From: Duncan <1i5t5.duncan@cox.net>
To: linux-btrfs@vger.kernel.org
Subject: Re: lost with degraded RAID1
Date: Fri, 31 Jan 2014 05:58:27 +0000 (UTC) [thread overview]
Message-ID: <pan$7d02a$d46a4ede$2832dda8$da2d2ddd@cox.net> (raw)
In-Reply-To: 8FF7CD63-3B07-42CB-88C3-E88D39280892@colorremedies.com
Chris Murphy posted on Thu, 30 Jan 2014 20:00:39 -0700 as excerpted:
> Another question for Johann is what exact balance command was to go back
> to single? Was there -dconvert and -mconvert? Both are required to go
> from raid1/raid to single/DUP or single/single, and actually get rid of
> the 2nd device. And looking at the man page, I'm not sure how we do that
> conversion and specify which multiple device we're dropping:
>
> [filesystem] balance start [options] <path>
>
> With a missing device, presumably this is obvious, but…
Of course the normal no-missing-device usage of a balance converting back
to single would be one of:
1) Multi-device raidN converting back to single with all devices present,
presumably either to get more space (from raid1/5/6/10), or to get back
/some/ device-loss tolerance when converting back from raid0 (presumably
a data-only conversion in that case, with metadata kept as the default
multi-device raid1).
2) Single-device, converting metadata-only to single from dup, either for
space reasons, or because the physical device is SSD (in which case
single-device metadata default is used due to some devices doing hardware/
firmware dedup with the effect of dup metadata on such a device thus not
entirely predictable anyway), possibly an ssd that does deduping in
hardware, so dup mode's not particularly useful and is somewhat
unpredictable in any case.
IOW, just because it's a conversion to single mode doesn't mean we're
dropping a device, and a rebalance to single mode wasn't in fact designed
to drop a device (that's what device delete is for), so it doesn't really
need a way to specify a device to drop. If one is missing, a rebalance
will obviously rebalance to existing devices, that's obvious, but
otherwise, balance isn't the command for that, device delete is.
But your question to Johann remains valid, since we don't know for sure
that he was doing the "correct" full balance -dconvert -mconvert, and so
far I've simply assumed he was.
--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman
next prev parent reply other threads:[~2014-01-31 5:58 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-01-29 19:16 lost with degraded RAID1 Johan Kröckel
2014-01-30 3:28 ` Duncan
2014-01-30 11:53 ` Johan Kröckel
2014-01-30 17:57 ` Chris Murphy
[not found] ` <CABgvyo9FYGSYpj+jL1oqCvtNUqsC8HZ+z=x-Gz7naWoEcCKWpQ@mail.gmail.com>
2014-01-30 19:32 ` Chris Murphy
2014-01-30 22:18 ` Duncan
2014-01-31 3:00 ` Chris Murphy
2014-01-31 5:58 ` Duncan [this message]
2014-01-31 6:10 ` Chris Murphy
2014-01-31 6:13 ` Chris Murphy
2014-01-31 7:37 ` Duncan
2014-01-31 2:19 ` Chris Murphy
2014-01-30 17:33 ` Chris Murphy
2014-01-30 17:58 ` Hugo Mills
2014-01-30 18:25 ` Chris Murphy
2014-02-03 20:55 ` Johan Kröckel
2014-02-03 21:08 ` Chris Murphy
2014-02-03 21:31 ` Johan Kröckel
2014-02-07 11:34 ` Johan Kröckel
2014-02-07 17:43 ` Chris Murphy
2014-02-08 11:09 ` Johan Kröckel
2014-02-09 5:40 ` Duncan
2014-02-10 14:05 ` Johan Kröckel
[not found] ` <CABgvyo822bOAHeA1GH28MPaBAU+Zdi72MD_uwL+dhopt+nwMig@mail.gmail.com>
2014-01-31 6:28 ` Fwd: " Johan Kröckel
2014-01-31 6:40 ` Johan Kröckel
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='pan$7d02a$d46a4ede$2832dda8$da2d2ddd@cox.net' \
--to=1i5t5.duncan@cox.net \
--cc=linux-btrfs@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).