Debian Bug report logs - #520923
live-initramfs: booting with a bad disk in the system takes forever

Package: live-initramfs; Maintainer for live-initramfs is Debian Live Project <debian-live@lists.debian.org>;

Reported by: michal.suchanek@ruk.cuni.cz

Date: Mon, 23 Mar 2009 16:18:01 UTC

Severity: normal

Done: Daniel Baumann <daniel@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, hramrach@centrum.cz, Debian Live <debian-live@lists.debian.org>:
Bug#520923; Package live-initramfs. (Mon, 23 Mar 2009 16:18:04 GMT) Full text and rfc822 format available.

Acknowledgement sent to michal.suchanek@ruk.cuni.cz:
New Bug report received and forwarded. Copy sent to hramrach@centrum.cz, Debian Live <debian-live@lists.debian.org>. (Mon, 23 Mar 2009 16:18:04 GMT) Full text and rfc822 format available.

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

From: michal.suchanek@ruk.cuni.cz
To: Debian Bug Tracking System <submit@bugs.debian.org>
Subject: live-initramfs: booting with a bad disk in the system takes forever
Date: Mon, 23 Mar 2009 17:09:03 +0100
Package: live-initramfs
Severity: normal


The initramfs scripts cause Linux to read quite a few sectors from disk
repeatedly. As a result the boot sequence is _very_ long if a disk is
bad and returns I/O errors with delay.


-- System Information:
Debian Release: 5.0
  APT prefers stable
  APT policy: (900, 'stable'), (500, 'unstable'), (1, 'experimental')
Architecture: i386 (i686)

Kernel: Linux 2.6.28-macmini (SMP w/2 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) (ignored: LC_ALL set to en_US.UTF-8)
Shell: /bin/sh linked to /bin/bash




Information forwarded to debian-bugs-dist@lists.debian.org, Debian Live <debian-live@lists.debian.org>:
Bug#520923; Package live-initramfs. (Mon, 23 Mar 2009 16:42:07 GMT) Full text and rfc822 format available.

Acknowledgement sent to Chris Lamb <lamby@debian.org>:
Extra info received and forwarded to list. Copy sent to Debian Live <debian-live@lists.debian.org>. (Mon, 23 Mar 2009 16:42:07 GMT) Full text and rfc822 format available.

Message #10 received at 520923@bugs.debian.org (full text, mbox):

From: Chris Lamb <lamby@debian.org>
To: michal.suchanek@ruk.cuni.cz, 520923@bugs.debian.org
Subject: Re: Bug#520923: live-initramfs: booting with a bad disk in the system takes forever
Date: Mon, 23 Mar 2009 16:39:36 +0000
[Message part 1 (text/plain, inline)]
michal.suchanek@ruk.cuni.cz wrote:

> The initramfs scripts cause Linux to read quite a few sectors from disk
> repeatedly. As a result the boot sequence is _very_ long if a disk is
> bad and returns I/O errors with delay.

Um, what do you suggest live-initramfs should do?


Regards,

-- 
      ,''`.
     : :'  :     Chris Lamb
     `. `'`      lamby@debian.org
       `-
[signature.asc (application/pgp-signature, attachment)]

Information forwarded to debian-bugs-dist@lists.debian.org, Debian Live <debian-live@lists.debian.org>:
Bug#520923; Package live-initramfs. (Mon, 23 Mar 2009 22:30:02 GMT) Full text and rfc822 format available.

Acknowledgement sent to Michael Prokop <mika@grml.org>:
Extra info received and forwarded to list. Copy sent to Debian Live <debian-live@lists.debian.org>. (Mon, 23 Mar 2009 22:30:02 GMT) Full text and rfc822 format available.

Message #15 received at 520923@bugs.debian.org (full text, mbox):

From: Michael Prokop <mika@grml.org>
To: Chris Lamb <lamby@debian.org>, 520923@bugs.debian.org
Cc: michal.suchanek@ruk.cuni.cz
Subject: Re: Bug#520923: live-initramfs: booting with a bad disk in the system takes forever
Date: Mon, 23 Mar 2009 23:28:16 +0100
[Message part 1 (text/plain, inline)]
* Chris Lamb <lamby@debian.org> [20090323 17:45]:
> michal.suchanek@ruk.cuni.cz wrote:

> > The initramfs scripts cause Linux to read quite a few sectors from disk
> > repeatedly. As a result the boot sequence is _very_ long if a disk is
> > bad and returns I/O errors with delay.

> Um, what do you suggest live-initramfs should do?

An approach could be to use bootfrom=.../live-media=... in such a
situation. Michal, does this help? Would be great if you could try
this in your 'I/O error' setup.

regards,
-mika-
[signature.asc (application/pgp-signature, inline)]

Information forwarded to debian-bugs-dist@lists.debian.org, Debian Live <debian-live@lists.debian.org>:
Bug#520923; Package live-initramfs. (Tue, 24 Mar 2009 21:30:02 GMT) Full text and rfc822 format available.

Acknowledgement sent to Michael Prokop <mika@grml.org>:
Extra info received and forwarded to list. Copy sent to Debian Live <debian-live@lists.debian.org>. (Tue, 24 Mar 2009 21:30:02 GMT) Full text and rfc822 format available.

Message #20 received at 520923@bugs.debian.org (full text, mbox):

From: Michael Prokop <mika@grml.org>
To: Michal Suchanek <michal.suchanek@ruk.cuni.cz>
Cc: 520923@bugs.debian.org
Subject: Re: Bug#520923: live-initramfs: booting with a bad disk in the system takes forever
Date: Tue, 24 Mar 2009 22:28:34 +0100
[Message part 1 (text/plain, inline)]
[Cc-ing 520923@bugs.debian.org]

* Michal Suchanek <michal.suchanek@ruk.cuni.cz> [20090324 18:53]:
> On 03/23/2009 11:28 PM, Michael Prokop wrote:

> >An approach could be to use bootfrom=.../live-media=... in such a
> >situation. Michal, does this help? Would be great if you could try
> >this in your 'I/O error' setup.

> I am booting from the network so the drives should not be read to find the 
> squashfs image, right?

Looking at the source code live-initramfs *shouldn't* be responsible
in this case. Might be some other script like lvm, mdadm,
cryptsetup,.. Do you have LVM, SW-RAID or something like that on
your box?

IMO it would be worth a try if you could add 'set -x' to
/usr/share/initramfs-tools/scripts/live and retry with a rebuilt
initramfs and check out what's going on.

regards,
-mika-
[signature.asc (application/pgp-signature, inline)]

Information forwarded to debian-bugs-dist@lists.debian.org, Debian Live <debian-live@lists.debian.org>:
Bug#520923; Package live-initramfs. (Tue, 31 Mar 2009 09:54:10 GMT) Full text and rfc822 format available.

Acknowledgement sent to Michal Suchanek <michal.suchanek@ruk.cuni.cz>:
Extra info received and forwarded to list. Copy sent to Debian Live <debian-live@lists.debian.org>. (Tue, 31 Mar 2009 09:54:10 GMT) Full text and rfc822 format available.

Message #25 received at 520923@bugs.debian.org (full text, mbox):

From: Michal Suchanek <michal.suchanek@ruk.cuni.cz>
To: Michael Prokop <mika@grml.org>
Cc: "520923@bugs.debian.org" <520923@bugs.debian.org>
Subject: Re: Bug#520923: live-initramfs: booting with a bad disk in the system takes forever
Date: Tue, 31 Mar 2009 11:33:28 +0200
On 03/24/2009 10:28 PM, Michael Prokop wrote:
> [Cc-ing 520923@bugs.debian.org]
>
> * Michal Suchanek<michal.suchanek@ruk.cuni.cz>  [20090324 18:53]:
>> On 03/23/2009 11:28 PM, Michael Prokop wrote:
>
>>> An approach could be to use bootfrom=.../live-media=... in such a
>>> situation. Michal, does this help? Would be great if you could try
>>> this in your 'I/O error' setup.
>
>> I am booting from the network so the drives should not be read to find the
>> squashfs image, right?
>
> Looking at the source code live-initramfs *shouldn't* be responsible
> in this case. Might be some other script like lvm, mdadm,
> cryptsetup,.. Do you have LVM, SW-RAID or something like that on
> your box?

Unfortunately I cannot reproduce the problem anymore because I have overwritten a part of the disk 
and it is not as bad as it used to be.

My guess is that this could be caused by some dm stuff which is standard part of Debian or  the 
live-initramfs searching for something like squashfs, persistent data or swap.

The bad thing is that Linux diligently retries each block three times (at least it does so for 
FUSE), and when a new tool tries to look at the same sector it retries again. The later does happen 
because there is about a dozen sectors tried, most of them twice.

As I do not use either the solution for me is probably to weed out dm and persistence.

Thanks

Michal

-- 
Michal Suchánek
Univerzita Karlova v Praze
Ǔstav Výpočetní Techniky
Ovocný trh 3
Praha 1




Reply sent to daniel@debian.org:
You have taken responsibility. (Mon, 15 Jun 2009 09:18:06 GMT) Full text and rfc822 format available.

Notification sent to michal.suchanek@ruk.cuni.cz:
Bug acknowledged by developer. (Mon, 15 Jun 2009 09:18:06 GMT) Full text and rfc822 format available.

Message #30 received at 520923-done@bugs.debian.org (full text, mbox):

From: Daniel Baumann <daniel@debian.org>
To: 520923-done@bugs.debian.org
Subject: Re: live-initramfs: booting with a bad disk in the system takes forever
Date: Mon, 15 Jun 2009 11:14:50 +0200
Hi,

first, in the case that you have a disk where you don't know that it's
broken, there's nothing we can do about it.

second, in the case where you know which disk is broken, you can either
boot with frommedia so that live-initramfs searches for the rootfs on a
particular device (a non broken one :).

third, if you don't use persistency, you can boot with nopersistent in
order to avoid accessing other devices than the device where the rootfs
lies.

Summarizing: there's nothing to do about it in live-initramfs, therefore
closing this bug.

Regards,
Daniel

-- 
Address:        Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist
Email:          daniel.baumann@panthera-systems.net
Internet:       http://people.panthera-systems.net/~daniel-baumann/




Bug archived. Request was from Debbugs Internal Request <owner@bugs.debian.org> to internal_control@bugs.debian.org. (Tue, 14 Jul 2009 07:27:40 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: Mon Apr 21 02:54:39 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.