Neil Brown
e5bddffd35
Release 2.6.3
2007-08-20 14:22:15 +10:00
Neil Brown
e1584dd2c7
Update changelog
2007-08-20 14:14:46 +10:00
Neil Brown
2295250a73
Report error when grow cannot be restarted.
...
Make sure that if --assemble find an array in the critical region
of a reshape, and cannot find the critical data to restart the
reshape, it gives an error message.
2007-08-20 14:14:42 +10:00
Neil Brown
eb6dae9850
Fix error message when adding a device that is too small.
...
We were multiplying MD_RESERVED_SECTORS by 1024 instead of 512,
so lots of 'is it too small' checks were wrong.
2007-08-20 14:14:28 +10:00
Neil Brown
005debfc11
Fix problem with add a device to a 1.x array created with older mdadm.
...
When adding new disk to an array, don't reserve so much bitmap
space that the disk cannot store the required data. (Needed when
1.x array was created with older mdadm).
2007-08-20 14:14:25 +10:00
Ian Dall
7d19ad0de3
Allow "--write-behind=" to be done in grow mode.
...
From: Ian Dall <ian@beware.dropbear.id.au>
I have a small patch to mdadm which allows the write-behind amount to be
set a array grow time (instead of currently only at grow or create
time). I have tested this fairly extensively on some arrays built out of
loop back devices, and once on a real live array.
2007-07-09 11:29:04 +10:00
Neil Brown
fffdbe5ed0
Release 2.6.2
2007-05-21 14:25:53 +10:00
Neil Brown
a77be58672
Make return code for "--detail --test" more reliable.
...
Missing devices as well as failed devices cause an error.
2007-05-21 14:25:50 +10:00
Neil Brown
1f48664b8e
Add --auto-detect for in-kernel autodetect.
...
This is equivalent to raidautorun that some distros provide.
2007-05-21 14:25:44 +10:00
Neil Brown
f752781f81
Fix --grow --add for linear arrays.
...
The new superblock needs to have a new disk.number. This is a bit of a hack...
Fix handling of negative bitmap offsets on 64bit hosts.
The bitmap offset is a signed 32bit number, so casting to (long)
isn't sufficient. We must cast to (int32_t).
Fix various problems with --grow --add for linear.
The code to add a drive to a live linear array had never
been tested properly and so was buggy. This tidies it up
and means that the new regression-test passes.
2007-05-21 14:25:37 +10:00
Neil Brown
b80da66161
Support failing and removed of detached and faulty devices.
...
This if you unplug a device and udev removes the entry from /dev,
you can still remove the device.
2007-05-11 16:13:03 +10:00
Neil Brown
e003092c07
Release 2.6.1
2007-02-22 15:04:59 +11:00
Neil Brown
00be0b12d2
Fixed old documentation in --grow --help
2007-02-22 14:59:22 +11:00
Neil Brown
48327135d6
Teach restripe to calculate Q syndrome for raid6.
...
This allows mdadm to correctly restart a raid6 grow that
crashed during the critcal phase.
2007-02-22 14:59:19 +11:00
Neil Brown
ae491d1e2c
Fix a warning about an uninitialised variable.
...
The case that doesn't initialise it is impossible,
so just return with an error..
2007-02-22 14:59:16 +11:00
Neil Brown
3e6944b2cc
Fix a bug that was causing incorrect warning from --monitor.
2007-02-22 14:59:13 +11:00
Neil Brown
f8409e5478
Release 2.6
2006-12-21 17:24:38 +11:00
Neil Brown
8382f19bdc
Add new mode: --incremental
...
--incremental allows arrays to be assembled one device at a time.
This is expected to be used with udev.
2006-12-21 17:10:52 +11:00
Neil Brown
3d3dd91e38
Support --uuid= with --create to choose your own UUID.
2006-12-14 17:33:10 +11:00
Neil Brown
ab5303d695
Don't hold md device open for so long in --monitor mode
...
map_dev can be slow and interferes with trying to stop the array.
2006-12-14 17:32:53 +11:00
Neil Brown
4a39c6f236
Check device is large enough before hot-add.
...
This improves quality of error message.
2006-12-14 17:32:49 +11:00
Neil Brown
c5a6f9a61d
Change handling for "--assemble --force" when two drives disappeared at once.
...
If two drives in a raid5 disappear at the same time, then "-Af"
will add them both in rather than just one and forcing the array
to 'clean'. This is slightly safer in some cases.
2006-12-14 17:31:41 +11:00
Neil Brown
bf4fb153a4
Fix and test --update=uuid
...
A number of odd bugs here, but now we have a regression test as well.
2006-12-14 17:31:29 +11:00
Neil Brown
b90c0e9a12
--wait or -W will wait for resync activity to finish on the given devices.
2006-12-14 17:31:22 +11:00
Neil Brown
e336254463
Change 'Device Size' to 'Used Dev Size'
...
because it only shows how much of each device is actually used, not
how big they are.
2006-12-14 17:31:19 +11:00
Neil Brown
758d3a8e7d
Increase raid456 stripe cache size if needed to --grow the array.
...
The setting used unfortunately requires intimate knowledge of the
kernel, and it not reset when the reshape finishes.
2006-12-14 17:31:16 +11:00
Neil Brown
583315d9c5
Give useful message if raid4/5/6 cannot be started because it is not clean and is also degraded.
2006-12-14 17:31:13 +11:00
Neil Brown
7572344665
Default to --auto=yes
...
so the array devices with 'standard' names
get created automatically, as this is almost always what is wanted.
2006-12-14 17:31:10 +11:00
Neil Brown
bee8ec56f4
Support --update=devicesize for cases where the underlying device can change size.
2006-12-14 17:31:03 +11:00
Neil Brown
67a8c82d60
Make Assemble/Force work on raid6 with 2 missing devices.
...
Previously it onl worked when one missing device.
Also split the "force" update_super method into two and it
is really serving two functions.
2006-12-14 17:30:51 +11:00
Neil Brown
434b77559e
--update=resync did exactly the wrong thing for version1 metadata.
2006-11-13 09:12:09 +11:00
Neil Brown
b7a708af6f
Fixed UUID printing in "--detail --brief" for version1 metadata.
2006-11-09 16:28:27 +11:00
Neil Brown
60e1bc1a33
Fixed problems that could cause infinitel loop with auto assemble.
...
If an auto-assembly attempt failes because the array cannot be
opened or because the array has already been created, then we
get into an infinite loop.
Reported-by: Dan Pascu <dan@ag-projects.com>
Fixes-debian-bug: 396582
2006-11-09 09:38:50 +11:00
Goswin Brederlow
2790ffe38a
Document that the default metadata version can be overridden in config file.
...
From: Goswin Brederlow <brederlo@informatik.uni-tuebingen.de>
This is already mentioned in the config documentation, but not in the
place when the normal default is mentioned.
Fixes-debian-bug: 396914
2006-11-09 09:37:45 +11:00
Neil Brown
d87d0978e2
Strdup the bitmap file name found in the config file
...
Otherwise we quickly lose it...
2006-11-09 09:37:10 +11:00
Neil Brown
209b742f4d
Release 2-5-5
2006-11-03 14:52:28 +11:00
Neil Brown
cbfbcb0b50
Improve error message when wrong --update option is given.
2006-10-23 08:56:27 +10:00
Neil Brown
702b557b1c
Fix bugs related to raid10 and the new offset layout.
...
Need to mask of bits above the bottom 16 when calculating number of
copies.
2006-10-19 16:38:29 +10:00
Luca Berra
1cac4c1b55
Don't trigger and error on -As if all arrays are already started.
...
From: Luca Berra <bluca@comedia.it>
Put another way, if we find any active arrays, then assume things are
going as planned.
2006-10-16 15:27:48 +10:00
Luca Berra
e60c27d04e
Don't fail md_open if array already active.
...
From: Luca Berra <bluca@comedia.it>
There is no need and it causes occasional problems.
2006-10-16 15:27:16 +10:00
Luca Berra
435d4ebb2e
Mdassemble improvements
...
From: Luca Berra <bluca@comedia.it>
- Fix a bug where mdassemble didn't close a filedescriptor and so couldn't assembele
stacked arrays.
- Allow mdassemble, when run a second time, to mark all arrays as writable.
This is useful if they are started read-only as is best at boot-time.
2006-10-16 15:26:53 +10:00
Neil Brown
b39827de21
Reduce max bitmap size when bitmap is in a file.
...
To ensure kernel never needs to allocate more than 1 page contiguously.
2006-10-16 15:26:43 +10:00
Neil Brown
0a816ef9fb
Don't #include blkpg.h
...
Just include the content directly. It is safer.
2006-10-16 14:38:13 +10:00
Neil Brown
8e64e04456
Release 2-5-4
2006-10-13 10:12:47 +10:00
Neil Brown
0430ed4868
Remove partitions from components of an md array
...
They do nothing but cause confusion.
2006-10-13 09:02:35 +10:00
Neil Brown
6ba83b5f5e
Allow a number after --auto=yes
...
So if the device name is /dev/md/d0, then the number of
partitions is as given. This is useful in 'CREATE' in
mdadm.conf
2006-10-13 08:23:16 +10:00
Paul Clements
10ae45c12c
Fixed endian problem with bitmap metdata
...
From: Paul Clements <paul.clements@steeleye.com>
2006-10-09 11:17:22 +10:00
Neil Brown
83205b6419
Fix some endian-ness issues with v1 superblocks.
2006-10-09 11:17:12 +10:00
Neil Brown
38098016ca
Allow symlink creation to be disabled from command line or mdadm.conf
2006-08-11 18:00:05 +10:00
Neil Brown
60248f7433
Tidyup automatic name choice for v-1 arrays
...
/dev/md_d0 now becomes '0', not '_d0'.
2006-08-07 11:30:02 +10:00