Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2012-01-21 Thread Adam D. Barratt 
tag 641155 + pending
thanks

On Wed, 2012-01-18 at 17:45 +0400, Michael Tokarev wrote:
 On 18.01.2012 17:34, Adam D. Barratt wrote:
 []
  So I think it is ready to be uploaded to pu.  Should I do so or
  should I post another debdiff for review?
  
  If the only change since the previous debdiff is the removal of the 
  duplicate changelog entry, please go ahead.  If there's anything else, an 
  updated debdiff and summary of the changes would be appreciated.
 
 Yes the only change is the removal of the changelog duplicate
 which you found out.  I rebuilt the package anyway, ofcourse.
 
 Now uploaded.

(As you know, but) for the record, this was accepted a few days ago.

Regards,

Adam




-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2012-01-18 Thread Michael Tokarev 
On 14.01.2012 20:40, Adam D. Barratt wrote:
[]
 +  * Schedule start/stop of mdadm-raid before/after filesystems are
 +checkedmounted/unmounted; thanks to Mario 'BitKoenig' Holbe
 +(closes: #611632).
 [...]
 +  * Make mdadm-raid init script depend on hostname; thanks to Mario
 +   'BitKoenig' Holbe (closes: #610421).

 Have these changes to the LSB headers been tested on squeeze systems?

 I tested these briefly, no extensive tests were done.  That's
 probably my mistake, but the problem is that it's difficult to
 come with some real testcases where this may go wrong.
 But lemme recheck it again, I'll send a follow-up email.
 
 Thanks.  I'm a little paranoid when it comes to changes that affect
 things like when RAID scripts run. :-)

I tested the resulting package in kvm virtual machine in several
different configurations, and it appears to work fine.  The boot
ordering is correct too, with and without lvm.  Sorry it took a
bit more time than I wanted it to take, -- testing is always
time-consuming and I had to install/reinstall several times to
test several configurations.

I also removed the duplicate changelog entry.

So I think it is ready to be uploaded to pu.  Should I do so or
should I post another debdiff for review?

Thanks,

/mjt



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2012-01-18 Thread Adam D. Barratt 

On 18.01.2012 12:55, Michael Tokarev wrote:

On 14.01.2012 20:40, Adam D. Barratt wrote:
[]
+  * Schedule start/stop of mdadm-raid before/after filesystems 
are

+checkedmounted/unmounted; thanks to Mario 'BitKoenig' Holbe
+(closes: #611632).
[...]
+  * Make mdadm-raid init script depend on hostname; thanks to 
Mario

+   'BitKoenig' Holbe (closes: #610421).

Have these changes to the LSB headers been tested on squeeze 
systems?

[...]

I tested the resulting package in kvm virtual machine in several
different configurations, and it appears to work fine.  The boot
ordering is correct too, with and without lvm.  Sorry it took a
bit more time than I wanted it to take, -- testing is always
time-consuming and I had to install/reinstall several times to
test several configurations.


Thanks.


I also removed the duplicate changelog entry.

So I think it is ready to be uploaded to pu.  Should I do so or
should I post another debdiff for review?


If the only change since the previous debdiff is the removal of the 
duplicate changelog entry, please go ahead.  If there's anything else, 
an updated debdiff and summary of the changes would be appreciated.


Regards,

Adam



--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2012-01-18 Thread Michael Tokarev 
On 18.01.2012 17:34, Adam D. Barratt wrote:
[]
 So I think it is ready to be uploaded to pu.  Should I do so or
 should I post another debdiff for review?
 
 If the only change since the previous debdiff is the removal of the duplicate 
 changelog entry, please go ahead.  If there's anything else, an updated 
 debdiff and summary of the changes would be appreciated.

Yes the only change is the removal of the changelog duplicate
which you found out.  I rebuilt the package anyway, ofcourse.

Now uploaded.

Thanks,

/mjt



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2012-01-14 Thread Adam D. Barratt 
On Tue, 2012-01-10 at 12:33 +0400, Michael Tokarev wrote:
 So, after quite some time, I'm attempting another upload of mdadm to
 stable-pu.  All the changes were sitting in testing for half a year,
 and were backported into stable/squeeze version.
 
 As before, the proposed changes are only in debian-specific areas of
 the package, fixing several bugs in supporting/maintenance scripts.
 Complete debdiff is attached.

Thanks for this.  A few (hopefully) quick queries.

+  * Fix checkarray script so that it does not die after scheduling the first
+device when there is no scheduling class specified; thanks to Mario
+'BitKoenig' Holbe (closes: #611627).

The fix for this includes this change:

-  ionice -p $resync_pid $arg
+  ionice -p $resync_pid $arg || :

I assume this is because $arg is unset and the ionice invocation thus
fails?  Wouldn't it better to check that the arguments are sane before
attempting to run ionice instead?

+  * Schedule start/stop of mdadm-raid before/after filesystems are
+checkedmounted/unmounted; thanks to Mario 'BitKoenig' Holbe
+(closes: #611632).
[...]
+  * Make mdadm-raid init script depend on hostname; thanks to Mario
+   'BitKoenig' Holbe (closes: #610421).

Have these changes to the LSB headers been tested on squeeze systems?

+  * Schedule start/stop of mdadm-raid before/after filesystems are
+checkedmounted/unmounted; thanks to Mario 'BitKoenig' Holbe
+(closes: #611632).

This change appears twice in the changelog.  Is this purely an editorial
issue or is there another change which isn't mentioned in the changelog?

Regards,

Adam




-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2012-01-14 Thread Michael Tokarev 
On 14.01.2012 19:50, Adam D. Barratt wrote:
 On Tue, 2012-01-10 at 12:33 +0400, Michael Tokarev wrote:
 So, after quite some time, I'm attempting another upload of mdadm to
 stable-pu.  All the changes were sitting in testing for half a year,
 and were backported into stable/squeeze version.

 As before, the proposed changes are only in debian-specific areas of
 the package, fixing several bugs in supporting/maintenance scripts.
 Complete debdiff is attached.
 
 Thanks for this.  A few (hopefully) quick queries.
 
 +  * Fix checkarray script so that it does not die after scheduling the first
 +device when there is no scheduling class specified; thanks to Mario
 +'BitKoenig' Holbe (closes: #611627).
 
 The fix for this includes this change:
 
 -  ionice -p $resync_pid $arg
 +  ionice -p $resync_pid $arg || :
 
 I assume this is because $arg is unset and the ionice invocation thus
 fails?  Wouldn't it better to check that the arguments are sane before
 attempting to run ionice instead?

No, it is not due to empty $arg.  For the full details see the discussion
in the bugreport mentioned:

 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=611627#17

It wasn't my bugfix, but now when I look at it, the whole logic in the
inner loop appears to be questionable, or racy:

  echo $action  $SYNC_ACTION_CTL
  [ $quiet -lt 1 ]  echo $PROGNAME: I: check queued for array $array. 
2
  ...
  resync_pid= wait=5
  while [ $wait -gt 0 ]; do
wait=$((wait - 1))
resync_pid=$(ps -ef | awk -v dev=$array 'BEGIN { pattern = ^\\[ dev 
_resync]$ } $8 ~ pattern { print $2 }')
if [ -n $resync_pid ]; then
  echo $PROGNAME: I: selecting $ionice I/O scheduling class for resync 
of $array. 2
  ionice -p $resync_pid $arg
  break
fi
sleep 1
  done

What it does here is triggers resync action and waits for
the kernel resync thread to appear, since it might appear
later.  It waits for the kernel thread for up to 5 secs.
Once it is found, checkarray tries to ionice this thread.
The change you're talking about is to fix one half of the
race condition: if the thread exits between the ps call but
before ionice, so ionice reports error and - due to set -e
in effect - whole thing terminates.

What the script does not take into account is that the thread
may start much later than 5 seconds, due to other arrays
being checked/resynced at the same time for example -- and
in this case no scheduling policy will be set for the thread
in question.

But this is still a minor race, in my opinion anyway.  I dont
want to change too much in there for squeeze at this stage.

The change you're talking about is harmless.

 +  * Schedule start/stop of mdadm-raid before/after filesystems are
 +checkedmounted/unmounted; thanks to Mario 'BitKoenig' Holbe
 +(closes: #611632).
 [...]
 +  * Make mdadm-raid init script depend on hostname; thanks to Mario
 +   'BitKoenig' Holbe (closes: #610421).
 
 Have these changes to the LSB headers been tested on squeeze systems?

I tested these briefly, no extensive tests were done.  That's
probably my mistake, but the problem is that it's difficult to
come with some real testcases where this may go wrong.
But lemme recheck it again, I'll send a follow-up email.

 +  * Schedule start/stop of mdadm-raid before/after filesystems are
 +checkedmounted/unmounted; thanks to Mario 'BitKoenig' Holbe
 +(closes: #611632).
 
 This change appears twice in the changelog.  Is this purely an editorial
 issue or is there another change which isn't mentioned in the changelog?

I haven't noticed.  This is purely due to merge error of two series of
changes for debian/changelog.  I'll fix this, thank you for spotting it.

Thank you!

/mjt



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2012-01-14 Thread Adam D. Barratt 
On Sat, 2012-01-14 at 20:23 +0400, Michael Tokarev wrote:
 On 14.01.2012 19:50, Adam D. Barratt wrote:
  +  * Fix checkarray script so that it does not die after scheduling the 
  first
  +device when there is no scheduling class specified; thanks to Mario
  +'BitKoenig' Holbe (closes: #611627).
  
  The fix for this includes this change:
  
  -  ionice -p $resync_pid $arg
  +  ionice -p $resync_pid $arg || :
  
  I assume this is because $arg is unset and the ionice invocation thus
  fails?  Wouldn't it better to check that the arguments are sane before
  attempting to run ionice instead?
 
 No, it is not due to empty $arg.  For the full details see the discussion
 in the bugreport mentioned:
 
  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=611627#17
[...]
 The change you're talking about is to fix one half of the
 race condition: if the thread exits between the ps call but
 before ionice, so ionice reports error and - due to set -e
 in effect - whole thing terminates.

Ah, thanks for the explanation.

  +  * Schedule start/stop of mdadm-raid before/after filesystems are
  +checkedmounted/unmounted; thanks to Mario 'BitKoenig' Holbe
  +(closes: #611632).
  [...]
  +  * Make mdadm-raid init script depend on hostname; thanks to Mario
  +   'BitKoenig' Holbe (closes: #610421).
  
  Have these changes to the LSB headers been tested on squeeze systems?
 
 I tested these briefly, no extensive tests were done.  That's
 probably my mistake, but the problem is that it's difficult to
 come with some real testcases where this may go wrong.
 But lemme recheck it again, I'll send a follow-up email.

Thanks.  I'm a little paranoid when it comes to changes that affect
things like when RAID scripts run. :-)

Regards,

Adam




-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2012-01-10 Thread Michael Tokarev 
In a previous life, on 11.09.2011 02:09, Philipp Kern wrote:
 On Sun, Sep 11, 2011 at 01:09:35AM +0400, Michael Tokarev wrote:
 We'd like to upload a bugfix release of mdadm package for the
 next squeeze point release.  There are mostly cosmetic changes,
 but some of the bugs are very annoying and already made quite
 some users unhappy - like the famous #598957, when mdadm monthly
 cron job generated useless email messages about i/o scheduling
 class for monthly array checks -- this bug alone prompted several
 NMU attempts already.
 
 See 4dbfa62f.1020...@debian.org ff.  We require changes to be in unstable
 first.  We really do.
 
 So fix your package in unstable, let it be tested, and we'll look at it.  So
 far all attempts have targetted stable only, which is the wrong approach.

So, after quite some time, I'm attempting another upload of mdadm to
stable-pu.  All the changes were sitting in testing for half a year,
and were backported into stable/squeeze version.

As before, the proposed changes are only in debian-specific areas of
the package, fixing several bugs in supporting/maintenance scripts.
Complete debdiff is attached.

Can we proceed with the update this time? :)

Thank you!

/mjt
diff -Nru mdadm-3.1.4/debian/changelog mdadm-3.1.4/debian/changelog
--- mdadm-3.1.4/debian/changelog2010-09-03 13:06:30.0 +0400
+++ mdadm-3.1.4/debian/changelog2012-01-10 12:25:03.0 +0400
@@ -1,3 +1,28 @@
+mdadm (3.1.4-1+8efb9d1+squeeze1) stable; urgency=low
+
+  [ martin f. krafft ]
+  * Do not print io rescheduling info message when run by cron
+(closes: #598957).
+  * Fix checkarray script so that it does not die after scheduling the first
+device when there is no scheduling class specified; thanks to Mario
+'BitKoenig' Holbe (closes: #611627).
+  * Schedule start/stop of mdadm-raid before/after filesystems are
+checkedmounted/unmounted; thanks to Mario 'BitKoenig' Holbe
+(closes: #611632).
+  * Work around a shell coding bug for cases when there are zero active
+devices (closes: #618561).
+  * Make mdadm-raid init script depend on hostname; thanks to Mario
+   'BitKoenig' Holbe (closes: #610421).
+  * Schedule start/stop of mdadm-raid before/after filesystems are
+checkedmounted/unmounted; thanks to Mario 'BitKoenig' Holbe
+(closes: #611632).
+
+  [ Michael Tokarev ]
+  * don't print W: auto-read-only in checkarray in quiet mode,
+thanks to Bernd Hanisch for the patch (Closes: #605722, #632797)
+
+ -- Michael Tokarev m...@tls.msk.ru  Sat, 10 Sep 2011 00:32:52 +0400
+
 mdadm (3.1.4-1+8efb9d1) unstable; urgency=low
 
   * Added patch with Makefile fix from upstream (commit 8efb9d1) to fix
diff -Nru mdadm-3.1.4/debian/checkarray mdadm-3.1.4/debian/checkarray
--- mdadm-3.1.4/debian/checkarray   2010-09-03 12:44:09.0 +0400
+++ mdadm-3.1.4/debian/checkarray   2011-09-10 14:30:22.0 +0400
@@ -159,7 +159,7 @@
   fi
 
   if [ $(cat /sys/block/$array/md/array_state) = 'read-auto' ]; then
-echo $PROGNAME: W: array $array in auto-read-only state, skipping... 2
+[ $quiet -lt 1 ]  echo $PROGNAME: W: array $array in auto-read-only 
state, skipping... 2
 continue
   fi
 
@@ -185,7 +185,7 @@
 low) arg='-c2 -n7';;
 high) arg='-c2 -n0';;
 realtime) arg='-c1 -n4';;
-*) break;;
+*) continue;;
   esac
 
   resync_pid= wait=5
@@ -193,8 +193,8 @@
 wait=$((wait - 1))
 resync_pid=$(ps -ef | awk -v dev=$array 'BEGIN { pattern = ^\\[ dev 
_resync]$ } $8 ~ pattern { print $2 }')
 if [ -n $resync_pid ]; then
-  echo $PROGNAME: I: selecting $ionice I/O scheduling class for 
resync of $array. 2
-  ionice -p $resync_pid $arg
+  [ $quiet -lt 1 ]  echo $PROGNAME: I: selecting $ionice I/O 
scheduling class for resync of $array. 2
+  ionice -p $resync_pid $arg || :
   break
 fi
 sleep 1
diff -Nru mdadm-3.1.4/debian/mdadm-raid mdadm-3.1.4/debian/mdadm-raid
--- mdadm-3.1.4/debian/mdadm-raid   2010-09-03 12:44:09.0 +0400
+++ mdadm-3.1.4/debian/mdadm-raid   2012-01-10 12:21:44.0 +0400
@@ -8,10 +8,12 @@
 #
 ### BEGIN INIT INFO
 # Provides:  mdadm-raid
-# Required-Start:mountkernfs
+# Required-Start:mountkernfs hostname
 # Should-Start:  udev devfsd
+# X-Start-Before:checkfs mountall
 # Required-Stop: mountkernfs
 # Should-Stop:   udev devfsd
+# X-Stop-After:  umountfs
 # Default-Start: S
 # Default-Stop:  0 6
 # Short-Description: MD array assembly
@@ -172,6 +174,7 @@
 
 log_action_begin_msg Generating udev events for MD arrays
 for uevent in /sys/block/md*/uevent; do
+  test -e $uevent || break
   sentinel=${uevent#/sys/block/}; sentinel=${sentinel%/uevent}-uevent
   test -e ${STATEDIR:-/doesnotexist}/$sentinel  continue
   test -w $uevent || continue


signature.asc
Description: OpenPGP digital signature

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2011-09-11 Thread Michael Tokarev 
On 11.09.2011 02:09, Philipp Kern wrote:
 On Sun, Sep 11, 2011 at 01:09:35AM +0400, Michael Tokarev wrote:
 We'd like to upload a bugfix release of mdadm package for the
 next squeeze point release.  There are mostly cosmetic changes,
 but some of the bugs are very annoying and already made quite
 some users unhappy - like the famous #598957, when mdadm monthly
 cron job generated useless email messages about i/o scheduling
 class for monthly array checks -- this bug alone prompted several
 NMU attempts already.
 
 See 4dbfa62f.1020...@debian.org ff.  We require changes to be in unstable
 first.  We really do.

Understand.  It was my biggest concern actually, I didn't think it is
a good idea to go stright to stable without being in testing first,
even with the trivial changes this set.

 So fix your package in unstable, let it be tested, and we'll look at it.  So
 far all attempts have targetted stable only, which is the wrong approach.

So, what's the right course of actions here?  Yesterday we were about to
upload a more recent upstream version to unstable, with all the fixes and
alot more.

Now, I'm not sure it is a good idea anymore, since that version will not
go to squeeze obviously.

So I have to upload this squeeze1 version to unstable, with one added
ugly fix for ftbfs with gcc-4.6.  Ugly because I don't want to apply a
more intrusive patch in order to fix -Wunused-but-set-variable (which
are all harmless) when targetting _stable_, instead I will have to remove
-Werror from the upstream makefile and let it build on _testing_, -- just
because I finally want it to appear in stable.

And once accepted into stable with this ugly change, I'll immediately revert
that -Werror change and upload new upstream into unstable.

Is that the right approach?

And what if we already had new upstream uploaded into unstable/testing?

I'm not complaining, I'm just trying to understand the rigth way... ;)

Thank you,

/mjt



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2011-09-11 Thread martin f krafft 
also sprach Michael Tokarev m...@tls.msk.ru [2011.09.11.0846 +0200]:
 So I have to upload this squeeze1 version to unstable,

No, we prepare 3.2.2-1 for unstable, which contains all the fixes we
propose for squeeze1. Once 3.2.2-1 moves to testing, we cherry-pick
the fixes into a new squeeze1 package.

-- 
 .''`.   martin f. krafft madduck@d.o  Related projects:
: :'  :  proud Debian developer   http://debiansystem.info
`. `'`   http://people.debian.org/~madduckhttp://vcs-pkg.org
  `-  Debian - when you have better things to do than fixing systems
 
even a stopped clock gives the right time twice a day.
-- orbital


digital_signature_gpg.asc
Description: Digital signature (see http://martin-krafft.net/gpg/sig-policy/999bbcc4/current)

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2011-09-10 Thread Michael Tokarev 
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Package: release.debian.org
User: release.debian@packages.debian.org
Usertags: pu
Severity: normal

Hello.

We'd like to upload a bugfix release of mdadm package for the
next squeeze point release.  There are mostly cosmetic changes,
but some of the bugs are very annoying and already made quite
some users unhappy - like the famous #598957, when mdadm monthly
cron job generated useless email messages about i/o scheduling
class for monthly array checks -- this bug alone prompted several
NMU attempts already.

There are a few other changes - some also minor/cosmetic (like
another two annoying monthly messages - #605722 and #618561) or
no monthly check at all in some cases (#611627) , and some are
more serious - like more appropriate dependency-based boot
ordering (#611632, #610421).

There are total of 6 bug fixes in this release.  The debdiff is
provided below.

The changes has been run-tested for the previously-buggy cases
and for common non-error cases and appears to work fine.

The upload has been prepared by me, Michael Tokarev, instead
of usual uploader martin f. krafft, -- I'm currently helping
with packaging mdadm.  I Cc'ed pkg-mdadm-devel@ to ensure
all mdadm maintainers will know about the proposed upload.

Please allow the upload to stable-p-u.

Thank you,

/mjt

diff -Nru mdadm-3.1.4/debian/changelog mdadm-3.1.4/debian/changelog
- --- mdadm-3.1.4/debian/changelog  2010-09-03 13:06:30.0 +0400
+++ mdadm-3.1.4/debian/changelog2011-09-11 01:00:23.0 +0400
@@ -1,3 +1,25 @@
+mdadm (3.1.4-1+8efb9d1+squeeze1) stable; urgency=low
+
+  [ martin f. krafft ]
+  * Do not print io rescheduling info message when run by cron
+(closes: #598957).
+  * Fix checkarray script so that it does not die after scheduling the first
+device when there is no scheduling class specified; thanks to Mario
+'BitKoenig' Holbe (closes: #611627).
+  * Schedule start/stop of mdadm-raid before/after filesystems are
+checkedmounted/unmounted; thanks to Mario 'BitKoenig' Holbe
+(closes: #611632).
+  * Make mdadm-raid init script depend on hostname; thanks to Mario
+   'BitKoenig' Holbe (closes: #610421).
+  * Work around a shell coding bug for cases when there are zero active
+devices (closes: #618561).
+
+  [ Michael Tokarev ]
+  * don't print W: auto-read-only in checkarray in quiet mode,
+thanks to Bernd Hanisch for the patch (Closes: #605722)
+
+ -- Michael Tokarev m...@tls.msk.ru  Sat, 10 Sep 2011 00:32:52 +0400
+
 mdadm (3.1.4-1+8efb9d1) unstable; urgency=low

   * Added patch with Makefile fix from upstream (commit 8efb9d1) to fix
diff -Nru mdadm-3.1.4/debian/checkarray mdadm-3.1.4/debian/checkarray
- --- mdadm-3.1.4/debian/checkarray 2010-09-03 12:44:09.0 +0400
+++ mdadm-3.1.4/debian/checkarray   2011-09-10 14:30:22.0 +0400
@@ -159,7 +159,7 @@
   fi

   if [ $(cat /sys/block/$array/md/array_state) = 'read-auto' ]; then
- -echo $PROGNAME: W: array $array in auto-read-only state, skipping... 
2
+[ $quiet -lt 1 ]  echo $PROGNAME: W: array $array in auto-read-only 
state, skipping... 2
 continue
   fi

@@ -185,7 +185,7 @@
 low) arg='-c2 -n7';;
 high) arg='-c2 -n0';;
 realtime) arg='-c1 -n4';;
- -*) break;;
+*) continue;;
   esac

   resync_pid= wait=5
@@ -193,8 +193,8 @@
 wait=$((wait - 1))
 resync_pid=$(ps -ef | awk -v dev=$array 'BEGIN { pattern = ^\\[ dev 
_resync]$ } $8 ~ pattern { print $2 }')
 if [ -n $resync_pid ]; then
- -  echo $PROGNAME: I: selecting $ionice I/O scheduling class for 
resync of $array. 2
- -  ionice -p $resync_pid $arg
+  [ $quiet -lt 1 ]  echo $PROGNAME: I: selecting $ionice I/O 
scheduling class for resync of $array. 2
+  ionice -p $resync_pid $arg || :
   break
 fi
 sleep 1
diff -Nru mdadm-3.1.4/debian/mdadm-raid mdadm-3.1.4/debian/mdadm-raid
- --- mdadm-3.1.4/debian/mdadm-raid 2010-09-03 12:44:09.0 +0400
+++ mdadm-3.1.4/debian/mdadm-raid   2011-09-11 00:46:23.0 +0400
@@ -8,10 +8,12 @@
 #
 ### BEGIN INIT INFO
 # Provides:  mdadm-raid
- -# Required-Start:mountkernfs
+# Required-Start:mountkernfs hostname
 # Should-Start:  udev devfsd
+# X-Start-Before:checkfs mountall
 # Required-Stop: mountkernfs
 # Should-Stop:   udev devfsd
+# X-Stop-After:  umountfs
 # Default-Start: S
 # Default-Stop:  0 6
 # Short-Description: MD array assembly
@@ -172,6 +174,7 @@

 log_action_begin_msg Generating udev events for MD arrays
 for uevent in /sys/block/md*/uevent; do
+  test -e $uevent || break
   sentinel=${uevent#/sys/block/}; sentinel=${sentinel%/uevent}-uevent
   test -e ${STATEDIR:-/doesnotexist}/$sentinel  continue
   test -w $uevent || continue
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.10 (GNU/Linux)

Bug#641155: pu: package mdadm/3.1.4-1+8efb9d1+squeeze1

2011-09-10 Thread Philipp Kern 
On Sun, Sep 11, 2011 at 01:09:35AM +0400, Michael Tokarev wrote:
 We'd like to upload a bugfix release of mdadm package for the
 next squeeze point release.  There are mostly cosmetic changes,
 but some of the bugs are very annoying and already made quite
 some users unhappy - like the famous #598957, when mdadm monthly
 cron job generated useless email messages about i/o scheduling
 class for monthly array checks -- this bug alone prompted several
 NMU attempts already.

See 4dbfa62f.1020...@debian.org ff.  We require changes to be in unstable
first.  We really do.

So fix your package in unstable, let it be tested, and we'll look at it.  So
far all attempts have targetted stable only, which is the wrong approach.

Kind regards
Philipp Kern


signature.asc
Description: Digital signature