Bug#243233: installation report (d-i 20040408)
Package: installation-reports
Debian-installer-version: Daily snapshot, from
http://gluck.debian.org/cdimage/testing/daily/i386/20040408/
uname -a: Linux
debtst3 2.4.25-1-386 #1 Tue Feb 24 08:11:13 EST 2004 i686
GNU/Linux
Date: <Date and time of the install>
2004.04.08
(afternoon/evening)
Method: <How did you install? What did you boot off? If
network
install, from where?
Proxied?>
Using the
Sarge netinst ISO image
(http://gluck.debian.org/cdimage/testing/daily/i386/20040408/),
booted off
the CD, ran the installer, completed via apt-get'ting remaining packages
via the
repositories
(ftp.debian.org)
Machine: <Description of machine (eg, IBM Thinkpad R32)>
x86
"wintel" clone: Gigabyte GA-7VT600-P-L motherboard, plus
PowerColor
Radeon 7500 AGP 64MB video card.
The
built-in SATA controller was not turned off.
Processor:
Duron 1.4
GHz, 266MHz FSB.
Memory:
256MB RAM
DDR (PC3200)
Root Device: <IDE? SCSI? Name of device?>
EIDE (1
HD)
WDC
WD400EB-00CPF0
Root Size/partition table: <Feel free to paste the full
partition
table, with notes on which partitions are
mounted where.>
Command
(m for help): p
Disk
/dev/hda: 40.0 GB, 40019582464 bytes
255 heads,
63 sectors/track, 4865 cylinders
Units =
cylinders of 16065 * 512 = 8225280 bytes
Device Boot
Start
End Blocks Id
System
/dev/hda1
1 1304
10474348+ c W95 FAT32 (LBA)
/dev/hda2
1305
1368 514080 82 Linux
swap
/dev/hda3
*
1369 4865
28089652+ 83 Linux
Command (m
for help): q
debtst3:~#
Output of lspci:
debtst3:~#
lspci
pcilib:
Cannot open /sys/bus/pci/devices
0000:00:00.0
Host bridge: VIA Technologies, Inc. VT8377 [KT400/KT600 AGP] Host Bridge
(rev 80)
0000:00:01.0
PCI bridge: VIA Technologies, Inc. VT8237 PCI Bridge
0000:00:0a.0
SCSI storage controller: Adaptec AHA-2940/2940W / AIC-7871 (rev 03)
0000:00:0c.0
Multimedia video controller: Brooktree Corporation Bt878 Video Capture
(rev 11)
0000:00:0c.1
Multimedia controller: Brooktree Corporation Bt878 Audio Capture (rev
11)
0000:00:0f.0
RAID bus controller: VIA Technologies, Inc.: Unknown device 3149 (rev
80)
0000:00:0f.1
IDE interface: VIA Technologies, Inc.
VT82C586A/B/VT82C686/A/B/VT823x/A/C/VT8235 PIPC Bus Master IDE (rev
06)
0000:00:10.0
USB Controller: VIA Technologies, Inc. VT6202 [USB 2.0 controller]
(rev
81)
0000:00:10.1
USB Controller: VIA Technologies, Inc. VT6202 [USB 2.0 controller]
(rev
81)
0000:00:10.2
USB Controller: VIA Technologies, Inc. VT6202 [USB 2.0 controller]
(rev
81)
0000:00:10.3
USB Controller: VIA Technologies, Inc. VT6202 [USB 2.0 controller]
(rev
81)
0000:00:10.4
USB Controller: VIA Technologies, Inc. USB 2.0 (rev 86)
0000:00:11.0
ISA bridge: VIA Technologies, Inc. VT8237 ISA bridge [K8T800 South]
0000:00:11.5
Multimedia audio controller: VIA Technologies, Inc. VT8233/A/8235/8237
AC97 Audio Controller (rev 60)
0000:00:13.0
Ethernet controller: Realtek Semiconductor Co., Ltd.
RTL-8139/8139C/8139C+ (rev 10)
0000:01:00.0
VGA compatible controller: ATI Technologies Inc Radeon RV200 QW [Radeon
7500]
debtst3:~#
Base System Installation Checklist:
Initial boot worked: [ yes ]
Configure network HW: [ yes ]
Config network: [ yes
]
Detect
CD:
[ yes ]
Load installer modules: [ yes, apparently! ]
Detect hard drives: [ yes ]
Partition hard drives: [ did manually, to keep pre-existing FAT32
partition ]
Create file systems: [ did manually ]
Mount partitions: [ did manually
]
Install base system: [ yes ]
Install boot loader: [ yes ]
Reboot:
[ ]
[O] = OK, [E] = Error (please elaborate below), [ ] = didn't try
it
Comments/Problems:
<Description of the install, in prose, and any thoughts,
comments
and ideas you had during the initial
install.>
A few notes, more or less in order:
(0) Bravo for the use of
kernel 2.4.25 and XF86 4.3: at least Debian Sarge is somewhat keeping up
with the Jones. I just hope that KDE 3.2.1+ will be part of it,
that Sarge will not use KDE 3.1.x.
(1) my tv card, an ATI TV
WONDER/VE was not detected, nor was the related kernel module
loaded.
(2) the PCMCIA kernel modules
were still loaded. I had to remove those two modules manually (pcmcia-cs
& kernel-pcmcia-modules-2.4.25-1-386). Considering new the installer
is supposed to automatically detect hardware, it should have detected I
have no PCMCIA hardware on my *nix box, as well as my TV card and not
loaded extra modules I do not need and loaded the missing one.
(2.1) on my first (or is it second?) run
with d-i 20040408, I have observed that the presence of the PCMCIA
modules can interfere with subsequent operations, more specifically with
proper operation of debconf. You'd loose the ability to specify options
(paper size, etc.) on the first debconf screen. If you wanted a
reason to make sure those modules are not loaded, that was one.
(3) Partitioning: because the
installer gives you only two choices (taking over the whole hard disk or
letting you edit the partition table manually), I had to go the manual
route to make sure my FAT32 partition did not get deleted. I wonder
if a third choice would not have been possible, to let me keep my FAT32
partition and take over the rest automagically. Anyway, so I got to the
"partition main screen" (a picture of which is at
http://www.thiesen.org/d-i/partition-3.png), and even though I've been
installing various x86 Unices (from Minix to the latest Linux distros,
including various BSDs... oh, and both Interactive Unix (before Sun
bought them and shut them down) and SCO), I found this screen confusing
when first encountered. Oh, *once you have used it a few times*, you can
zip through it at a normal clip, but how many *normal* people will have
practice runs at installing their Linux distro? It should be
understandable *at first glance*. I will add a few specific
comments about this screen:
(3.1) the 4 options ("Finish
partitioning and write changes to disk", "Automatically
partition a drive", "Undo changes to partitions" &
"Configure the Logical Volume Manager") should be with the two
buttons at the bottom ("Go back" & "Continue")
and not be visually with the partition table. In fact they should be *the
bottom menu* that would somehow be like this: "Go back",
"Finish partitioning..." (the new "Continue"),
"Automatically...", "Undo..." and "Configure
the...". And there should be a visual clue, a frame, a
scrollable (sp?) list delimiting the partition table.
(3.1.1) Pray tell, what do the little symbols
(smiley, lightning) in that screen do mean, exactly?
(3.2) When you select a specific
partition in the main partition screen, you get to a sub-screen where you
manipulate/set up said partition. Again, this is a screen that is
confusing at first encounter:
(3.2.1) It is not easy to *get out*/complete the
screen once one has finished defining the partition (filesystem,
(3.2.2) In the sub-screen detailing the setup of a
specific partition, there are more or less two parts: specs of the
partition and actions. The three options ("Done setting up...",
"Copy data from..." and "Delete..." should be
"buttons" at the bottom of the screen along with the "go
back" option, instead of visually being with the specs.
(3.2.3) Maybe a bit of help text could be added to the partitioning
menus, especially in the partition settings screens. At first I had no
clue what "Usage method" meant. Do we really need to go to yet
another sub-screen to specify what we want to do with that
partition? I somehow feel that all possible options for a specific
partition could have been specified on the same screen, using hotkeys to
select the setting (with prompting/description at the bottom of the
display).
(3.2.4) That screen too should been reorganised: we
are just dealing with a simple menu that could be streamlined with just
the 4 options ("Format...", "Keep and use the...",
"Use the partition as a..." and "Do not...") and a
"back to main partitioning screen". I must admit, though,
I haven't fully thought this very precise detail fully, as I am trying to
submit these comments as fast as I can, because I have other things to
tend to.
/* I guess by now you should have figured out that I feel the
partitioning step is the weak point of the installer. */
(4) There should have been an
explicit option to have the NIC configured via DHCP (automatically) or
via a static address (manually). We should not have to type
"expert" at the "boot:" prompt for d-i not to use
DHCP automatically, want it or not. Worse even, I had to press
<F1> and then <F3> to figure out what options I had. They
should have been on the first help screen, like you have with
Knoppix.
(5) Assuming grub and use a
graphical splashscreen like lilo can, I am surprised you did not make
grub use one to make the boot process prettier. If we want the
"unwashed masses" to use Linux, we might was well make it
pretty. And that includes a graphical grub menu as well as using the
bootsplash kernel patch. I should not have to patch & recompile
my kernel to make my machine not scream "geek!".
(6) After the initial reboot,
during "part two" of the install, tasksel asks me how I want to
configure my machine. The only way to have just an FTP & telnet
server (maybe I should have used ssh...), I have to select
"conventional Unix server" and get Apache (& more?) in the
process. What if I do not want Apache et al and just want FTP &
telnet/ssh servers? Why can't I do a sub-selection?
(7) The "cdrom" and
"audio" groups should have been correctly configured by d-i so
that the "normal user (s)" could hear music. Not everyone would
know about this (I knew this because of previous experiences with
Debian).
(8) the d-i, after detecting
my hardware, should have detected that my motherboard can initiate a full
shutdown of the machine and should have installed apmd. I had to do this
manually for me just to have to type "shutdown -h now" and have
my machine power off by itself.
(9) In tasksel, I did not
select a graphical DE because I wanted to install KDE 3.2.1...
Nonetheless, I feel that both read-edid and mdetect should have been
loaded and set up so that my subsequent set up of XF86 could have been
fully automated.
(10) After ajusting APT to point to
the unstable archives and for proper "pinning" of packages,
installing "kdebase" from unstable should have also installed
"x-window-system-core" automatically.
(11) Configuring XF86 was painless
due to the fact I had read-edid + mdetect installed and rebooted
afterwards, before installing XF86. The installer (or is it
debconf?) made it possible. The only gripe I would have is the lack, that
I remember, of a "test" button to make sure the chosen settings
do look ok. But if you do not have read-edid & mdetect
installed, ouch.
(12) My other comments are with KDE
3.2.1 and will be submitted seperatly.
I just hope I was not too late to help for the next release of d-i (beta
4).
Cheers.
Bruno Majewski
bruno@pubnix.qc.ca
Reply to: