Debian Bug report logs - #315004
raidtools2: Ignoring /etc/raidtab breaks upgrade to mdadm

version graph

Package: raidtools2; Maintainer for raidtools2 is (unknown);

Reported by: Adam C Powell IV <hazelsct@debian.org>

Date: Mon, 20 Jun 2005 00:18:01 UTC

Severity: grave

Tags: sarge

Found in version 1.00.4

Done: Lucas Nussbaum <lucas@lucas-nussbaum.net>

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, Matthijs Mohlmann <matthijs@cacholong.nl>:
Bug#315004; Package raidtools2. Full text and rfc822 format available.

Acknowledgement sent to Adam C Powell IV <hazelsct@debian.org>:
New Bug report received and forwarded. Copy sent to Matthijs Mohlmann <matthijs@cacholong.nl>. Full text and rfc822 format available.

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

From: Adam C Powell IV <hazelsct@debian.org>
To: Debian Bugs <submit@bugs.debian.org>
Subject: raidtools2: Ignoring /etc/raidtab breaks upgrade to mdadm
Date: Sun, 19 Jun 2005 20:02:50 -0400
Package: raidtools2
Version: 1.00.4
Severity: grave
Justification: breaks reboot for sites with multiple RAID arrays
Tags: sarge

Greetings,

The transition from raidtools2 to mdadm breaks all installations with
more than one RAID array because /etc/raidtab is ignored during
autodetection.  I discovered this surprise on reboot, when the order of
my site's RAID arrays was switched, and disabling autodetection (which
intuition said would use /etc/raidtab) of course resulted in none of
them being mounted.  So I tried to remove mdadm but of course raidtools2
depends on it (since raidtools2 is empty).

(Actually, a grad student noticed the problem when booting our server
after a scheduled power outage while I was away.  He of course had no
idea of what was happening, so our site was down for the better part of
a day as a result!  The only clue I had was that mdadm had been
installed as a dependency of raidtools2...)

That the transitional raidtools2 package entered sarge just days before
the release (and that sarge had zero testing cycles, unlike potato or
woody) means that there was just about zero testing for it, and it is
too late for those admins who like me have since rebooted had the same
problem.  But for those who have not needed to reboot, please upload a
fix (translate raidtab to something mdadm understands?) or a *prominent*
debconf warning (maybe even a warning in the raidtools2 description) to
testing-proposed-updates.

If the right place for this is mdadm then please reassign it.  I would
be happy to produce a candidate patch.

Thanks,

-Adam
-- 
GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Welcome to the best software in the world today cafe!
http://www.take6.com/albums/greatesthits.html



Information forwarded to debian-bugs-dist@lists.debian.org:
Bug#315004; Package raidtools2. Full text and rfc822 format available.

Acknowledgement sent to Matthijs Mohlmann <matthijs@cacholong.nl>:
Extra info received and forwarded to list. Full text and rfc822 format available.

Message #10 received at 315004@bugs.debian.org (full text, mbox):

From: Matthijs Mohlmann <matthijs@cacholong.nl>
To: Adam C Powell IV <hazelsct@debian.org>, 315004@bugs.debian.org
Subject: Re: Bug#315004: raidtools2: Ignoring /etc/raidtab breaks upgrade to mdadm
Date: Tue, 21 Jun 2005 21:03:00 +0200
[Message part 1 (text/plain, inline)]
Hi,

We know that there was no testing cycle for Sarge (badly enough) but we
made a release note for this and you can find it here:
http://www.debian.org/releases/stable/i386/release-notes/ch-upgrading.en.html#s-mdadm

If we make a patch that does the transition from /etc/raidtab to
mdadm.conf, i'm afraid that we get more trouble.

Regards,

Matthijs Mohlmann

Adam C Powell IV wrote:
> Package: raidtools2
> Version: 1.00.4
> Severity: grave
> Justification: breaks reboot for sites with multiple RAID arrays
> Tags: sarge
> 
> Greetings,
> 
> The transition from raidtools2 to mdadm breaks all installations with
> more than one RAID array because /etc/raidtab is ignored during
> autodetection.  I discovered this surprise on reboot, when the order of
> my site's RAID arrays was switched, and disabling autodetection (which
> intuition said would use /etc/raidtab) of course resulted in none of
> them being mounted.  So I tried to remove mdadm but of course raidtools2
> depends on it (since raidtools2 is empty).
> 
> (Actually, a grad student noticed the problem when booting our server
> after a scheduled power outage while I was away.  He of course had no
> idea of what was happening, so our site was down for the better part of
> a day as a result!  The only clue I had was that mdadm had been
> installed as a dependency of raidtools2...)
> 
> That the transitional raidtools2 package entered sarge just days before
> the release (and that sarge had zero testing cycles, unlike potato or
> woody) means that there was just about zero testing for it, and it is
> too late for those admins who like me have since rebooted had the same
> problem.  But for those who have not needed to reboot, please upload a
> fix (translate raidtab to something mdadm understands?) or a *prominent*
> debconf warning (maybe even a warning in the raidtools2 description) to
> testing-proposed-updates.
> 
> If the right place for this is mdadm then please reassign it.  I would
> be happy to produce a candidate patch.
> 
> Thanks,
> 
> -Adam
[signature.asc (application/pgp-signature, attachment)]

Information forwarded to debian-bugs-dist@lists.debian.org, Matthijs Mohlmann <matthijs@cacholong.nl>:
Bug#315004; Package raidtools2. Full text and rfc822 format available.

Acknowledgement sent to Martin Michlmayr <tbm@cyrius.com>:
Extra info received and forwarded to list. Copy sent to Matthijs Mohlmann <matthijs@cacholong.nl>. Full text and rfc822 format available.

Message #15 received at 315004@bugs.debian.org (full text, mbox):

From: Martin Michlmayr <tbm@cyrius.com>
To: Adam C Powell IV <hazelsct@debian.org>
Cc: 315004@bugs.debian.org
Subject: Re: raidtools2: Ignoring /etc/raidtab breaks upgrade to mdadm
Date: Tue, 21 Jun 2005 22:12:50 +0100
* Adam C Powell IV <hazelsct@debian.org> [2005-06-19 20:02]:
> The transition from raidtools2 to mdadm breaks all installations with
> more than one RAID array

> That the transitional raidtools2 package entered sarge just days before
> the release (and that sarge had zero testing cycles, unlike potato or
> woody) means that there was just about zero testing for it, and it is
> too late for those admins who like me have since rebooted had the same
> problem.  But for those who have not needed to reboot, please upload a
> fix (translate raidtab to something mdadm understands?) or a *prominent*
> debconf warning (maybe even a warning in the raidtools2 description) to
> testing-proposed-updates.

The fact that raidtools2 entered sarge relatively late doesn't really
have much to do with it since it was pretty clear what would break and
what wouldn't.  However, our assumption was that most people would use
RAID devices that are set to autoconfiguration (and therefore don't
need a configuration file for mdadm to work) and that the rest would
be experienced enough to read the release notes or NEWS.Debian after
an upgrade.  While there's no debconf message, there is a clear
NEWS.Debian message which explains exactly what you need to do.

I'm not sure about what to do.  Obviously, the situation isn't ideal
but I thought a dummy package depending on mdadm and including a
NEWS.Debian was better than nothing at all - since it will work for
most people.  In any case, I'm not sure how many people will not have
upgraded before r1 given that kernel security fixes are schedulded to
come out before.

If you have any idea about a good solution that is acceptable for r1,
I'd certainly like to hear it.  But personally I'm not sure what to
do.
-- 
Martin Michlmayr
http://www.cyrius.com/



Information forwarded to debian-bugs-dist@lists.debian.org, Matthijs Mohlmann <matthijs@cacholong.nl>:
Bug#315004; Package raidtools2. Full text and rfc822 format available.

Acknowledgement sent to Adam C Powell IV <hazelsct@debian.org>:
Extra info received and forwarded to list. Copy sent to Matthijs Mohlmann <matthijs@cacholong.nl>. Full text and rfc822 format available.

Message #20 received at 315004@bugs.debian.org (full text, mbox):

From: Adam C Powell IV <hazelsct@debian.org>
To: Martin Michlmayr <tbm@cyrius.com>
Cc: 315004@bugs.debian.org
Subject: Re: raidtools2: Ignoring /etc/raidtab breaks upgrade to mdadm
Date: Wed, 22 Jun 2005 18:00:05 -0400
Greetings,

First, sorry about the very negative tone of my message.  I had meant to
start it with "First the good news: thanks for the transitional
raidtools2 package!"  But with emotions running high and a rush to get
this out at the end of the day, I neglected this important part of the
report.

And thank you Matthijs for the link into the release notes, I should
have caught that.

On Tue, 2005-06-21 at 22:12 +0100, Martin Michlmayr wrote:
> * Adam C Powell IV <hazelsct@debian.org> [2005-06-19 20:02]:
> > The transition from raidtools2 to mdadm breaks all installations with
> > more than one RAID array
> 
> > That the transitional raidtools2 package entered sarge just days before
> > the release (and that sarge had zero testing cycles, unlike potato or
> > woody) means that there was just about zero testing for it, and it is
> > too late for those admins who like me have since rebooted had the same
> > problem.  But for those who have not needed to reboot, please upload a
> > fix (translate raidtab to something mdadm understands?) or a *prominent*
> > debconf warning (maybe even a warning in the raidtools2 description) to
> > testing-proposed-updates.
> 
> The fact that raidtools2 entered sarge relatively late doesn't really
> have much to do with it since it was pretty clear what would break and
> what wouldn't.  However, our assumption was that most people would use
> RAID devices that are set to autoconfiguration (and therefore don't
> need a configuration file for mdadm to work) and that the rest would
> be experienced enough to read the release notes or NEWS.Debian after
> an upgrade.  While there's no debconf message, there is a clear
> NEWS.Debian message which explains exactly what you need to do.

I see.  Unfortunately, I focused on the "Issues to be aware of for
sarge" section, and since it wasn't there, assumed my upgrade would go
smoothly, particularly since I had been using sarge on this server since
late March.  Perhaps this notice, or a reference to it, should go in
section 5 for r1?

With 104 NEWS.Debian files on my system, it's probably best to assume
that people will not read those during the upgrade.

> I'm not sure about what to do.  Obviously, the situation isn't ideal
> but I thought a dummy package depending on mdadm and including a
> NEWS.Debian was better than nothing at all - since it will work for
> most people.  In any case, I'm not sure how many people will not have
> upgraded before r1 given that kernel security fixes are schedulded to
> come out before.
> 
> If you have any idea about a good solution that is acceptable for r1,
> I'd certainly like to hear it.  But personally I'm not sure what to
> do.

I understand.  For the most part it's "spilt milk" at this point.

Would you accept a patch with a prominent debconf warning?  Although
there are hundreds of debconf dialogs during a typical upgrade, at least
one can be sure that this crosses the admin's screen, in contrast to
NEWS.Debian.

I suppose this illustrates why we need to be more careful with testing
during the release process next time...

Thanks,

-Adam
-- 
GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Welcome to the best software in the world today cafe!
http://www.take6.com/albums/greatesthits.html



Information forwarded to debian-bugs-dist@lists.debian.org, Matthijs Mohlmann <matthijs@cacholong.nl>:
Bug#315004; Package raidtools2. Full text and rfc822 format available.

Acknowledgement sent to Martin Michlmayr <tbm@cyrius.com>:
Extra info received and forwarded to list. Copy sent to Matthijs Mohlmann <matthijs@cacholong.nl>. Full text and rfc822 format available.

Message #25 received at 315004@bugs.debian.org (full text, mbox):

From: Martin Michlmayr <tbm@cyrius.com>
To: Adam C Powell IV <hazelsct@debian.org>, bam@debian.org, agustin.martin@hispalinux.es
Cc: 315004@bugs.debian.org
Subject: Re: raidtools2: Ignoring /etc/raidtab breaks upgrade to mdadm
Date: Tue, 9 Aug 2005 20:44:50 +0100
Hi Adam (et al.),

Sorry for not responding earlier but I've been travelling and quite
busy and completely forgot about this issue.  As I said, I don't know
what the ideal situation is.  The nice thing about the transition
package is that it provides some kind of upgrade path: it depends on
the package you need and has a NEWS entry telling you what to do.  It
will work for most (easy) configurations.  It will not work for
complex systems.  I assumed the sys admins of such people would either
read the docs or could deal with breakage.  But who knows.  Maybe a
debconf message should be added; maybe the package should be removed
altogether.

If you think you have a good solution, please talk to the stable
release manager and NMU this package at will.
-- 
Martin Michlmayr
http://www.cyrius.com/



Information forwarded to debian-bugs-dist@lists.debian.org, Matthijs Mohlmann <matthijs@cacholong.nl>:
Bug#315004; Package raidtools2. Full text and rfc822 format available.

Acknowledgement sent to Adam C Powell IV <hazelsct@debian.org>:
Extra info received and forwarded to list. Copy sent to Matthijs Mohlmann <matthijs@cacholong.nl>. Full text and rfc822 format available.

Message #30 received at 315004@bugs.debian.org (full text, mbox):

From: Adam C Powell IV <hazelsct@debian.org>
To: Martin Michlmayr <tbm@cyrius.com>
Cc: bam@debian.org, agustin.martin@hispalinux.es, 315004@bugs.debian.org
Subject: Re: raidtools2: Ignoring /etc/raidtab breaks upgrade to mdadm
Date: Thu, 11 Aug 2005 12:15:22 -0400
Hello,

On Tue, 2005-08-09 at 20:44 +0100, Martin Michlmayr wrote:
> Hi Adam (et al.),
> 
> Sorry for not responding earlier but I've been travelling and quite
> busy and completely forgot about this issue.  As I said, I don't know
> what the ideal situation is.  The nice thing about the transition
> package is that it provides some kind of upgrade path: it depends on
> the package you need and has a NEWS entry telling you what to do.  It
> will work for most (easy) configurations.  It will not work for
> complex systems.  I assumed the sys admins of such people would either
> read the docs or could deal with breakage.  But who knows.  Maybe a
> debconf message should be added; maybe the package should be removed
> altogether.
> 
> If you think you have a good solution, please talk to the stable
> release manager and NMU this package at will.

Thanks, I'll produce a patch by next week and give this a try.

-Adam
-- 
GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Welcome to the best software in the world today cafe!
http://www.take6.com/albums/greatesthits.html



Information forwarded to debian-bugs-dist@lists.debian.org, Matthijs Mohlmann <matthijs@cacholong.nl>:
Bug#315004; Package raidtools2. Full text and rfc822 format available.

Acknowledgement sent to "Hampton Connell" <Hampton.Connell@asenasystems.net>:
Extra info received and forwarded to list. Copy sent to Matthijs Mohlmann <matthijs@cacholong.nl>. Full text and rfc822 format available.

Message #35 received at 315004@bugs.debian.org (full text, mbox):

From: "Hampton Connell" <Hampton.Connell@asenasystems.net>
To: <315004@bugs.debian.org>
Subject: Etymology MFr politicus
Date: Fri, 20 Apr 2007 10:41:34 +0100
These tree controls are never visible in the parent window.

AN ALLE FINANZINVESTOREN!
DIESE AKTIE WIRD DURCHSTARTEN!
FREITAG 20. APRIL STARTET DIE HAUSSE!
REALISIERTER KURSGEWINN VON 400%+ IN 5 TAGEN!

Symbol: G7Q.F
Company: COUNTY LINE ENERGY
5 Tages Kursziel: 0.95
Schlusskurs: 0.21
WKN:  A0J3B0
ISIN: US2224791077
Markt: Frankfurt

LASSEN SIE SICH DIESE CHANCE NICHT ENTGEHEN!
G7Q WIRD WIE EINE RAKETE DURCHSTARTEN!
UNSERE ERWARTUNGEN WIRD G7Q.F UBERTREFFEN!

Would I ever find out what had happened to this man, who had entrusted me, as it were, with the task of telling the story of his life and death.




Bug closed, send any further explanations to Adam C Powell IV <hazelsct@debian.org> Request was from Lucas Nussbaum <lucas@lucas-nussbaum.net> to control@bugs.debian.org. (Wed, 17 Oct 2007 22:30:06 GMT) Full text and rfc822 format available.

Bug archived. Request was from Debbugs Internal Request <owner@bugs.debian.org> to internal_control@bugs.debian.org. (Thu, 15 Nov 2007 07:25:26 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: Fri Apr 18 08:16:30 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.