debian.org>:
Bug#598862; Package hdparm. (Fri, 08 Oct 2010 21:51:06 GMT) Full text and rfc822 format available.

Acknowledgement sent to Cesare Leonardi <celeonar@gmail.com>:
Extra info received and forwarded to list. Copy sent to Stephen Gran <sgran@debian.org>. (Fri, 08 Oct 2010 21:51:06 GMT) Full text and rfc822 format available.

Message #15 received at 598862@bugs.debian.org (full text, mbox):

From: Cesare Leonardi <celeonar@gmail.com>
To: 598862@bugs.debian.org
Subject: Re: hdparm: Nasty messages from the kernel on boot
Date: Fri, 08 Oct 2010 23:46:24 +0200
I've filed a bug upstream, hoping they have some idea:
http://sourceforge.net/tracker/?func=detail&aid=3084002&group_id=136732&atid=736682

Cesare.




Information forwarded to debian-bugs-dist@lists.debian.org, Stephen Gran <sgran@debian.org>:
Bug#598862; Package hdparm. (Sat, 23 Oct 2010 00:54:03 GMT) Full text and rfc822 format available.

Acknowledgement sent to Cesare Leonardi <celeonar@gmail.com>:
Extra info received and forwarded to list. Copy sent to Stephen Gran <sgran@debian.org>. (Sat, 23 Oct 2010 00:54:03 GMT) Full text and rfc822 format available.

Message #20 received at 598862@bugs.debian.org (full text, mbox):

From: Cesare Leonardi <celeonar@gmail.com>
To: 598862@bugs.debian.org
Subject: Re: hdparm: Nasty messages from the kernel on boot
Date: Sat, 23 Oct 2010 02:51:54 +0200
As i reported upstream, today this problem happened with a kernel 
different from the Squeeze one: 2.6.36-rc6.
For the rest, all the same.

Cesare.




Information forwarded to debian-bugs-dist@lists.debian.org:
Bug#598862; Package hdparm. (Mon, 25 Oct 2010 08:15:11 GMT) Full text and rfc822 format available.

Acknowledgement sent to Stephen Gran <sgran@debian.org>:
Extra info received and forwarded to list. (Mon, 25 Oct 2010 08:15:12 GMT) Full text and rfc822 format available.

Message #25 received at 598862@bugs.debian.org (full text, mbox):

From: Stephen Gran <sgran@debian.org>
To: Cesare Leonardi <celeonar@gmail.com>, 598862@bugs.debian.org
Subject: Re: Bug#598862: hdparm: Nasty messages from the kernel on boot
Date: Mon, 25 Oct 2010 09:10:02 +0100
[Message part 1 (text/plain, inline)]
This one time, at band camp, Cesare Leonardi said:
> As i reported upstream, today this problem happened with a kernel
> different from the Squeeze one: 2.6.36-rc6.

Hi,

Did this setting you're using ever work without provoking errors from
the drive?

Cheers,
-- 
 -----------------------------------------------------------------
|   ,''`.                                            Stephen Gran |
|  : :' :                                        sgran@debian.org |
|  `. `'                        Debian user, admin, and developer |
|    `-                                     http://www.debian.org |
 -----------------------------------------------------------------
[signature.asc (application/pgp-signature, inline)]

Information forwarded to debian-bugs-dist@lists.debian.org, Stephen Gran <sgran@debian.org>:
Bug#598862; Package hdparm. (Mon, 25 Oct 2010 17:57:03 GMT) Full text and rfc822 format available.

Acknowledgement sent to Cesare Leonardi <celeonar@gmail.com>:
Extra info received and forwarded to list. Copy sent to Stephen Gran <sgran@debian.org>. (Mon, 25 Oct 2010 17:57:03 GMT) Full text and rfc822 format available.

Message #30 received at 598862@bugs.debian.org (full text, mbox):

From: Cesare Leonardi <celeonar@gmail.com>
To: Stephen Gran <sgran@debian.org>
Cc: 598862@bugs.debian.org
Subject: Re: Bug#598862: hdparm: Nasty messages from the kernel on boot
Date: Mon, 25 Oct 2010 19:52:33 +0200
On 10/25/2010 10:10 AM, Stephen Gran wrote:
> Did this setting you're using ever work without provoking errors from
> the drive?

I've seen the first errors on the first reboot after the upgrade from 
9.27-2.1 to 9.32-1.

The strange is that nobody else seems to have reported a similar 
problem, so now i've decided for another test: i've downgraded hdparm to 
9.27-2.1 and i'll test it for about a month. I'll report here if the 
errors happen the same (and so something bad has happened by coincidence 
to my drive near the upgrade period) or they will not happen anymore.

Ciao.

Cesare.




Information forwarded to debian-bugs-dist@lists.debian.org, Stephen Gran <sgran@debian.org>:
Bug#598862; Package hdparm. (Sun, 05 Dec 2010 16:03:03 GMT) Full text and rfc822 format available.

Acknowledgement sent to Cesare Leonardi <celeonar@gmail.com>:
Extra info received and forwarded to list. Copy sent to Stephen Gran <sgran@debian.org>. (Sun, 05 Dec 2010 16:03:03 GMT) Full text and rfc822 format available.

Message #35 received at 598862@bugs.debian.org (full text, mbox):

From: Cesare Leonardi <celeonar@gmail.com>
To: 598862@bugs.debian.org
Subject: Re: Bug#598862: hdparm: Nasty messages from the kernel on boot
Date: Sun, 05 Dec 2010 16:58:35 +0100
On 10/25/2010 07:52 PM, Cesare Leonardi wrote:
> The strange is that nobody else seems to have reported a similar
> problem, so now i've decided for another test: i've downgraded hdparm to
> 9.27-2.1 and i'll test it for about a month. I'll report here if the
> errors happen the same (and so something bad has happened by coincidence
> to my drive near the upgrade period) or they will not happen anymore.

After over a mounth of test with 9.27-2.1, i've seen the ATA freezes two 
times.

So now i don't know what to think: kernel? hd failure? hdparm bug 
triggered by changes in the kernel?
Upstream was completely unresponsive.

If it is a bug someone else will notice.
Game over.

Cesare.




Information forwarded to debian-bugs-dist@lists.debian.org, Stephen Gran <sgran@debian.org>:
Bug#598862; Package hdparm. (Mon, 13 Jun 2011 20:36:06 GMT) Full text and rfc822 format available.

Acknowledgement sent to Cesare Leonardi <celeonar@gmail.com>:
Extra info received and forwarded to list. Copy sent to Stephen Gran <sgran@debian.org>. (Mon, 13 Jun 2011 20:36:07 GMT) Full text and rfc822 format available.

Message #40 received at 598862@bugs.debian.org (full text, mbox):

From: Cesare Leonardi <celeonar@gmail.com>
To: 598862@bugs.debian.org
Subject: Update on this bug
Date: Mon, 13 Jun 2011 22:34:02 +0200
This bug keeps on hitting me. But in the meantime i've discovered some 
things:

- hdparm seems not responsible for this problem and this bug can be 
closed (or passed to kernel-team?).

- Looks like a problem with the firmware (drive?, laptop?) that activate 
some energy saving mechanism without notifying the acpi subsystem. You 
can find more infos, that i think are related, here:
https://wiki.archlinux.org/index.php/Samsung_N140 (SATA exceptions, 
freeze, reset)
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/285892

- The Arch link reports a Suse kernel patch that seems to solve the 
problem, but indeed it wasn't merged upstream.

- This freezes happens to a lot of people (Ubuntu link), many with 
Samsung hardware (both laptop and hard disk or a combination of these).

- I don't know any workaround, but looks like these freezes are harmless 
to the drive, even if they are logged by SMART. I haven't tried 
disabling the apm=254, because i need it. Without it the drive parks 
drive's heads too frequently, reducing its life.

- For me the freezes happens only (and not always) when hdparm set the 
above apm property. But others see freezes in different situation, as 
you can read in the links above.

Cesare.




Reply sent to Michael Meskes <meskes@debian.org>:
You have taken responsibility. (Tue, 09 Oct 2012 11:21:03 GMT) Full text and rfc822 format available.

Notification sent to Cesare Leonardi <celeonar@gmail.com>:
Bug acknowledged by developer. (Tue, 09 Oct 2012 11:21:03 GMT) Full text and rfc822 format available.

Message #45 received at 598862-done@bugs.debian.org (full text, mbox):

From: Michael Meskes <meskes@debian.org>
To: 598862-done@bugs.debian.org
Subject: Closing
Date: Tue, 9 Oct 2012 13:08:20 +0200
As the latest info in this report already says, hdparm is not repsonsible for
the problem. Therefore I don't see a reason for keeping this report open and
thus close it. If you believe it should stay open feel free to re-open and
explain your reasoning.

Michael
-- 
Michael Meskes
Michael at Fam-Meskes dot De, Michael at Meskes dot (De|Com|Net|Org)
Michael at BorussiaFan dot De, Meskes at (Debian|Postgresql) dot Org
Jabber: michael.meskes at gmail dot com
VfL Borussia! Força Barça! Go SF 49ers! Use Debian GNU/Linux, PostgreSQL



Bug archived. Request was from Debbugs Internal Request <owner@bugs.debian.org> to internal_control@bugs.debian.org. (Wed, 07 Nov 2012 07:27:16 GMT) Full text and rfc822 format available.

Send a report that this bug log contains spam.


Debian bug tracking system administrator <owner@bugs.debian.org>. Last modified: Sun Apr 20 06:22:57 2014; Machine Name: buxtehude.debian.org

Debian Bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.