On Sun, 10 Feb 2008 01:14:51 GMT, Jim Beard wrote:
> I finally got around to installing the 2008.1 Beta I
> downloaded a week ago. Installation was interesting.
> Details below, if interested.
>
> Questions:
>
> I got nothing in updates via the installed update
> routines, but when I added cooker repositories from
> the usc mirror, almost everything on the system has
> an update.

Yep, 2008.1 is cooker

> Should I install all these updates?

If you want fixes, and plan on reporting bugs, yes.


> Or should I keep the Beta and just update things I am interested in?

As long as you are not going to be reporting problems it is your call.
I think it is abusive to have the bug hunter looking into a problem
might not exist if your system had all updates.

> One main reason for trying the Beta is because it is
> supposed to have support for some multimedia keyboards.

Yes, and my HP multimedia keyboard is not working as I thought the
"feature list" indicated.


> What is the threshhold on reporting problems?

A problem should get fixed if reported and is reproducible.

> I have run into a bunch of problems that seem to appear and
> disappear as I change things.

You better have all updates in before reporting from what I gather
from your problems so far.


> So far, I have not set up an easy way to mail from the Beta, so I
> am not reporting,


Should be an easy fix there. Appended to /etc/postfix/main.cf, I have


# my changes appened to main.cf for my LAN nodes

default_destination_concurrency_limit = 1

masquerade_exceptions = root
mydestination = $myhostname localhost.$mydomain localhost $mydomain
mynetworks = 192.168.1.0/24, 127.0.0.0/8
relayhost = outgoing.verizon.net
relay_domains =
inet_interfaces = all
unknown_local_recipient_reject_code = 550
smtp_host_lookup = dns, native

# I use virtual for local user email (root->bittwister, plus others)
virtual_alias_maps = hash:/etc/postfix/virtual

# These munge headers so mail from users on wm81.home.test to outside
# world get changed to my junk accounts at yahoo and hotmail.com
sender_canonical_maps = hash:/etc/postfix/canonical_sender
smtp_generic_maps = hash:/etc/postfix/generic

# My ISP wants a login id/pw, so I have
smtp_sasl_auth_enable = yes
smtp_sasl_security_options =
smtp_sasl_password_maps = hash:/etc/postfix/smtp_auth

#*********** end main.cf **************************

Via cron,
You can use fetchyahoo to pull yahoo email into an account on your system.
You can use getlive to pull hotmail email into an account on your system.


> I decided to install the Gnome desktop to start with,
> Shifted to KDE and had a problem installing

I installed both and run KDE via startx at runlevel 3 default install
security level. Been pretty stable. Have not tried gnome.

> An attempt to run rpm -qa died; error message not
> written down and forgotten (dgi ?).

> avahi and beagle are now removed. I hope this will
> allow the machine to boot in reasonable time.

Hmmm, I did not install beagle/kerry but I only disabled the avahi deamon.
Too many packages seemed to want avahi.

> The real interesting bits resulted from my fat-fingering
> or failure to remember the final digit on the name of
> the partition I intended to install to.

Hehehe, been there, done that. Glad it was my hot backup partition.

What I have done is use "e2label" to label all my partitions.
Prior to install, I format the target partition then label it.
During partition phase of install, I toggle to expert mode.
Now I can click the partition, label, cut, Ok, Mount, /(paste)
I leave the install partition for last and unchecked format to preserve
the label.

Example Snippet follows:

# blkid /dev/sd*
/dev/sda1: UUID="CE18854F1885378B" LABEL="HP_PAVILION" TYPE="ntfs"
/dev/sda10: LABEL="2008_1" UUID="2cdce39f
/dev/sda11: LABEL="2007_1" UUID="2f77be38
/dev/sda12: LABEL="fc7" UUID="e5cefdc9-59
/dev/sda13: LABEL="kubu7" UUID="94512f8b-
/dev/sda14: LABEL="2008_0" UUID="5dbfc846
/dev/sda2: LABEL="HP_RECOVERY" UUID="4B6E
/dev/sda5: SEC_TYPE="msdos" UUID="466C-5E
/dev/sda6: UUID="9061a2d8-48f9-4a43-b94b-
/dev/sda7: LABEL="accounts" UUID="3342816
/dev/sda8: LABEL="free_a08" UUID="54b176f
/dev/sda9: LABEL="bk_up" UUID="6c14250c-c
/dev/sdb1: LABEL="local" UUID="01441543-d
/dev/sdb2: LABEL="spool" UUID="276c013c-f
/dev/sdb3: LABEL="hotbu" UUID="1bd9c57c-0
/dev/sdb5: LABEL="fc8" UUID="8ee4849e-a57

Doing the labeling with cut/pasting gets me a fstab with

$ grep -v '^#' /2008_1/etc/fstab
LABEL=2008_1 / ext3 relatime 1 1
LABEL=local /local ext3 relatime 1 2
LABEL=2007_1 /2007_1 ext3 user,noauto,relatime 1 2
LABEL=2008_0 /2008_0 ext3 user,noauto,relatime 1 2
LABEL=accounts /accounts ext3 relatime 1 2
LABEL=bk_up /bk_up ext3 user,noauto,relatime 1 2
LABEL=fc7 /fc7 ext3 user,noauto,relatime 1 2
LABEL=fc8 /fc8 ext3 user,noauto,relatime 1 2
LABEL=free_a08 /free_a08 ext3 user,noauto,relatime 1 2
LABEL=hotbu /hotbu ext3 user,noauto,relatime 1 2
LABEL=kubu7 /kubu7 ext3 user,noauto,relatime 1 2
/dev/cdrom /media/cdrom auto umask=0022,users,iocharset=utf8,noauto,ro,exec 0 0
UUID=CE18854F1885378B /win_c ntfs user,noauto,nls=utf8,ro 0 0
UUID=466C-5EB8 /win_d vfat user,noauto,iocharset=utf8 0 0
UUID=4B6E-6BC0 /win_e vfat user,noauto,iocharset=utf8 0 0
none /proc proc defaults 0 0
LABEL=spool /spool ext3 relatime 1 2
UUID=9061a2d8-48f9-4a43-b94b-92f119fb07e5 swap swap defaults 0 0


> The installer decided that my disks should be identified
> in fstab by UUID number rather than the conventional
> /dev/sd.. and my system claimed it was booting from
> the cdrom rather than from the hard disk, even though
> there was nothing in the cdrom. I changed fstab names
> back to the familiar /dev/....

Fought that battle in the alpha 1 testshot.
As you have seen, I now use labels if fstab.

> For the first time ever (since Mandrake 8.1) I have
> had to put myself in the audio group for sound to work.
> I do not know why it always worked before, nor why it
> failed to work this time until I was a proper member
> of audio, but such is life.

I noticed sound does not work on a startx login but will come up later.
/var/log/messages has some reasons and I have not been back to see if
they fixed the problem yet or not. I did check at the time and it had
not been reported.

As usual you need to read all the bug reports to see if your problems
exists, you have updated your system and can recreate the problem.

Main difference I seen with your install and mine,
I did the 32 bit install from iso on HD not from cdrom and I picked
all groups except LSB and installed KDE and GNOME.

Bit Twister wrote:
> Yep, 2008.1 is cooker

> As usual you need to read all the bug reports to see if your problems
> exists, you have updated your system and can recreate the problem.
>
> Main difference I seen with your install and mine,
> I did the 32 bit install from iso on HD not from cdrom and I picked
> all groups except LSB and installed KDE and GNOME.

Looks like I need to update the system, and set up a means
to send mail, and then see how things go. My problem with
reproducible is twofold -- what to do when the problem(s)
seems to come and go as you make frequent changes to other
things? what to do when it is possible things are not
correctly installed & configured? Well, I'll deal with that
when I get to it. If I get to it.

And oddly enough, despite problems with GNOME aborting
my first attempt to install, after an install that
was supposed to provide KDE both are available. Everything
on the DVD installed, whether I asked for it or not.

Of course, life can never go so simply. A hard drive has
started making unh,unh,unh,unh scratchy sounds. I have my
main install and user files on one hard drive, with the
Beta and backup files on another. I may be moving to the
other, despite my preferences to keep things as they are.

And of course it is income tax time, and I still do my own
on my machine. Time to back up that directory again!

Cheers!

jim b.


--
UNIX is not user-unfriendly; it merely
expects users to be computer-friendly.

On Sun, 10 Feb 2008 17:20:54 GMT, Jim Beard wrote:

> Looks like I need to update the system, and set up a means
> to send mail, and then see how things go. My problem with
> reproducible is twofold -- what to do when the problem(s)
> seems to come and go as you make frequent changes to other
> things?

Pretty sure it will only get fixed if you can provide the
steps to reproduce the problem.



> what to do when it is possible things are not
> correctly installed & configured?

Those should be reproducible assuming it is not some unique hardware gotcha.


> And oddly enough, despite problems with GNOME aborting
> my first attempt to install, after an install that
> was supposed to provide KDE both are available. Everything
> on the DVD installed, whether I asked for it or not.

I'm back from a 2008_1 update. 344 packages.
Noticed beagle was installed.


FlighGear still does not work.

>
> And of course it is income tax time, and I still do my own
> on my machine. Time to back up that directory again!

Yes, played with my taxes Friday. I have an old 900mhz box to run
turbo tax. Makes it nicer if you can point to last years tax file.

On Sat, 09 Feb 2008 20:14:51 -0500, Jim Beard wrote:

> One main reason for trying the Beta is because it is
> supposed to have support for some multimedia keyboards.
> I have a Logitech Internet Pro, but choosing the

Just fyi. I'm using the same keyboard, and in 2008.0, it is supported
directly by xorg.

In /etc/X11/xorg.conf, I have
Option "XkbModel" "logiik"

Running xev, and selecting the + volume key shows the key identified as
keysym 0x1008ff13, XF86AudioRaiseVolume, while the - volume shows as
keysym 0x1008ff11, XF86AudioLowerVolume.

To get these keys to work in kde, run kcontrol. Select "Regional &
Accessibility", then "Input actions". Expand the entry for "Mentor Office
Wireless Keyboard" (I know, that's not the right keyboard, but it has the
desired actions defined.). Select "Decrease Volume". Click on the
"Keyboard Shortcut" tab, and press the - volume key. Click on the
"DCOP Call settings". Change the arguments to 2 (that's the pcm mixer
channel). Click on apply, open the kmix mixer window, and try the
volume - key. Then apply the same changes for the Increase volume.
The other keys work in a similar manner, you just have to decide which
application to associate, with each key.

While 2008.1 may be supposed to do the above automatically, it can't
hurt to try the above, and see if it still works.

The keyboard models available are listed in
/usr/share/X11/xkb/rules/base.lst

Regards, Dave Hodgins

--
Change nomail.afraid.org to ody.ca to reply by email.
(nomail.afraid.org has been set up specifically for
use in usenet. Feel free to use it yourself.)

David W. Hodgins wrote:
> On Sat, 09 Feb 2008 20:14:51 -0500, Jim Beard wrote:
>
>> One main reason for trying the Beta is because it is
>> supposed to have support for some multimedia keyboards.
>
> Just fyi. I'm using the same keyboard, and in 2008.0, it is supported
> directly by xorg.
>
> In /etc/X11/xorg.conf, I have
> Option "XkbModel" "logiik"
>
> Running xev, and selecting the + volume key shows the key identified as
> keysym 0x1008ff13, XF86AudioRaiseVolume, while the - volume shows as
> keysym 0x1008ff11, XF86AudioLowerVolume.
>
> To get these keys to work in kde, run kcontrol. Select "Regional &
> Accessibility", then "Input actions". Expand the entry for "Mentor Office
> Wireless Keyboard" (I know, that's not the right keyboard, but it has the
> desired actions defined.). Select "Decrease Volume". Click on the
> "Keyboard Shortcut" tab, and press the - volume key. Click on the
> "DCOP Call settings". Change the arguments to 2 (that's the pcm mixer
> channel). Click on apply, open the kmix mixer window, and try the
> volume - key. Then apply the same changes for the Increase volume.
> The other keys work in a similar manner, you just have to decide which
> application to associate, with each key.
>
> While 2008.1 may be supposed to do the above automatically, it can't
> hurt to try the above, and see if it still works.
>
> The keyboard models available are listed in
> /usr/share/X11/xkb/rules/base.lst

I have not yet tried it in 2008.1, but it definitely works in 2008.0.
I have the two eight-notes launching Amarok, the envelope launching
Thunderbird, home locking the screen, and the increase volume,
decrease volume, and mute keys working. That leave the play/pause
key not mapped to anything, but I will leave that for later.

The odd feature is that if I have the kmix speaker in the toolbar
and bring it up, it blocks the keys from working. Also, changes made
using the keys do not affect the toolbar kmix slider, but do change
the volume level. If you put the slider to top, then lower the
volume with the key while the slider is minimized, when you bring the
slider up again it will be at the top and you cannot increase the
volume.

Cheers!

jim b.

--
UNIX is not user-unfriendly; it merely
expects users to be computer-friendly.

Jim Beard wrote:
>> Questions:
>>
>> I got nothing in updates via the installed update
>> routines, but when I added cooker repositories from
>> the usc mirror, almost everything on the system has
>> an update.
>
>> Should I install all these updates?

Bit Twister wrote:
> If you want fixes, and plan on reporting bugs, yes.

My first bug report is filed, Nr 36760. rpm returns an
error message upon each attempt to use it, and then
disappears. I tried using gftp to get an update
package and install with rpm, but no joy.

If I get real enthusiastic, I may copy over rpm
from 2008.0 and see if it will run.

Using mcc, at one point, I had 1507 updates listed, with a
notice that none could be installed. AT other points, mcc
and urpmi both told me my system was all up to date. I have
my doubts about the accuracy of the long list, but something
is definitely not working as expected.

Incidentally, booting 2008.1 is still taking several minutes.
Shorewall loads, and the machine then sits there and
contemplates its naval or something for a few minutes before
something times out and it goes on to finish booting. Any
idea what it might be? With beagle and avahi removed,
the choices seem to be:

S53xfs
S54netfs
S54nfs-common
S55keytable
S55nfs-server
S56ntpd
S56xinetd
S80postfix
S90crond
S95jexec
S95kheader
S99local

I suppose I could put come echo commands in the
init scripts and see which ones get printed...

Cheers!

jim b.

--
UNIX is not user-unfriendly; it merely
expects users to be computer-friendly.

On Sun, 10 Feb 2008 22:55:57 GMT, Jim Beard wrote:
>
> Using mcc, at one point, I had 1507 updates listed, with a
> notice that none could be installed.

Does sound like a bit much, unless you have both 64 and 32 bit mirrors set.

> AT other points, mcc
> and urpmi both told me my system was all up to date.

Yes, I reported a problem in alpha.
urpmi --wget --update --auto-update --auto
did no updates.
urpmi.update --wget -a
urpmi --wget --auto-select --auto
did the updates.


> I have
> my doubts about the accuracy of the long list, but something
> is definitely not working as expected.

I am going to have to guess it is 64 bit install gotcha.

> Incidentally, booting 2008.1 is still taking several minutes.
> Shorewall loads, and the machine then sits there and
> contemplates its naval or something for a few minutes before
> something times out and it goes on to finish booting. Any
> idea what it might be? With beagle and avahi removed,
> the choices seem to be:

If I had to guess, avahi not installed is causing the grief.
Then again, it can be the 64 bit install. I installed 32 bit.

Of course maybe your bad disk is helping you into the ditch.

I did have one big pause but come to find out it was a full check on my
84 gig partition.

If I were you, I would have to do a 32 bit install to prove it is a 64
bit problem.




Due to a problem found in alpha release, my admin install diary now has
the following:

mcc
click, Configure media sources for install and update
Disable all removable media
Set wget in Options Global
quit mcc

If mcc is broke for disable media.
edt /etc/urpmi/urpmi.cfg
add
downloader: wget
verify-rpm: 1
to first stanza

and add
ignore
above key-ids: for each removable media (/dev/cdrom)


Click Save/Quit

ping -c1 yahoo.com

/local/bin/set_mirror <---- my script to add mirrors based on release


urpmi --wget wget --auto
urpmi --wget rpm --auto
urpmi --wget perl-URPM --auto
urpmi --wget urpmi --wget --auto
urpmi --wget rpmdrake --auto
urpmi --wget rpm-helper --auto
service httpd stop
urpmi --wget --auto-select --auto

Jim Beard writes:

>Jim Beard wrote:
> >> Questions:
> >>
> >> I got nothing in updates via the installed update
> >> routines, but when I added cooker repositories from
> >> the usc mirror, almost everything on the system has
> >> an update.
> >
> >> Should I install all these updates?

>Bit Twister wrote:
>> If you want fixes, and plan on reporting bugs, yes.

>My first bug report is filed, Nr 36760. rpm returns an
>error message upon each attempt to use it, and then
>disappears. I tried using gftp to get an update
>package and install with rpm, but no joy.

>If I get real enthusiastic, I may copy over rpm
>from 2008.0 and see if it will run.

>Using mcc, at one point, I had 1507 updates listed, with a
>notice that none could be installed. AT other points, mcc
>and urpmi both told me my system was all up to date. I have
>my doubts about the accuracy of the long list, but something
>is definitely not working as expected.

>Incidentally, booting 2008.1 is still taking several minutes.
>Shorewall loads, and the machine then sits there and
>contemplates its naval or something for a few minutes before
>something times out and it goes on to finish booting. Any
>idea what it might be? With beagle and avahi removed,
>the choices seem to be:

Uh, you can see what is next. It is probably shorewall doing some external
address search which is timing out.
Note that postfix is often the culprit, since it does do address
resolution, but it is too late in the sequence, unless you were really not
watching the screen to see what had finished running before the hangup.


>S53xfs
>S54netfs
>S54nfs-common
>S55keytable
>S55nfs-server
>S56ntpd
>S56xinetd
>S80postfix
>S90crond
>S95jexec
>S95kheader
>S99local

>I suppose I could put come echo commands in the
>init scripts and see which ones get printed...

>Cheers!

>jim b.

>--
>UNIX is not user-unfriendly; it merely
> expects users to be computer-friendly.

Well, the copy of the Beta I have is not Beta-ready.
I cannot update, and I am abandoning the effort.

Yesterday, much of my home directory stuff had been
reassigned to owner nobody. Still group jim. I changed
that back yesterday, but today many of them had gone back.
Not nice. With the directory owned by nobody, I could
not load KDE due to inability to set the .auth* file.
Changed all that stuff to owner/group jim.

I mounted my partions with the 2008.0 system commands,
and the old rpm works, at least enough to do a rpm -qa.
Hoping that would allow me to updata, I tried urpmi
and then mcc. I still cannot get an update for anything.

Given that BitTwister thought removing Avahi might have
had negative effects, I decided to reinstall. When the
splash came up and asked me if I wanted to install or
to update 2008.0 or 2008.1, I decided that I would avail
myself of this nifty shortcut, and went for the upgrade
2008.1.

The installer ran for a while and then gave me an estimated
time to completion of over 3 hours, installing from the DVD.
I let it run a while, to see if it was just miscalibrated,
but progress on the progress bar over time suggested that
would be about right. I hit the soft reboot button, and
went back to my normal 2008.0.

Only to find that the installer had somehow mucked with
my 2008.0. 2008.1 was on partition sda6, 2008.0 scattered
across a half dozen partitions on sdb. The installer should
not have touched sdb, with the possible exception of
/boot/grub/menu.lst on sdb1. End of story.

Except, to get everything back operating smoothly, I had
to do a full system restore.

I am not inclined to dedicate a second machine to Beta
testing (it would have to be a Pentium 4, or buy another
64-bit machine -- neither is attractive to me) nor do a
complete backup each time I want to try the Beta and
expect to do a complete restore after tinkering with it.
That is more than I feel like these days. Maybe when I
am retired, but that may be some time away.

No cheers.

jim b.


--
UNIX is not user-unfriendly; it merely
expects users to be computer-friendly.

On Tue, 12 Feb 2008 01:01:57 GMT, Jim Beard wrote:
> Well, the copy of the Beta I have is not Beta-ready.
> I cannot update, and I am abandoning the effort.
>
> I am not inclined to dedicate a second machine to Beta
> testing (it would have to be a Pentium 4, or buy another
> 64-bit machine -- neither is attractive to me) nor do a
> complete backup each time I want to try the Beta and
> expect to do a complete restore after tinkering with it.

Dang, all I can say is my 2008.1 32 bit is without all the problems
you have indicated.

After seeing your post, I booted 2008.1 and did another 96 package
update and checked junk to verify uid/gid was still 500/500.

I never do the update, always pick custom clean install.

Doing the install from Hard Drive is faster than from DVD.

I started the 64 bit beta download and will try it to see if I have
your problems.

Bit Twister wrote:
> On Tue, 12 Feb 2008 01:01:57 GMT, Jim Beard wrote:
>> Well, the copy of the Beta I have is not Beta-ready.
>> I cannot update, and I am abandoning the effort.

> I never do the update, always pick custom clean install.

Yes, but since all I was really after was the two packages
I had removed, I thought it might go faster. The initial
install was to a clean partition. Of course, the installer
decided to overwrite everything....

> Doing the install from Hard Drive is faster than from DVD.

Yes, but I had not bothered to make a bootable cd to start
things off with.

> I started the 64 bit beta download and will try it to see if I have
> your problems.

Part of the problem may be I had (note past tense) a week-old
copy. If the updated packages are now in the Beta download,
all may go well.

I'll be waiting to see how your's goes.

Cheers!

jim b.

--
UNIX is not user-unfriendly; it merely
expects users to be computer-friendly.

On Tue, 12 Feb 2008 02:23:15 GMT, Jim Beard wrote:

> Part of the problem may be I had (note past tense) a week-old
> copy.

ISOs are released only once. They are not updated until next iso build.
No idea if we bet beta 2 or it will be rc1.

I assume I downloaded the same iso image as you. Time stamp set same
as found on carrol mirror.

$ ls -l *64*iso
2008-01-25 03:46 mandriva-linux-2008-spring-free-ophrys-dvd-x86_64.iso

My system hardware:
cpu AMD Athlon(tm) 64 Processor 3800+
memory 2 129 ns 512 MB (Single-bank Connection)
video RS480 [Radeon Xpress 200G Series] (on board)
sda disk Seagate ATA ST3200827AS 200 gig [sata_sil]
sdb disk Seagate ATA ST3500641AS 500 gig [sata_sil]

Clean install using Custom during partition phase.
/dev/sdb6 13G 5.2G 6.5G 45% /20081_64


Partition formatted and labeled under 2008.0 then unchecked "format" in
partition phase install to keep label mounting for /.

fstab has label mounting not UID partition ID.
Everything is under /.
Swap is shared across all installs.

Default Security level: High.

Initial package install took about 23 minutes using a Hard Drive install.
All package groups selected except LSB, and only picked KDE, GNOME desktops.
Went through and added a bunch of stuff plus all games and database.
Disabled avahi-daemon during Configuration phase of install. See list
of active daemon/service set on boot end of reply.

On first boot, KDE desktop, checked rpm -qa, it worked.

Did my usual manual wget, rpm, urpmi, perl-URPM, rpmdrake, rpm-helper package
pre-update then did the full auto update with
urpmi --wget --auto-select --auto
517 package updates were installed.
One missing package. Same thing on 32 bit today.
One Gnome package failure because of a package bug.

Checked rpm -qa again, still worked.
No apparent corruption on other partitions.

Have not executed all cron jobs, yet, to check for your nobody and partition
corruption problems.

I am going to assume you md5 checked your iso download before burn and
md5 after iso burn.

If so, first guess is heat problem on your system which is odd because
I would have assumed your disk restore would heat up the cpu more so
than normal operation.

Based on that assumption, that would leave some instruction or one of
the daemon/services dinked up your system. Here is my daemon list:
Forgot to set runlevel 5 so system booted at 5.

$ cat chkconfig.list_orig
acpi 0ff 1ff 2n 3n 4n 5n 6ff
acpid 0ff 1ff 2ff 3n 4n 5n 6ff
alsa 0ff 1ff 2n 3n 4n 5n 6ff
atd 0ff 1ff 2ff 3n 4n 5n 6ff
avahi-daemon 0ff 1ff 2ff 3ff 4ff 5ff 6ff
consolekit 0ff 1ff 2n 3n 4n 5n 6ff
crond 0ff 1ff 2n 3n 4n 5n 6ff
ct_sync 0ff 1ff 2n 3n 4n 5n 6ff
cups 0ff 1ff 2n 3n 4n 5n 6ff
dkms 0ff 1ff 2ff 3n 4n 5n 6ff
dm 0ff 1ff 2ff 3ff 4ff 5n 6ff
haldaemon 0ff 1ff 2ff 3n 4n 5n 6ff
harddrake 0ff 1ff 2ff 3n 4n 5n 6ff
httpd 0ff 1ff 2ff 3n 4n 5n 6ff
iptables 0ff 1ff 2n 3n 4n 5n 6ff
keytable 0ff 1ff 2n 3n 4n 5n 6ff
kheader 0ff 1ff 2n 3n 4ff 5n 6ff
ksysguard 0ff 1ff 2ff 3ff 4ff 5ff 6ff
lisa 0ff 1ff 2ff 3ff 4ff 5ff 6ff
mandi 0ff 1ff 2n 3n 4n 5n 6ff
messagebus 0ff 1ff 2n 3n 4n 5n 6ff
mysqld 0ff 1ff 2ff 3ff 4ff 5ff 6ff
netconsole 0ff 1ff 2ff 3ff 4ff 5ff 6ff
netfs 0ff 1ff 2ff 3n 4n 5n 6ff
network 0ff 1ff 2n 3n 4n 5n 6ff
network-up 0ff 1ff 2n 3n 4n 5n 6ff
nfs-common 0ff 1ff 2ff 3n 4n 5n 6ff
nfs-server 0ff 1ff 2ff 3n 4n 5n 6ff
ntpd 0ff 1ff 2n 3n 4n 5n 6ff
numlock 0ff 1ff 2ff 3n 4n 5n 6ff
partmon 0ff 1ff 2ff 3ff 4ff 5ff 6ff
portmap 0ff 1ff 2ff 3n 4n 5n 6ff
postfix 0ff 1ff 2n 3n 4n 5n 6ff
proftpd 0ff 1ff 2ff 3n 4n 5n 6ff
resolvconf 0ff 1ff 2n 3n 4n 5n 6ff
shorewall 0ff 1ff 2n 3n 4n 5n 6ff
smb 0ff 1ff 2ff 3ff 4ff 5ff 6ff
sound 0ff 1ff 2n 3n 4n 5n 6ff
sshd 0ff 1ff 2n 3n 4n 5n 6ff
syslog 0ff 1ff 2n 3n 4n 5n 6ff
ucarp 0ff 1ff 2n 3n 4n 5n 6ff
webmin 0ff 1ff 2ff 3ff 4ff 5ff 6ff
xinetd 0ff 1ff 2ff 3n 4n 5n 6ff

xinetd based services:
cups-lpd: off
cvs: off
proftpd-xinetd: off
rsync: off
sshd-xinetd: off

Final WAG, you have some Spooky Dooky going on in your hardware.

Do consider closing the bug reports you opened.

On Tue, 12 Feb 2008 06:55:02 +0000 (UTC), Bit Twister wrote:

Ok, finished the 2008.1 beta 64 bit install.

Ran all /etc/cron.* jobs and no disk partition corruption and
/home accounts did not flip over to nobody ownership.

No abnormal pauses during boot at runlevel 3.

After sleeping, I thought maybe you have hardware problems with
dvd/sata drive interference or something.

But, if your restore was from same media hardware used during iso
install I have no good guess as to where to look for the problems
you had.

Bit Twister wrote:
> On Tue, 12 Feb 2008 02:23:15 GMT, Jim Beard wrote:
>
>> Part of the problem may be I had (note past tense) a week-old
>> copy.
>
> ISOs are released only once. They are not updated until next iso build.
> No idea if we bet beta 2 or it will be rc1.

My copy should be the same as yours, then.
>
> I assume I downloaded the same iso image as you. Time stamp set same
> as found on carrol mirror.
Four minutes difference in time stamp.
>
> $ ls -l *64*iso
> 2008-01-25 03:46 mandriva-linux-2008-spring-free-ophrys-dvd-x86_64.iso
4345438208 2008-01-25 03:50
mandriva-linux-2008-spring-free-ophrys-dvd-x86_64.iso

>
> My system hardware:
> cpu AMD Athlon(tm) 64 Processor 3800+
AMD Athlon 64 two dual-core 5000+
> memory 2 129 ns 512 MB (Single-bank Connection)
1 GB, ?533 MHz? not sure what the ns is.
> video RS480 [Radeon Xpress 200G Series] (on board)
nVidia built into the motherboard but using nVidia 7300 nvidia card
> sda disk Seagate ATA ST3200827AS 200 gig [sata_sil]
> sdb disk Seagate ATA ST3500641AS 500 gig [sata_sil]
Two Western Digital SATA 250 gig
>
> Clean install using Custom during partition phase.
> /dev/sdb6 13G 5.2G 6.5G 45% /20081_64
Installed clean to /dev/sda6 (where non-used XP lives, plus
storage; all 2008.0 partitions with executables on sdb.
>
>
> Partition formatted and labeled under 2008.0 then unchecked "format" in
> partition phase install to keep label mounting for /.

Initially, I had the installer format sda6. I did have to tinker
with fstab inconsequence.
>
> fstab has label mounting not UID partition ID.
> Everything is under /.
> Swap is shared across all installs.

I don't know where the installer put swap for 2008.1. It is
sdb5 for 2008.0.

> Default Security level: High.

I dropped this to standard on my system.
>
> Initial package install took about 23 minutes using a Hard Drive install.
> All package groups selected except LSB, and only picked KDE, GNOME desktops.
> Went through and added a bunch of stuff plus all games and database.
> Disabled avahi-daemon during Configuration phase of install. See list
> of active daemon/service set on boot end of reply.

Installed from a DVD. md5sum checked for the iso file on disk
before burning and again for the dvd burned.
>
> On first boot, KDE desktop, checked rpm -qa, it worked.
>
> Did my usual manual wget, rpm, urpmi, perl-URPM, rpmdrake, rpm-helper package
> pre-update then did the full auto update with
> urpmi --wget --auto-select --auto
> 517 package updates were installed.
> One missing package. Same thing on 32 bit today.
> One Gnome package failure because of a package bug.

My first attempt was to install Gnome, and that aborted due
to package failure. Second attempt, with KDE selected, also
had a package failure, but I don't think that was significant.
>
> Checked rpm -qa again, still worked.
> No apparent corruption on other partitions.

Unscrambling fstab for 2008.1 was a chore. Unfamiliarity with
labels/UUIDs did not help.
>
> Have not executed all cron jobs, yet, to check for your nobody and partition
> corruption problems.
>
> I am going to assume you md5 checked your iso download before burn and
> md5 after iso burn. Yep.
>
> If so, first guess is heat problem on your system which is odd because
> I would have assumed your disk restore would heat up the cpu more so
> than normal operation.
>
> Based on that assumption, that would leave some instruction or one of
> the daemon/services dinked up your system. Here is my daemon list:
> Forgot to set runlevel 5 so system booted at 5.
>
> $ cat chkconfig.list_orig
> acpi 0ff 1ff 2n 3n 4n 5n 6ff


I no longer have 2008.1 on the system, but changes were first to
disable avahi and beagle, and later to remove them. Not long
before I removed 2008.1 totally, error messages complained that
Console-kit could not be found.
>
> Final WAG, you have some Spooky Dooky going on in your hardware.

Always possible, but 2008.0 seems to work fine. And, I cannot see
how dinking up fstab would have messed up things without making
it totally obvious. The OS was finding the correct partitions and
programs (after I got fstab straightened out).
>
> Do consider closing the bug reports you opened.

The only thing opened was the rpm problem 36760 I think it was.
That was a problem from the git-go, and copying over and running
rpm for 2008.0 under 2008.1 worked, so I am inclined to leave it.
It is possible that the 2008.1 rpm was corrupted somehow, or the lib
file that contained the symbol missing could have been corrupted,
but this seems unlikely.

I may take another try at it this coming weekend.

Cheers!

jim b.

--
UNIX is not user-unfriendly; it merely
expects users to be computer-friendly.

On Tue, 12 Feb 2008 23:22:56 GMT, Jim Beard wrote:
> Bit Twister wrote:

> AMD Athlon 64 two dual-core 5000+

Forgot you were dual core.

>> memory 2 129 ns 512 MB (Single-bank Connection)
> 1 GB, ?533 MHz? not sure what the ns is.

I'll guess Nano Second for speed.
Values are from mcc Hardware

>> Clean install using Custom during partition phase.
>> /dev/sdb6 13G 5.2G 6.5G 45% /20081_64
> Installed clean to /dev/sda6 (where non-used XP lives, plus
> storage; all 2008.0 partitions with executables on sdb.

Dang, misremembered start of thread, thought you installed on sdb.

>
> Initially, I had the installer format sda6. I did have to tinker
> with fstab inconsequence.

Good chance for setting wrong one there unless you cut/pasted comment
above UUID.

>
> I don't know where the installer put swap for 2008.1. It is
> sdb5 for 2008.0.

I never have had to set swap location except day-one install.
Custom always found/set it for me.

>
>> Default Security level: High.
>
> I dropped this to standard on my system.

Hehehe, you sure are making it hard on re-creating your problem.


>
> My first attempt was to install Gnome, and that aborted due
> to package failure. Second attempt, with KDE selected, also
> had a package failure, but I don't think that was significant.

I had no package until update. On next test shot, try picking both KDE,GNOME

>
> Unscrambling fstab for 2008.1 was a chore. Unfamiliarity with
> labels/UUIDs did not help.

Did you dink with fstab before updates?
Did you reboot after fstab changes?


> I no longer have 2008.1 on the system, but changes were first to
> disable avahi and beagle, and later to remove them. Not long
> before I removed 2008.1 totally, error messages complained that
> Console-kit could not be found.

Recommend install, updates, reboot, then start dinking around.

>>
>> Final WAG, you have some Spooky Dooky going on in your hardware.
>
> Always possible, but 2008.0 seems to work fine.

Yes, but could be a dual-core bug in 2008.1


> And, I cannot see
> how dinking up fstab would have messed up things without making
> it totally obvious. The OS was finding the correct partitions and
> programs (after I got fstab straightened out).

I will agree, ONLY if fstab changes were done with unmounted partitions.


>
> The only thing opened was the rpm problem 36760 I think it was.
> That was a problem from the git-go, and copying over and running
> rpm for 2008.0 under 2008.1 worked, so I am inclined to leave it.

What can I say, it worked after boot for me. I do remember two rpm
updates from cooker so you might have snagged a bad one.


> I may take another try at it this coming weekend.

My suggestion, use e2label to label your partitions prior to install
to bypass the UUID setup.

I recommend my type install to reduce differences so that you might
get a known starting point for troubleshooting.

Go ahead and create/format your 2008.1 partition sda6 and then label it.
e2label /dev/sda6 2008_1

I picked cooker 2008.1 main, contrib non-free from easyurpm and saved
results into mm in a 2008.0 partition and chmod +x mm


If your iso is still on hard drive, loop mount the iso and
burn the ~12M /cdrom/x86_64/install/images/boot.iso to cd.

I have Pocket CD-RW cds for those small burns. (186 MB/21 MINUTE)
burns quick.

On boot, pick Hard drive
pick drive where iso resides.
pick partition where iso resides.
If in sub-directory, enter sub-directory name.
pick iso from list.

If not, use DVD.
I just wanted you to be able to rule out DVD and it's hardware.

Pick Install, and Custom in partition phase.
Leave 2008.1 partition for last.
At bottom of gui partition tool, click Toggle Expert mode.
Now you click each formatted partition,
click label, and control c to copy each label, OK
click mount point, click in box, control v to paste label.

If hard disk install, skip partition where iso resides.
Installer creates its own mount fstab entry.
After booting you can add correct values and delete the old one.

Now click sda6 partition, click Label, ctl c, ok
Click Mount point, / ok.
Note, we did not paste or change mount point for /.

Click Done.
Next screen, uncheck format box for / on sda6

Otherwise it will format sda6, loose label and give you grief on
boot.

In package selection you might also add kernel latest desktop source
and dkms to get closer to my install.

Take default Security Level.
Do click network configure and set your gateway if not dhcp

In Services config, disable
avahi-daemon
ct_sync
httpd
ksysguard
lisa
mandi
mysqld
named
netconsole
partmon
smb
webmin


Now if after boot and rpm is broke, then there are much less areas to
check for problem.

Next, set mirrors and manually update download/rpm packages

/wherever/mm
urpmi --wget wget --auto
urpmi --wget rpm --auto
urpmi --wget perl-URPM --auto
urpmi --wget urpmi --wget --auto
urpmi --wget rpmdrake --auto
urpmi --wget rpm-helper --auto
service httpd stop
service postfix stop
service cups stop

# now get all updates

urpmi --wget --auto-select --auto

check the rpm -qa again.

Cut/paste 2008.0 stanza into menu.lst, and reboot just to keep
everyone honest.

At this point your install would look pretty close to mine and should work.
That assumes the dual-core and Murphy did not get together.

Now you start dinking with fstab and whatnot.

On Wed, 13 Feb 2008 00:46:47 +0000 (UTC), Bit Twister wrote:
>
> At this point your install would look pretty close to mine and should work.
> That assumes the dual-core and Murphy did not get together.

Forgot to add, all 2008.1 partitions are in /,
bin home lib opt proc sys boot initrd lost+found root tmp dev media
usr etc mnt sbin var

swap can be shared with other installs assuming a clean shutdown
before new install.

Bit Twister wrote:
> On Tue, 12 Feb 2008 23:22:56 GMT, Jim Beard wrote:
>> Bit Twister wrote:
>
>> AMD Athlon 64 two dual-core 5000+
> Forgot you were dual core.
>
>>> memory 2 129 ns 512 MB (Single-bank Connection)
>> 1 GB, ?533 MHz? not sure what the ns is.
>
> I'll guess Nano Second for speed.
> Values are from mcc Hardware

Interesting. mcc Hardware lists memory controllers
but not the RAM itself. The invoice for the system
says 1 GB 533 MHz DDR2 Memory (1x1GB Module).

>
>>> Clean install using Custom during partition phase.
>>> /dev/sdb6 13G 5.2G 6.5G 45% /20081_64
>> Installed clean to /dev/sda6 (where non-used XP lives, plus
>> storage; all 2008.0 partitions with executables on sdb.
>
> Dang, misremembered start of thread, thought you installed on sdb.

2008.0 is on sdb. 2008.1 went onto sda6. XP on sda1
>
>> Initially, I had the installer format sda6. I did have to tinker
>> with fstab inconsequence.
>
> Good chance for setting wrong one there unless you cut/pasted comment
> above UUID.

I simply reentered from old fstab entries (backup) since I had not
changed any partitions, other than formatting sda6.
>
>> I don't know where the installer put swap for 2008.1. It is
>> sdb5 for 2008.0.
>
> I never have had to set swap location except day-one install.
> Custom always found/set it for me.
>
>>> Default Security level: High.
>> I dropped this to standard on my system.
>
> Hehehe, you sure are making it hard on re-creating your problem.
>
>
>> My first attempt was to install Gnome, and that aborted due
>> to package failure. Second attempt, with KDE selected, also
>> had a package failure, but I don't think that was significant.
>
> I had no package until update. On next test shot, try picking both KDE,GNOME

This is odd, as I was not given an option to select both. When
things started installing, I likewise did not have any choice about
what installed. The installer just went to work, without so much as
a by-your-leave.
>
>> Unscrambling fstab for 2008.1 was a chore. Unfamiliarity with
>> labels/UUIDs did not help.
>
> Did you dink with fstab before updates?

Yes, but I do not see how that could have caused rpm -qa
(actually, any invocation of rpm) to be unable to find a
symbol.

> Did you reboot after fstab changes?

Yes.
>
>> I no longer have 2008.1 on the system, but changes were first to
>> disable avahi and beagle, and later to remove them. Not long
>> before I removed 2008.1 totally, error messages complained that
>> Console-kit could not be found.
>
> Recommend install, updates, reboot, then start dinking around.

I have cleaned sda6 out, formatted it ext3 (installer set it to
native Linux, ext2).

Should I try the mirrors listed in 2008.1 or go to
easyurpmi.zarb.org? 2008.1 listed gatech, while easyurpmi gave
usc. Neither would work for me.
>
>>> Final WAG, you have some Spooky Dooky going on in your hardware.
>> Always possible, but 2008.0 seems to work fine.
>
> Yes, but could be a dual-core bug in 2008.1
>
>
>> And, I cannot see
>> how dinking up fstab would have messed up things without making
>> it totally obvious. The OS was finding the correct partitions and
>> programs (after I got fstab straightened out).
>
> I will agree, ONLY if fstab changes were done with unmounted partitions.

Point. Changes were to partitions not in use, but they were
mounted. Made changes. rebooted.
>
>> The only thing opened was the rpm problem 36760 I think it was.
>> That was a problem from the git-go, and copying over and running
>> rpm for 2008.0 under 2008.1 worked, so I am inclined to leave it.
>
> What can I say, it worked after boot for me. I do remember two rpm
> updates from cooker so you might have snagged a bad one.
>
>
>> I may take another try at it this coming weekend.
>
> My suggestion, use e2label to label your partitions prior to install
> to bypass the UUID setup.

I'll have to man e2label on that. I labeled hda6 2008beta in mcc.
Guess I will change that.
>
> I recommend my type install to reduce differences so that you might
> get a known starting point for troubleshooting.
>
> Go ahead and create/format your 2008.1 partition sda6 and then label it.
> e2label /dev/sda6 2008_1
>
> I picked cooker 2008.1 main, contrib non-free from easyurpm and saved
> results into mm in a 2008.0 partition and chmod +x mm
>
>
> If your iso is still on hard drive, loop mount the iso and
> burn the ~12M /cdrom/x86_64/install/images/boot.iso to cd.

Hmmm. I mounted the DVD and copied the boot.iso to my hard drive.
Then I simply used k3b to burn it to cd. K3b asked if I wanted to
burn it direct as an ISO, and I gave it a yes. Three times k3b
complained that the original and the burned copy differed, but
md5sum for boot.iso and for /dev/hda matched, so I assume the
copy was good. Will it boot? Or was copoying boot.iso from the
DVD illegit? Guess I will find out.

I think I remember instructions on how to loop mount an iso,
some months ago I think but Google archives will have it if needed.

> I have Pocket CD-RW cds for those small burns. (186 MB/21 MINUTE)
> burns quick.

Burning 11 MB or whatever it was took 2 or 3 minutes, K3B quits
adding to the CD when it runs out of bytes.
>
> On boot, pick Hard drive
> pick drive where iso resides.
> pick partition where iso resides.
> If in sub-directory, enter sub-directory name.
> pick iso from list.
>
> If not, use DVD.
> I just wanted you to be able to rule out DVD and it's hardware.
>
> Pick Install, and Custom in partition phase.
> Leave 2008.1 partition for last.
> At bottom of gui partition tool, click Toggle Expert mode.
> Now you click each formatted partition,
> click label, and control c to copy each label, OK
> click mount point, click in box, control v to paste label.

This makes mount point the label, right?
>
> If hard disk install, skip partition where iso resides.
> Installer creates its own mount fstab entry.
> After booting you can add correct values and delete the old one.
>
> Now click sda6 partition, click Label, ctl c, ok
> Click Mount point, / ok.
> Note, we did not paste or change mount point for /.
>
> Click Done.
> Next screen, uncheck format box for / on sda6
>
> Otherwise it will format sda6, loose label and give you grief on
> boot.
>
> In package selection you might also add kernel latest desktop source
> and dkms to get closer to my install.

If given the opportunity to select. Did not happen first time round.
>
> Take default Security Level.
> Do click network configure and set your gateway if not dhcp

Plain vanilla eth0 dhcp.
>
> In Services config, disable
> avahi-daemon
> ct_sync
> httpd
> ksysguard
> lisa
> mandi
> mysqld
> named
> netconsole
> partmon
> smb
> webmin
>
>
> Now if after boot and rpm is broke, then there are much less areas to
> check for problem.
>
> Next, set mirrors and manually update download/rpm packages
>
> /wherever/mm
> urpmi --wget wget --auto
> urpmi --wget rpm --auto
> urpmi --wget perl-URPM --auto
> urpmi --wget urpmi --wget --auto
> urpmi --wget rpmdrake --auto
> urpmi --wget rpm-helper --auto
> service httpd stop
> service postfix stop
> service cups stop
>
> # now get all updates
>
> urpmi --wget --auto-select --auto
>
> check the rpm -qa again.
>
> Cut/paste 2008.0 stanza into menu.lst, and reboot just to keep
> everyone honest.
>
> At this point your install would look pretty close to mine and should work.
> That assumes the dual-core and Murphy did not get together.
>
> Now you start dinking with fstab and whatnot.

For the moment, I'll print out the above and think about it for
a while. I'm almost prepared to do a full backup and then
contemplate a second try.

jim b.

--
UNIX is not user-unfriendly; it merely
expects users to be computer-friendly.

On Wed, 13 Feb 2008 02:06:21 GMT, Jim Beard wrote:
> Bit Twister wrote:
>
> Interesting. mcc Hardware lists memory controllers
> but not the RAM itself.

Hmmm, Under Memory, I clicked on Memory Module (A0)
Speed: was towards bottom.


>> I had no package until update. On next test shot, try picking both KDE,GNOME
>
> This is odd, as I was not given an option to select both. When
> things started installing,

First package screen has Radio selections
KDE
GNOME
Custom Install, I picked last one,

>> Did you dink with fstab before updates?
>
> Yes, but I do not see how that could have caused rpm -qa
> (actually, any invocation of rpm) to be unable to find a
> symbol.

Was thinking about crap spread into other partitions problem you had.

> I have cleaned sda6 out, formatted it ext3 (installer set it to
> native Linux, ext2).

Check that format pick list and set it Journalised FS: ext3


Click sda6 to reformat that 2008.1 partition
Click Toggle to Expert mode.
Click Type.
Select Journalised FS: ext3
Set label
Then format it.

>
> Should I try the mirrors listed in 2008.1 or go to
> easyurpmi.zarb.org? 2008.1 listed gatech, while easyurpmi gave
> usc. Neither would work for me.

My mm from easy

# cat /local/doc/mm
urpmi.addmedia main ftp://carroll.cac.psu.edu/pub/linux/...a/main/release with media_info/hdlist.cz
urpmi.addmedia contrib ftp://carroll.cac.psu.edu/pub/linux/...ontrib/release with media_info/hdlist.cz
urpmi.addmedia non-free ftp://carroll.cac.psu.edu/pub/linux/...n-free/release with media_info/hdlist.cz

>
>> I will agree, ONLY if fstab changes were done with unmounted partitions.
>
> Point. Changes were to partitions not in use, but they were
> mounted. Made changes. rebooted.

I make it a habit when changing fstab entry to umount first, make
change, mount it to verify change works.
No need to find a problem during boot.

> I'll have to man e2label on that. I labeled hda6 2008beta in mcc.
> Guess I will change that.

label text will not matter as long as it does not start with /

Since I have all install commands in my admin diary the command line
cut/paste is faster than mcc.


> Hmmm. I mounted the DVD and copied the boot.iso to my hard drive.
> Then I simply used k3b to burn it to cd. K3b asked if I wanted to
> burn it direct as an ISO, and I gave it a yes. Three times k3b
> complained that the original and the burned copy differed, but
> md5sum for boot.iso and for /dev/hda matched, so I assume the
> copy was good. Will it boot? Or was copoying boot.iso from the
> DVD illegit? Guess I will find out.

I burnt boot.iso from looped back iso mount.
Your k3b complaint would seem to indicate bad burn.

My 2008.0 32 bit k3b will not verify burnt cds.



> I think I remember instructions on how to loop mount an iso,
> some months ago I think but Google archives will have it if needed.


$ cat /local/bin/mount_iso
#************************************************* ******
#*
#* mount_iso - mount iso in loopback to /iso mount point
#*
#************************************************* ******

if [ $# -ne 1 ] ; then
tput clear
echo "Usage: $0 image.iso"
exit 1
fi

ISO_FN=$1

if [ ! -d /iso ] ; then
mkdir /iso
fi

if [ ! -r "$ISO_FN" ] ; then
echo "$ISO_FN does not exists or permissions set not readable"
exit 1
fi


mount -t auto -o users,ro,loop=/dev/loop0 $ISO_FN /iso

ls -al /iso

echo " "
echo "$iso_fn Mounted at /iso"


#************************ end mount_iso *********************



>> At bottom of gui partition tool, click Toggle Expert mode.
>> Now you click each formatted partition,
>> click label, and control c to copy each label, OK
>> click mount point, click in box, control v to paste label.
>
> This makes mount point the label, right?

That cut/paste does it. No reason that label and mount point have to
be the same value. Seems the click "Label" with a value is what
toggles installer to use labels instead of UUID or /dev/xxxx

# head -4 /20081_64/etc/fstab
LABEL=20081_64 / ext3 relatime 1 1
LABEL=accounts /accounts ext3 relatime 1 2
LABEL=bk_up /bk_up ext3 user,noauto,relatime 1 2
LABEL=fc7 /fc7 ext3 user,noauto,relatime 1 2

Just makes install faster, less typing/errors



>> In package selection you might also add kernel latest desktop source
>> and dkms to get closer to my install.
>
> If given the opportunity to select. Did not happen first time round.

Hmmm, Ok, pick Custom Install here
http://www.mandrake.tips.4.free.fr/i...mdv2008/17.png
There is where I pick major package groups and bottom of screen here
http://www.mandrake.tips.4.free.fr/i...v08samos16.png
set check box just above Help button bottom left.
Next screen will then allow you to do custom add/delete.

Do not bother with Beagle package. Auto update of some package will
reinstall it again later.

>
> Plain vanilla eth0 dhcp.
>>
>> /wherever/mm
>> urpmi --wget wget --auto
>> urpmi --wget rpm --auto
>> urpmi --wget perl-URPM --auto
>> urpmi --wget urpmi --wget --auto
>> urpmi --wget rpmdrake --auto
>> urpmi --wget rpm-helper --auto
>> service httpd stop
>> service postfix stop
>> service cups stop
>>
>> # now get all updates
>>
>> urpmi --wget --auto-select --auto
>>
>> check the rpm -qa again.
>>
>> Cut/paste 2008.0 stanza into menu.lst, and reboot just to keep
>> everyone honest.
>>
>
> For the moment, I'll print out the above and think about it for
> a while. I'm almost prepared to do a full backup and then
> contemplate a second try.

Keep an ascii copy.
Hey, put my urpmi, service commands at bottom of mm.

Just after boot, and
rpm -qa 'kernel*'
should work then if
ping -c1 yahoo.com
works, do a
/wherever/mm
and lean back and watch screen for about an hour.

On Wed, 13 Feb 2008 03:17:00 +0000 (UTC), Bit Twister wrote:
>
> Just after boot, and
> rpm -qa 'kernel*'
> should work then if
> ping -c1 yahoo.com
> works,

Get into mcc and disable your removable media selection in
Software Management -->
Configure media sources for install and update
Quit


> Then do a
> /wherever/mm
> and lean back and watch screen for about an hour.