When updating uuid, update the bitmap as well - external bitmaps

Signed-off-by: Neil Brown <neilb@suse.de>

Fix a couple of typos in the man page.

Signed-off-by: Neil Brown <neilb@suse.de>

When updating uuid, update the bitmap as well - internal bitmaps.

Otherwise when you "--update==uuid" an array with an internal
bitmap, it will break, badly.

Signed-off-by: Neil Brown <neilb@suse.de>

Detail checks if array has been started or not and includes that in report.

Signed-off-by: Neil Brown <neilb@suse.de>

Allow default creation info to to be stored in mdadm.conf

Default owner, group, mode and 'auto' flag can be given in a 'CREATE' line.

Signed-off-by: Neil Brown <neilb@suse.de>

Limit size of bitmap to 2million chunks.

When creating a file bitmap, choose a default size that
results in fewer than 2^21 chunks.  Without this kmalloc
failure in the kernel becomes likely.

Signed-off-by: Neil Brown <neilb@suse.de>

Fix problem with sector/KB size confuision for bitmap sizing.

Signed-off-by: Neil Brown <neilb@suse.de>

Support new offset layout for raid10

Requires 2.6.18.

Signed-off-by: Neil Brown <neilb@suse.de>

Assume "DEVICE partitions" if no DEVICE line present.

This means the output of "mdadm -Es" can be used as a complete
mdadm.conf file .... if you really want to do that.

Signed-off-by: Neil Brown <neilb@suse.de>

Arrange that SparesMissing events generate an email too.

Signed-off-by: Neil Brown <neilb@suse.de>

Support 'mailfrom' line in mdadm.conf so the From: line in alert emails can be explicitly set.

Signed-off-by: Neil Brown <neilb@suse.de>

Use O_DIRECT to read bitmap files.

A pending patch to the kernel causes bitmap file updates
to not go through the page cache, so O_DIRECT is needed to
ensure that we read current data.

Signed-off-by: Neil Brown <neilb@suse.de>

[PATCH] Make sure update_super returns correct value.

For 'force' and 'assemble', update_super must return true
if anything was changed.

Also fix a bug with wonly handling in super0.

Signed-off-by: Neil Brown <neilb@suse.de>
### Diffstat output
 ./super0.c |   18 ++++++++++++++----
 ./super1.c |    7 +++++++
 2 files changed, 21 insertions(+), 4 deletions(-)

diff ./super0.c~current~ ./super0.c

[PATCH] Get the name for partitioned devices in sysfs correct.

There are 64 partitions per device, not 16.

Signed-off-by: Neil Brown <neilb@suse.de>
### Diffstat output
 ./sysfs.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff ./sysfs.c~current~ ./sysfs.c

Release 2.4.1

Annoying alignment bug in version-1 supers on 64bit..

Signed-off-by: Neil Brown <neilb@suse.de>

Fix alignment problem in version-1 superblocks.

NOTE: This is an incompatable change affecting raid5 reshape.
If you want to reshape a raid5 using version-1 superblocks,
use 2.6.17-rc2 or later, and mdadm-2.4.1 or later.

Signed-off-by: Neil Brown <neilb@suse.de>

Add tests/05r1-grow-internal-1

fix write mostly for add and re-add

The following patch makes it possible to tag a device as write-mostly on
--add and --re-add with a non-persistent superblock array. Previously,
this was not working.

Signed-Off-By: Paul Clements <paul.clements@steeleye.com>
Signed-off-by: Neil Brown <neilb@suse.de>

Add ANNOUNCE-2.4

Release 2.4

Signed-off-by: Neil Brown <neilb@suse.de>

Add tests/05r1-grow-internal-1

Updates to md.4

Particularly restiping and sysfs, but a few other bits too.

Signed-off-by: Neil Brown <neilb@suse.de>

Bug fixes to pass test suite

Signed-off-by: Neil Brown <neilb@suse.de>

Support 'bitmap=' in mdadm.conf for auto-assembling arrays with write-intent bitmaps in separate files.

Signed-off-by: Neil Brown <neilb@suse.de>

Man page tidyup.

mdadm.8 improved (I hope).
Rearrange some option documentation and add --backup-file, and other
general improvements.

Signed-off-by: Neil Brown <neilb@suse.de>

Changelog Updates

Signed-off-by: Neil Brown <neilb@suse.de>

Kill old superblocks on create.

Make sure old-version superblocks are not around to confuse anything
when a new array is created.

Signed-off-by: Neil Brown <neilb@suse.de>

Reduce dependance on MD_SB_DISKS

--monitor should now work with arrays larger than 28 devices.

Signed-off-by: Neil Brown <neilb@suse.de>

Add information about reshape to --detail

Also fix problems with dev names and symlinks

Signed-off-by: Neil Brown <neilb@suse.de>

Don't try to get bitmap info on old kernels.

.. this causes silly warnings.
Unfortuantely there are 2.6 kernels which will still produce
warnings.  We cannot reliable detect those.

Signed-off-by: Neil Brown <neilb@suse.de>

Create missing /dev files where needed.

Whenever we need a device file to open, if one cannot be found in /dev,
create a temporary one.

Signed-off-by: Neil Brown <neilb@suse.de>

Fix messy --detail output when there is a failed device.

The missing devices show as garbage.

Signed-off-by: Neil Brown <neilb@suse.de>

Make scanning of /dev work even if it is a symlink

We generally don't want to follow symlinks in /dev, but if
/dev itself is a symlink, we want to follow it.
So nftw needs a bit of help.

Signed-off-by: Neil Brown <neilb@suse.de>

Allow resize to backup to a file.

To support resizing an array without a spare, mdadm now understands
  --backup-file=
which should point to a file for storing a backup of critical data.
This can be given to --grow which will create the file, or
--assemble which will restore from the file if needed.

Signed-off-by: Neil Brown <neilb@suse.de>

Remove ident arg from getinfo_super;

Add a 'name' field to 'info' to compensate.

Signed-off-by: Neil Brown <neilb@suse.de>

Put a 'canary' block in front of the backup

.. so corruptio can be detected.

Signed-off-by: Neil Brown <neilb@suse.de>

Add ANNOUNCE-2.4-pre1

Release 2.4-pre1

Signed-off-by: Neil Brown <neilb@suse.de>

Fix a few issues with the reshape patches.

Signed-off-by: Neil Brown <neilb@suse.de>

Support restarting of a reshape on --assemble

Signed-off-by: Neil Brown <neilb@suse.de>

Initial reshape support

Needs work for other levels etc.

Signed-off-by: Neil Brown <neilb@suse.de>

Add ANNOUNCE-2.3.1

Release some compile fixes.

Signed-off-by: Neil Brown <neilb@suse.de>

Update ChangeLog and release 2.3

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Neil Brown <neilb@suse.de>

Some little bits and pieces...

I'm obviously bored of writing changelog entries.

Signed-off-by: Neil Brown <neilb@suse.de>

Fix printing of size of reiserfs filesystem.

 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=318230

Signed-off-by: Neil Brown <neilb@suse.de>

Sort mdstat entries so that composites are well-ordered.

This means that "-Ds" lists arrays in an approprate order
for assembly.

Signed-off-by: Neil Brown <neilb@suse.de>

Prefer version-1 superblocks for v.large devices.

Signed-off-by: Neil Brown <neilb@suse.de>

Report device size correctly in --detail for Large devices.

Signed-off-by: Neil Brown <neilb@suse.de>

Stuff like..

- report Intent Bitmap in --detail
- report internal bitmap in --examine
- pass' --force through to --grow --bitmap
- support v.large arrays in --grow --bitmap

Signed-off-by: Neil Brown <neilb@suse.de>

Report bitmap offset when examining v1 superblock.

Signed-off-by: Neil Brown <neilb@suse.de>

Work towards allowing larger internal bitmaps in version1 superblocks.

Signed-off-by: Neil Brown <neilb@suse.de>

Support --help --assemble etc.

i.e. allow '--help' to be first and still give
context sensitive help.

Also don't print truncated device-size of very large arrays.

Signed-off-by: Neil Brown <neilb@suse.de>

Have --examine report reshape details.

Signed-off-by: Neil Brown <neilb@suse.de>

Provide error message if --examine does find a valid superblock

Signed-off-by: Neil Brown <neilb@suse.de>

Enable support for v.large raid1.

clean up 'long long' usage for size of array, so that
with v-1 superblocks a raid1 larger than 2TB is possible.

Signed-off-by: Neil Brown <neilb@suse.de>

Further improvements to online help

diff .//ReadMe.c~current~ .//ReadMe.c

Make some improvements to online help.

Signed-off-by: Neil Brown <neilb@suse.de>

Minor fixes

Signed-off-by: Neil Brown <neilb@suse.de>

Assorted fixes...

Signed-off-by: Neil Brown <neilb@suse.de>

uhm.. assort fixes, particularly for recovery offset...

Signed-off-by: Neil Brown <neilb@suse.de>

Allow /etc/mdadm/mdadm.conf as an alternate to /etc/mdadm.conf

This provide compatability with Debian.

Signed-off-by: Neil Brown <neilb@suse.de>

Release 2.2

Signed-off-by: Neil Brown <neilb@suse.de>

Guides on how to use mdadm with initramfs

Signed-off-by: Neil Brown <neilb@suse.de>

Support updating of uuid during --assemble.

Signed-off-by: Neil Brown <neilb@suse.de>

Allow event counters to be off by one

.. even without --force

Signed-off-by: Neil Brown <neilb@suse.de>

syslog support for monitor mode

From: ross@jose.lug.udel.edu (Ross Vandegrift)

Hi Neil,

While adding the text message mode, I saw a FIXME asking for syslog
support in monitor mode.

This patch adds exactly that.
Signed-off-by: Neil Brown <neilb@suse.de>

Report which device failed in a Fail message.

Signed-off-by: Neil Brown <neilb@suse.de>

Support --assume-clean for --create

Signed-off-by: Neil Brown <neilb@suse.de>

Improve option parsing

stuff

====Do Not Remove====
Status: ok
Signed-off-by: Neil Brown <neilb@suse.de>

Allow scanning of devices listed in /proc/partitions independant of /dev

If a device found in /proc/partitions isn't listed in /dev, then
mknod a temporary name and open that.

Signed-off-by: Neil Brown <neilb@suse.de>

Remvoe blank line from '--examine --brief' output.

Signed-off-by: Neil Brown <neilb@suse.de>

Support bitmaps with raid10

And a couple of other little things

Signed-off-by: Neil Brown <neilb@suse.de>

Make sure mdadm -S returns correct error code

If a bad file names was given, exit status still 0.
If some devices couldn't be stopped for "mdadm -Ss"
exit status still 0.

Thanks: Daniel Hottinger <hotti@google.com>

Signed-off-by: Neil Brown <neilb@suse.de>

Create version-4 bitmaps if kernel supports it.

Version-3 bitmaps are host-endian.  Version-4 are little-endian
and so more portable.

Signed-off-by: Neil Brown <neilb@suse.de>

Add 'setup' option to 'test'

Signed-off-by: Neil Brown <neilb@suse.de>

Just some minor stuff...

Signed-off-by: Neil Brown <neilb@suse.de>

Release 2-0

With little fix for -O2 compile

Signed-off-by: Neil Brown <neilb@suse.de>

Add tests/05r1-grow-internal-1

Refine some tests.

Signed-off-by: Neil Brown <neilb@suse.de>

Fix byteordr covnerion ofr 'events' that was '32' instead of '64'

Signed-off-by: Neil Brown <neilb@suse.de>

Fix remaining problems with hot-add bitmap to version-1 superblock

Also some more tests - r5 and r6 bitmaps

Signed-off-by: Neil Brown <neilb@suse.de>

Fix assembling of array with spares when verion-1 superblock is used.

As spared don't have a position in the raid array with verion-1 superblocks,
we need to handle them a bit differently.

Signed-off-by: Neil Brown <neilb@suse.de>

Fix assembling of raid10 in the face of missing devices.

We now check if enough devices are present properly, so
--force can be used to good effect.

Signed-off-by: Neil Brown <neilb@suse.de>

Release!!

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>

Add tests/05r1-internalbitmap-v1c

Add tests/05r1-internalbitmap-v1b

Add tests/05r1-internalbitmap-v1a

Passes all tests, nearly ready for release.

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>

Getting ready for 2.0 release...

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>

remove ANNOUNCE-2.0-devel-?

remove ANNOUNCE-1*

Add raid10 doco to mdadm.8

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>

Lots of little manpage updates.

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>

Add RAID10 and other stuff to md.4

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>

Support internal bitmaps with format-1 superblocks.

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>

Check-in swap_super.c

and update a test that uses it.

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>

Support cross-compiling in Makefile

Thanks to "Anthony Russello" <russello@iomega.com>

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>

Print descriptive level when warning about pre-existing arrays.

Use map_num to get the name, so 'linear' doens't appear as '-1'.

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>

Fix bug on new --detail code

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>