[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: wait for the boot device



also sprach Marco d'Itri <md@Linux.IT> [2009.07.20.1023 +0200]:
> > 1. use --run to start arrays as soon as possible, or only start an
> > array when it's completely assembled?
> Why wait?

If you run before it's assembled, any additional members added
afterwards will have to be resynchronised.

> > 2. when can we assume that the arrays are ready to be started?
> When all members have appeared.

Okay, so this is means *not* to use --run, but it also means that
the current behaviour, which is to assemble the array if it is
degraded, will no longer work. If you lose a disk and reboot,
currently the array assembles fine. With the incremental approach,
it would need manual interaction *or* --run, which introduced the
aforementioned problem.

> > 4. How do we properly deal with the device node change? What
> > previously was /dev/md2 with normal assembly is now /dev/md_d2?
> How do other distributions do?

Either they take any sort of choice away from the user (Ubuntu), do
not (seem to) have a migration path (Ubuntu, Suse), or just don't
deal with it at all.

> > It's not quite as easy as just deprecating udevsettle on
> > a distro scale.
> I am quite sure that I always discouraged the use of udevsettle by
> other packages, but policy does not allow me to shoot people who
> do it...

I think you did. However, there was no other way to do it for mdadm
until --incremental was implemented.

-- 
 .''`.   martin f. krafft <madduck@d.o>      Related projects:
: :'  :  proud Debian developer               http://debiansystem.info
`. `'`   http://people.debian.org/~madduck    http://vcs-pkg.org
  `-  Debian - when you have better things to do than fixing systems
 
"zwei monologe, die sich gegenseitig
 immer und immer wieder störend unterbrechen,
 nennt man eine diskussion."
                                                    -- charles tschopp

Attachment: digital_signature_gpg.asc
Description: Digital signature (see http://martin-krafft.net/gpg/)


Reply to: