From: NeilBrown <neilb@suse.de>
To: linux-raid@vger.kernel.org
Subject: [md PATCH 00/10] md patches for 3.5: RAID10 reshape
Date: Tue, 03 Apr 2012 15:53:01 +1000 [thread overview]
Message-ID: <20120403054656.19495.36380.stgit@notabene.brown> (raw)
The following patches are now in my queue for 3.5.
(I may add some more, but to avoid patch-bombing I sending these
now as I think they are ready).
This is all about reshaping arrays and particularly RAID10.
As you probably know, RAID5 reshaping often needs a backup file.
That is a pain for a number of reasons and is slow.
So RAID10 does it without a backup file.
This is achieved by modifying the data_offset during the reshape,
so there is never any need to write data over data that is still
live.
Eventually I'll make RAID5 reshape work this way too (optionally)
however as 0.90 metadata doesn't allow data_offset to be changed,
this is a v1.x feature only.
I have initial patches to support some of this with mdadm, but I don't
know when they'll make it into a release. I have a very strong
feeling that mdadm needs some major refactoring work before I can feel
happy about it again, and that might have to come first, or at least
it will not come last.
(I'll probably put the patches in my git tree sometime soon for
experimentation).
Review/comments/etc always welcome.
You can find all of these in my 'for-next' branch (and so in
linux-next in a day or so) if you would prefer to look at them that
way.
Thanks,
NeilBrown
---
NeilBrown (10):
md/raid10: add reshape support
md/raid10: split out interpretation of layout to separate function.
md/raid10: Introduce 'prev' geometry to support reshape.
md: use resync_max_sectors for reshape as well as resync.
md: teach sync_page_io about new_data_offset.
md/raid10: collect some geometry fields into a dedicated structure.
md/raid5: allow for change in data_offset while managing a reshape.
md/raid5: Use correct data_offset for all IO.
md: add possibility to change data-offset for devices.
md: allow a reshape operation to be reversed.
drivers/md/md.c | 299 ++++++++++-
drivers/md/md.h | 8
drivers/md/raid1.c | 4
drivers/md/raid10.c | 1236 ++++++++++++++++++++++++++++++++++++++++-----
drivers/md/raid10.h | 35 +
drivers/md/raid5.c | 206 +++++---
drivers/md/raid5.h | 6
include/linux/raid/md_p.h | 15 -
8 files changed, 1565 insertions(+), 244 deletions(-)
--
Signature
next reply other threads:[~2012-04-03 5:53 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-04-03 5:53 NeilBrown [this message]
2012-04-03 5:53 ` [md PATCH 01/10] md: allow a reshape operation to be reversed NeilBrown
2012-10-01 9:11 ` Roman Mamedov
2012-10-01 10:15 ` Roman Mamedov
2012-10-02 2:41 ` NeilBrown
2012-04-03 5:53 ` [md PATCH 04/10] md/raid5: allow for change in data_offset while managing a reshape NeilBrown
2012-04-03 5:53 ` [md PATCH 06/10] md: teach sync_page_io about new_data_offset NeilBrown
2012-04-03 5:53 ` [md PATCH 05/10] md/raid10: collect some geometry fields into a dedicated structure NeilBrown
2012-04-03 5:53 ` [md PATCH 03/10] md/raid5: Use correct data_offset for all IO NeilBrown
2012-04-03 5:53 ` [md PATCH 02/10] md: add possibility to change data-offset for devices NeilBrown
2012-04-03 5:53 ` [md PATCH 09/10] md/raid10: split out interpretation of layout to separate function NeilBrown
2012-04-03 5:53 ` [md PATCH 10/10] md/raid10: add reshape support NeilBrown
2012-04-03 5:53 ` [md PATCH 08/10] md/raid10: Introduce 'prev' geometry to support reshape NeilBrown
2012-04-03 5:53 ` [md PATCH 07/10] md: use resync_max_sectors for reshape as well as resync NeilBrown
2014-03-01 5:08 ` [md PATCH 00/10] md patches for 3.5: RAID10 reshape Phillip Susi
2014-03-01 7:03 ` NeilBrown
2014-03-01 19:33 ` Phillip Susi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20120403054656.19495.36380.stgit@notabene.brown \
--to=neilb@suse.de \
--cc=linux-raid@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).