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

Re: Cannot get Etch installer to accept partitions -- success



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Douglas A. Tutty wrote:

Summary of what I have snipped from my original post which Douglas
answered.  I had created *six* RAID1 arrays, with only one partition in
each array, and listed the details in my post.

> I've never considered having so many raid arrays on two disks.  I have
> two arrays (both raid1, one for /boot, the other for LVM).
> 
> Since this is all raid1, I'd do it with two raid1s (one for /boot, the
> other for LVM), and then use LVs to split up the raid1, then mark the
> LVs as physical device for encryption.

Following Douglas's advice, I now have -- finally, after much
experimentation -- the following RAID arrays and partitions:

	LVM VG VG1,  LV home - 865.1 GB Linux device-mapper
			#1 865.1 GB    K crypto      (VG1-home_crypt)
    	Encrypted volume (VG1-home_crypt) - 865.1 GB Linux device-mapper
			#1 865.1 GB    F xfs         /home
    	LVM VG VG1,  LV root -  10.7 GB Linux device-mapper
			#1  10.7 GB    F xfs         /
    	LVM VG VG1,  LV swap -  2.1 GB Linux device-mapper
			#1   2.1 GB    K crypto      (VG1-swap_crypt)
    	Encrypted volume (VG1-swap_crypt) - 2.1 GB Linux device-mapper
			#1   2.1 GB    F swap        swap
	LVM VG VG1,  LV tmp  - 524.3 MB Linux device-mapper
			#1 524.3 MB    K crypto      (VG1-tmp_crypt)
	Encrypted volume (VG1-tmp_crypt)  - 524.3 MB Linux device-mapper
    		#1 524.3 MB    F xfs        /tmp
	LVM VG VG1 LV var    -   4.1 GB Linux device-mapper
			#1   4.1 GB    F xfs        /var
    	RAID1 device #0 -  98.6 MB Software RAID device
			#1  98.6 MB    F xfs        /boot
    	RAID1 device #1 -   2.0 GB Software RAID device
			#1   1.0 T     K lvm
    	SCS12 (0,0,0) (sda) - 1.0 TB ATA SAMSUNG HD103UJ
			#1 primary    98.7 MB B K raid
			#2 primary     1.0 TB   K raid
	SCS12 (0,1,0) (sdb) - 1.0 TB ATA SAMSUNG HD103UJ
			#1 primary    98.7 MB B K raid
			#2 primary     1.0 TB   K raid

Note that I now have only two RAID1 arrays, one for boot and the other
for everything else for LVs, as he has.

I have encrypted three of the LVs: home, swap and tmp.  Home has a LUKS
passphrase.  The other two have random keys.

Something strange happened when I was setting up the three encrypted
LVs.  I set up home and and swap without any problems.  When I came to
  set up tmp, I discovered that the installer had already configured it
to be a *swap* configured LV, in spite of the fact that I had already
configured another LV as an encrypted swap.

The other strange thing happens on boot.  One of the messages which
screens by on boot says:

	XFS mounting file system dm-1
	failed (in red)

Dm-1 is the /var partition.

Likewise, on shutdown, the following message appears:

Stopping MD array md-1 failed - busy  (last two words in red)

These messages do not seem to have any effect on performance. This Etch
installation has been running for three weeks now. There have been --
and still are -- problems; but so far nothing to hang permanently the box.

There are however occasions where this box -- and the other box with
Etch which my spouse uses -- hang to the point where they have to be
rebooted  -- approximately three time a week for each one.

> Have you submitted an installation report as per the installation manual
>  and subscribed to debian-boot?  The debian-boot people wrote the
>  installer and should know what's up.

No, I have not.  Perhaps I should report not only the foregoing but also
comment on the Installation Guide.  I have spent a significant part of
my career writing technical manuals.  The Guide violates just about
every good practice recommended for such documents.

Ken Heard
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.0 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFJQUiQlNlJzOkJmTcRAkuqAJ9elAYFZs4qBXM/H3si3Lq3vK3cYwCeKd04
N1hEIVjqtaRd+eknrAQvQlA=
=4P7U
-----END PGP SIGNATURE-----


Reply to: