12:25 p.m.
Has F17 been tested to see whether it will install on a Pentium 4?
Recently I installed F15 because I could not get the F16 installer to boot.
I can get the F13, F14 and F15 installers to boot without media.
The F16 installer was a no go, both without media and with a minimal CD.
In both cases, I got crashes that required pushing the reset button.
I never got to tell them about the iso file.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
12:32 p.m.
On 29 March 2012 11:25, Michael Hennebry <hennebry(a)web.cs.ndsu.nodak.edu> wrote:
Has F17 been tested to see whether it will install on a Pentium 4?
Recently I installed F15 because I could not get the F16 installer to boot.
I can get the F13, F14 and F15 installers to boot without media.
The F16 installer was a no go, both without media and with a minimal CD.
In both cases, I got crashes that required pushing the reset button.
I never got to tell them about the iso file.
What kind of Pentium IV. There were several versions and also multiple
motherboards all of which will affect whether this question can be
answered.
--
Stephen J Smoogen.
"The core skill of innovators is error recovery, not failure avoidance."
Randy Nelson, President of Pixar University.
"Years ago my mother used to say to me,... Elwood, you must be oh
so smart or oh so pleasant. Well, for years I was smart. I
recommend pleasant. You may quote me." —James Stewart as Elwood P. Dowd
2:30 p.m.
On Thu, 29 Mar 2012, Stephen John Smoogen wrote:
On 29 March 2012 11:25, Michael Hennebry
<hennebry(a)web.cs.ndsu.nodak.edu> wrote:
> Has F17 been tested to see whether it will install on a Pentium 4?
> Recently I installed F15 because I could not get the F16 installer to boot.
> I can get the F13, F14 and F15 installers to boot without media.
> The F16 installer was a no go, both without media and with a minimal CD.
> In both cases, I got crashes that required pushing the reset button.
> I never got to tell them about the iso file.
What kind of Pentium IV. There were several versions and also multiple
motherboards all of which will affect whether this question can be
answered.
Stepping 9.
Apparently it doesn't have NX.
At least it doesn't show up in /proc/cpuinfo
and there seems to be no BIOS switch for it.
It does have hyperthreading.
The machine is a DakTech with a D865GBF motherboard.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
3:15 p.m.
On 2012/03/29 14:30 (GMT-0500) Michael Hennebry composed:
Stephen John Smoogen wrote:
> What kind of Pentium IV. There were several versions and also
multiple
> motherboards all of which will affect whether this question can be
> answered.
Stepping 9.
Stepping doesn't say anything useful without family & model for context. I'm
looking at two different Hyperthreading P4 cpuinfo screens ATM, stepping of 3
on one, 5 on the other, but one is model 4 family 5 2048 cache (2.8GHz
Prescott I think), the other model 6 family 5 2048 cache (3.4GHz Cedar Mill
for sure).
Apparently it doesn't have NX.
At least it doesn't show up in /proc/cpuinfo
On the two I'm looking at NX shows on the 2nd output line between pbe & lm.
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
5:29 p.m.
On Thu, 29 Mar 2012, Felix Miata wrote:
On 2012/03/29 14:30 (GMT-0500) Michael Hennebry composed:
> Stephen John Smoogen wrote:
>> What kind of Pentium IV. There were several versions and also multiple
>> motherboards all of which will affect whether this question can be
>> answered.
> Stepping 9.
Stepping doesn't say anything useful without family & model for context. I'm
looking at two different Hyperthreading P4 cpuinfo screens ATM, stepping of 3
on one, 5 on the other, but one is model 4 family 5 2048 cache (2.8GHz
Prescott I think), the other model 6 family 5 2048 cache (3.4GHz Cedar Mill
for sure).
Family 15, model 2.
Here is /proc/cpuinfo:
processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 3.20GHz
stepping : 9
cpu MHz : 3200.000
cache size : 512 KB
physical id : 0
siblings : 2
core id : 0
cpu cores : 1
apicid : 0
initial apicid : 0
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts
acpi mmx fxsr sse sse2 ss ht tm pbe pebs bts cid xtpr
bogomips : 6384.10
clflush size : 64
cache_alignment : 128
address sizes : 36 bits physical, 32 bits virtual
power management:
processor : 1
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 3.20GHz
stepping : 9
cpu MHz : 3200.000
cache size : 512 KB
physical id : 0
siblings : 2
core id : 0
cpu cores : 1
apicid : 1
initial apicid : 1
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts
acpi mmx fxsr sse sse2 ss ht tm pbe pebs bts cid xtpr
bogomips : 6383.09
clflush size : 64
cache_alignment : 128
address sizes : 36 bits physical, 32 bits virtual
power management:
> Apparently it doesn't have NX.
> At least it doesn't show up in /proc/cpuinfo
On the two I'm looking at NX shows on the 2nd output line between pbe & lm.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
6:26 p.m.
On 2012/03/29 17:29 (GMT-0500) Michael Hennebry composed:
On Thu, 29 Mar 2012, Felix Miata wrote:
> Stepping doesn't say anything useful without family&
model for context. I'm
> looking at two different Hyperthreading P4 cpuinfo screens ATM, stepping of 3
> on one, 5 on the other, but one is model 4 family 5 2048 cache (2.8GHz
> Prescott I think), the other model 6 family 5 2048 cache (3.4GHz Cedar Mill
> for sure).
Those were supposed to read family 15 (but all Pentium 4 are 15 AFAICT).
Here is /proc/cpuinfo:
processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 3.20GHz
stepping : 9
cpu MHz : 3200.000
cache size : 512 KB
flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts
acpi mmx fxsr sse sse2 ss ht tm pbe pebs bts cid xtpr
address sizes : 36 bits physical, 32 bits virtual
Except for the apparent CPU speed error between URL and body text,
http://ark.intel.com/products/27493/Intel-Pentium-4-Processor-supporting-...
seems to describe what you have, a late production (socket 478) Northwood.
https://en.wikipedia.org/wiki/Pentium_4
If you want NX for your existing motherboard you'd probably have to find a
Prescott 478, which may never have existed, and even if it does, might not be
supported by your BIOS or any available upgrade for it.
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
8:34 p.m.
On Thu, 29 Mar 2012, Felix Miata wrote:
If you want NX for your existing motherboard you'd probably have
to find a
Prescott 478, which may never have existed, and even if it does, might not be
supported by your BIOS or any available upgrade for it.
I'm not looking for NX.
I only mentioned it because in another thread someone suggested that
NX might be the reason I couldn't get the F16 installer to boot.
For now, I'm going to settle for F15. That installer booted for me.
If the F17 installer won't boot for me,
then when F15 goes EOL, so does fedora.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
9:22 p.m.
On Mar 29, 2012 9:34 PM, "Michael Hennebry"
<hennebry(a)web.cs.ndsu.nodak.edu>
wrote:
On Thu, 29 Mar 2012, Felix Miata wrote:
> If you want NX for your existing motherboard you'd probably have to find
a
Prescott 478, which may never have existed, and even if it does, might
not be supported by your BIOS or any available upgrade for it.
I'm not looking for NX.
I only mentioned it because in another thread someone suggested that
NX might be the reason I couldn't get the F16 installer to boot.
For now, I'm going to settle for F15. That installer booted for me.
If the F17 installer won't boot for me,
then when F15 goes EOL, so does fedora.
"Crashes" isn't very helpful. If you would like developers to help
You should probably file a bug with what the crashes actually are.
Josh
12:42 a.m.
On Thu, 29 Mar 2012, Josh Boyer wrote:
On Mar 29, 2012 9:34 PM, "Michael Hennebry"
<hennebry(a)web.cs.ndsu.nodak.edu>
wrote:
>
> On Thu, 29 Mar 2012, Felix Miata wrote:
>
>> If you want NX for your existing motherboard you'd probably have to find
a Prescott 478, which may never have existed, and even if it does, might
not be supported by your BIOS or any available upgrade for it.
>
>
> I'm not looking for NX.
> I only mentioned it because in another thread someone suggested that
> NX might be the reason I couldn't get the F16 installer to boot.
> For now, I'm going to settle for F15. That installer booted for me.
> If the F17 installer won't boot for me,
> then when F15 goes EOL, so does fedora.
"Crashes" isn't very helpful. If you would like developers to help
You should probably file a bug with what the crashes actually are.
bug 808300 https://bugzilla.redhat.com/show_bug.cgi?id=808300
I gave it component anaconda because I couldn't find anything better.
It never gets to anaconda.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
10:08 p.m.
On Thu, 2012-03-29 at 20:34 -0500, Michael Hennebry wrote:
On Thu, 29 Mar 2012, Felix Miata wrote:
> If you want NX for your existing motherboard you'd probably have to find a
> Prescott 478, which may never have existed, and even if it does, might not be
> supported by your BIOS or any available upgrade for it.
I'm not looking for NX.
I only mentioned it because in another thread someone suggested that
NX might be the reason I couldn't get the F16 installer to boot.
For now, I'm going to settle for F15. That installer booted for me.
If the F17 installer won't boot for me,
then when F15 goes EOL, so does fedora.
I doubt your CPU has anything to do with 16 not booting. There's no
intentional change to 16 which would cause that, and I'm pretty sure we
still have quite a lot of people running on P4s.
It's much more likely to be some kind of BIOS issue or just a bad boot
medium. Why not just try a 17 image and see if it boots? Bandwidth is
cheap. Well, mostly.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
11:07 p.m.
On 03/29/2012 11:08 PM, Adam Williamson wrote:
On Thu, 2012-03-29 at 20:34 -0500, Michael Hennebry wrote:
> On Thu, 29 Mar 2012, Felix Miata wrote:
>
>> If you want NX for your existing motherboard you'd probably have to find a
>> Prescott 478, which may never have existed, and even if it does, might not be
>> supported by your BIOS or any available upgrade for it.
>
> I'm not looking for NX.
> I only mentioned it because in another thread someone suggested that
> NX might be the reason I couldn't get the F16 installer to boot.
> For now, I'm going to settle for F15. That installer booted for me.
> If the F17 installer won't boot for me,
> then when F15 goes EOL, so does fedora.
I doubt your CPU has anything to do with 16 not booting. There's no
intentional change to 16 which would cause that, and I'm pretty sure we
still have quite a lot of people running on P4s.
It's much more likely to be some kind of BIOS issue or just a bad boot
medium. Why not just try a 17 image and see if it boots? Bandwidth is
cheap. Well, mostly.
Indeed. One of my P4's is a Family 15, Model 2, Stepping 9.
F17 boots OK.
-Mike
7:47 a.m.
On Thu, Mar 29, 2012 at 11:08 PM, Adam Williamson <awilliam(a)redhat.com> wrote:
On Thu, 2012-03-29 at 20:34 -0500, Michael Hennebry wrote:
> On Thu, 29 Mar 2012, Felix Miata wrote:
>
> > If you want NX for your existing motherboard you'd probably have to find a
> > Prescott 478, which may never have existed, and even if it does, might not be
> > supported by your BIOS or any available upgrade for it.
>
> I'm not looking for NX.
> I only mentioned it because in another thread someone suggested that
> NX might be the reason I couldn't get the F16 installer to boot.
> For now, I'm going to settle for F15. That installer booted for me.
> If the F17 installer won't boot for me,
> then when F15 goes EOL, so does fedora.
I doubt your CPU has anything to do with 16 not booting. There's no
intentional change to 16 which would cause that, and I'm pretty sure we
still have quite a lot of people running on P4s.
Actually, there was a change in the ACPICA code that broke a number of
P4 motherboards entirely that went out with F16. That was in 2.6.39,
and F16 shipped with 3.1, and it wasn't fixed until last month.
I have no idea if that is the issue Michael is seeing with F16, but
it's possible
josh
10:39 a.m.
On Fri, 30 Mar 2012, Josh Boyer wrote:
Actually, there was a change in the ACPICA code that broke a number
of
P4 motherboards entirely that went out with F16. That was in 2.6.39,
and F16 shipped with 3.1, and it wasn't fixed until last month.
I have no idea if that is the issue Michael is seeing with F16, but
it's possible
I'll betcha. I got my DVD iso here:
http://mirrors.reflected.net/fedora/linux//releases/16/Fedora/i386/iso/
Index of /fedora/linux//releases/16/Fedora/i386/iso/
Icon Name Last modified Size Description[DIR]
Parent Directory -
[ ] Fedora-16-i386-CHECKSUM 04-Nov-2011 10:01 1.1K
[ ] Fedora-16-i386-DVD.iso 02-Nov-2011 22:18 3.5G
[ ] Fedora-16-i386-netinst.iso 02-Nov-2011 22:16 266M
Note the dates.
How do I find the most recent image?
I'd previously thought that that was the most recent image.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
11:24 a.m.
How do I find the most recent image?
There is only one Release image ever made and distributed by Fedora,
so if the checksum matches then you have the latest official image.
Updates are distributed electronically via yum in the repository
that is designated in /etc/yum.repos.d/fedora-updates.repo .
You can make your own updated .iso; for instance see
https://fedorahosted.org/pungi/wiki/PungiDocs and
https://bugzilla.redhat.com/show_bug.cgi?id=789101 .
This can be done from one release back [on Fedora 15 for Fedora 16]
by installing the forward versions of the packages pungi, anaconda,
fedora-kickstarts (previously spin-kickstarts), and [perhaps] lorax.
I build my own .isos a few times shortly before alpha, beta, and release.
After downloading the packages, it takes me about 25 minutes for Fedora 16,
10 minutes for Fedora 17. Use ext4 for /var/cache/pungi and the output
directory, a tmpfs for the work directory, and have >=4GB RAM; else
it will take a while longer.
--
4:15 p.m.
On Fri, 2012-03-30 at 08:47 -0400, Josh Boyer wrote:
On Thu, Mar 29, 2012 at 11:08 PM, Adam Williamson
<awilliam(a)redhat.com> wrote:
> On Thu, 2012-03-29 at 20:34 -0500, Michael Hennebry wrote:
>> On Thu, 29 Mar 2012, Felix Miata wrote:
>>
>> > If you want NX for your existing motherboard you'd probably have to
find a
>> > Prescott 478, which may never have existed, and even if it does, might not
be
>> > supported by your BIOS or any available upgrade for it.
>>
>> I'm not looking for NX.
>> I only mentioned it because in another thread someone suggested that
>> NX might be the reason I couldn't get the F16 installer to boot.
>> For now, I'm going to settle for F15. That installer booted for me.
>> If the F17 installer won't boot for me,
>> then when F15 goes EOL, so does fedora.
>
> I doubt your CPU has anything to do with 16 not booting. There's no
> intentional change to 16 which would cause that, and I'm pretty sure we
> still have quite a lot of people running on P4s.
Actually, there was a change in the ACPICA code that broke a number of
P4 motherboards entirely that went out with F16. That was in 2.6.39,
and F16 shipped with 3.1, and it wasn't fixed until last month.
I have no idea if that is the issue Michael is seeing with F16, but
it's possible
Then that's the *motherboard* that's the problem, not the CPU. =)
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
1:26 a.m.
In bugzilla bug 808300, Josh Boyer wrote:
D865GBF is one of the motherboards that was broken due to an ACPICA
change.
It was fixed in an updated kernel, but the F16 media cannot be updated.
See bugs 73007 727865.
Can you please try the nightly F17 live images, or the upcoming F17 beta?
I can try, but right now, I don't know where to find things.
The F17 alpha didn't work for me. The checksum was
d37440807791aaa595a4b9e4c2c9a4e381ef0770571b17eac552107e3fb6aeec.
Also, I don't have much of a clue about pungi,
which I gather is what I would have to use
to make an F16 installer with the new kernel.
F15 is my latest installed system and the pungi documentation I
could find didn't mention how to do forward versions.
All the above refer to hard drive install attempts without media.
My DVD burner doesn't work any more.
With F15, I can get to the installer without media.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
1:46 a.m.
In my most recent efforts, I added noapci to the kernel command line.
In the case of F17, the last line I could read before catatonia
included "unpacking initramfs".
F16 was similar, but it was to fast for me to read well.
Here are the grub stanzas:
title Install Fedora17 from iso
root (hd0,2)
pause have rooted (hd0,2)
find /isolinux17/vmlinuz
pause have finded
kernel /isolinux17/vmlinuz askmethod noacpi
pause have kerneled
find /isolinux17/initrd.img
pause have finded
initrd /isolinux17/initrd.img
pause have initrded
title Install Fedora16 from iso
root (hd0,2)
pause have rooted (hd0,2)
find /isolinux16/vmlinuz
pause have finded
kernel /isolinux16/vmlinuz askmethod noacpi
pause have kerneled
find /isolinux16/initrd.img
pause have finded
initrd /isolinux16/initrd.img
pause have initrded
In each case, all the pauses happened.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
8:30 a.m.
On Sat, Mar 31, 2012 at 01:26:26 -0500,
Michael Hennebry <hennebry(a)web.cs.ndsu.nodak.edu> wrote:
In bugzilla bug 808300, Josh Boyer wrote:
>D865GBF is one of the motherboards that was broken due to an ACPICA change.
>It was fixed in an updated kernel, but the F16 media cannot be updated.
>See bugs 73007 727865.
>Can you please try the nightly F17 live images, or the upcoming F17 beta?
I can try, but right now, I don't know where to find things.
The nightly live images are at:
http://alt.fedoraproject.org/pub/alt/nightly-composes/
4:39 p.m.
On Sat, 31 Mar 2012, Bruno Wolff III wrote:
On Sat, Mar 31, 2012 at 01:26:26 -0500,
Michael Hennebry <hennebry(a)web.cs.ndsu.nodak.edu> wrote:
> In bugzilla bug 808300, Josh Boyer wrote:
>> D865GBF is one of the motherboards that was broken due to an ACPICA
>> change.
>> It was fixed in an updated kernel, but the F16 media cannot be updated.
>> See bugs 73007 727865.
>> Can you please try the nightly F17 live images, or the upcoming F17 beta?
>
> I can try, but right now, I don't know where to find things.
The nightly live images are at:
http://alt.fedoraproject.org/pub/alt/nightly-composes/
Thanks.
Can they be booted without media?
I tried the following stanza:
title Install Fedora17-3-28 from iso
root (hd0,2)
pause have rooted (hd0,2)
find /isolinux17-3-28/vmlinuz0
pause have finded
kernel /isolinux17-3-28/vmlinuz0 askmethod noacpi root=/dev/ram0
pause have kerneled
find /isolinux17-3-28/initrd0.img
pause have finded
initrd /isolinux17-3-28/initrd0.img
pause have initrded
I got a dracut warning: signal caught
attempt to kill inint
no or empty root argument
It didn't boot.
Adding root=/dev/ram0 to the kernel line got me told no such device.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
9:34 a.m.
On Sat, Mar 31, 2012 at 16:39:56 -0500,
Michael Hennebry <hennebry(a)web.cs.ndsu.nodak.edu> wrote:
On Sat, 31 Mar 2012, Bruno Wolff III wrote:
Can they be booted without media?
Possibly, but I am not sure how to do it. I usually use livecd-iso-to-disk
to put images on flash drives and boot from them.
10:49 a.m.
On Sun, 1 Apr 2012, Bruno Wolff III wrote:
On Sat, Mar 31, 2012 at 16:39:56 -0500,
Michael Hennebry <hennebry(a)web.cs.ndsu.nodak.edu> wrote:
> On Sat, 31 Mar 2012, Bruno Wolff III wrote:
> Can they be booted without media?
Possibly, but I am not sure how to do it. I usually use livecd-iso-to-disk
to put images on flash drives and boot from them.
That could be a problem.
My machine won't boot with a USB drive attached and my burner does not work.
Can one boot using a partition that is an iso image?
I've mounted such partitions, but that is not the same as booting from one.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
11:21 a.m.
On 1 April 2012 09:49, Michael Hennebry <hennebry(a)web.cs.ndsu.nodak.edu> wrote:
On Sun, 1 Apr 2012, Bruno Wolff III wrote:
> On Sat, Mar 31, 2012 at 16:39:56 -0500,
> Michael Hennebry <hennebry(a)web.cs.ndsu.nodak.edu> wrote:
>>
>> On Sat, 31 Mar 2012, Bruno Wolff III wrote:
>> Can they be booted without media?
>
>
> Possibly, but I am not sure how to do it. I usually use livecd-iso-to-disk
> to put images on flash drives and boot from them.
That could be a problem.
My machine won't boot with a USB drive attached and my burner does not work.
Can one boot using a partition that is an iso image?
I've mounted such partitions, but that is not the same as booting from one.
Some BIOS's will, others won't. From the differences you have listed
with your hardware, I would go with no on yours.
--
Stephen J Smoogen.
"The core skill of innovators is error recovery, not failure avoidance."
Randy Nelson, President of Pixar University.
"Years ago my mother used to say to me,... Elwood, you must be oh
so smart or oh so pleasant. Well, for years I was smart. I
recommend pleasant. You may quote me." —James Stewart as Elwood P. Dowd
11:48 a.m.
On Sun, Apr 01, 2012 at 10:49:56AM -0500, Michael Hennebry wrote:
On Sun, 1 Apr 2012, Bruno Wolff III wrote:
>On Sat, Mar 31, 2012 at 16:39:56 -0500,
> Michael Hennebry <hennebry(a)web.cs.ndsu.nodak.edu> wrote:
>>On Sat, 31 Mar 2012, Bruno Wolff III wrote:
>>Can they be booted without media?
>
>Possibly, but I am not sure how to do it. I usually use livecd-iso-to-disk
>to put images on flash drives and boot from them.
That could be a problem.
My machine won't boot with a USB drive attached and my burner does not work.
Can one boot using a partition that is an iso image?
I've mounted such partitions, but that is not the same as booting from one.
I did not try that with the current anaconda but I expect that this, or
something similar, should work. At least Fedora 16 and earlier can be
booted to an installation image with the following grub2.cfg entry
(adjust all of that to a need and taste - this is only an example):
menuentry 'Anaconda' {
load_video
set gfxpayload=keep
insmod gzio
insmod part_msdos
insmod ext2
linux (hd0,msdos2)/boota/vmlinuz
repo=hd:/dev/sda<X>:/<path_to_a_directory_with_an_iso_image>/
initrd (hd0,msdos2)/boota/initrd.img
}
Obviously things like "/boota/" and
"/<path_to_a_directory_with_an_iso_image>/"
are relative to corresponding disk devices and not to /. 'vmlinux' and
'initrd.img' refer to files you will find, for example, in 'pxeboot'
directory of your distro.
It is a good idea not to keep your iso image on a partition you are
going to clobber while installing. :-) Making '/boota/' different from
'/boot/' and on a "set-aside" device could be handy, especially if you
are going to repeat an exercise above, but it is not strictly required.
Michal
2:26 p.m.
On Sun, 1 Apr 2012, Michal Jaegermann wrote:
On Sun, Apr 01, 2012 at 10:49:56AM -0500, Michael Hennebry wrote:
> My machine won't boot with a USB drive attached and my burner
does not work.
> Can one boot using a partition that is an iso image?
> I've mounted such partitions, but that is not the same as booting from one.
I did not try that with the current anaconda but I expect that this, or
something similar, should work. At least Fedora 16 and earlier can be
booted to an installation image with the following grub2.cfg entry
(adjust all of that to a need and taste - this is only an example):
menuentry 'Anaconda' {
load_video
set gfxpayload=keep
insmod gzio
insmod part_msdos
insmod ext2
linux (hd0,msdos2)/boota/vmlinuz
repo=hd:/dev/sda<X>:/<path_to_a_directory_with_an_iso_image>/
initrd (hd0,msdos2)/boota/initrd.img
}
Is there an equivalent in grub 1?
Obviously things like "/boota/" and
"/<path_to_a_directory_with_an_iso_image>/"
are relative to corresponding disk devices and not to /. 'vmlinux' and
'initrd.img' refer to files you will find, for example, in 'pxeboot'
directory of your distro.
It is a good idea not to keep your iso image on a partition you are
going to clobber while installing. :-) Making '/boota/' different from
'/boot/' and on a "set-aside" device could be handy, especially if you
are going to repeat an exercise above, but it is not strictly required.
The plan was to start with an F17 live CD.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
3:13 p.m.
On Sun, Apr 01, 2012 at 02:26:45PM -0500, Michael Hennebry wrote:
On Sun, 1 Apr 2012, Michal Jaegermann wrote:
>
>menuentry 'Anaconda' {
> load_video
> set gfxpayload=keep
> insmod gzio
> insmod part_msdos
> insmod ext2
> linux (hd0,msdos2)/boota/vmlinuz
repo=hd:/dev/sda<X>:/<path_to_a_directory_with_an_iso_image>/
> initrd (hd0,msdos2)/boota/initrd.img
>}
Is there an equivalent in grub 1?
Sure. It does not matter what a bootloader. The only important things
are files to use and kernel boot options. Just write yourself a
corresponding entry and off you go.
With older grub you do not write "(hd0,msdos2)" but "(hd0,1)" as this
is not that specific but syntax details are not essential here.
Michal
4:53 p.m.
On Sun, 1 Apr 2012, Michal Jaegermann wrote:
On Sun, Apr 01, 2012 at 02:26:45PM -0500, Michael Hennebry wrote:
> On Sun, 1 Apr 2012, Michal Jaegermann wrote:
>
> >
> >menuentry 'Anaconda' {
> > load_video
> > set gfxpayload=keep
> > insmod gzio
> > insmod part_msdos
> > insmod ext2
> > linux (hd0,msdos2)/boota/vmlinuz
repo=hd:/dev/sda<X>:/<path_to_a_directory_with_an_iso_image>/
> > initrd (hd0,msdos2)/boota/initrd.img
> >}
>
> Is there an equivalent in grub 1?
Sure. It does not matter what a bootloader. The only important things
are files to use and kernel boot options. Just write yourself a
corresponding entry and off you go.
I should have been more specific.
I suspect that module is the grub1 equivalent of insmod,
but load_video, set and even linux stump me.
The grub1 info manual doesn't give me much n the way of suspects.
In the case of linux, I strongly suspect that there is one.
I've often been told to type linux <options> at a boot: prompt.
With older grub you do not write "(hd0,msdos2)" but
"(hd0,1)" as this
is not that specific but syntax details are not essential here.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
8:30 p.m.
On Sun, Apr 01, 2012 at 04:53:32PM -0500, Michael Hennebry wrote:
>>Is there an equivalent in grub 1?
OK, so what do you really mean by "grub 1"? You either have
grub-0.97-... or grub2-1.99-... a.k.a. grub2. I thought that by
"grub 1" you meant the former.
I should have been more specific.
I suspect that module is the grub1 equivalent of insmod,
but load_video, set and even linux stump me.
Forget about stuff you do not want to bother to check in documentation.
Write a boot entry in configuration file of something you are using
and that will be about it. If you are using grub-0.97 then this
will look somewhat of this sort (only adjust _all_ settings to your
circumstances) and it will be in a file to which /etc/grub.conf
is a symlink:
title Anaconda
kernel (hd0,1)/boota/vmlinuz
repo=hd:/dev/sda<X>:/<path_to_a_directory_with_an_iso_image>/
initrd (hd0,1)/boota/initrd.img
If you are using some other bootloader then change all of that
accordingly.
Michal
10:05 p.m.
On Sun, 1 Apr 2012, Michal Jaegermann wrote:
On Sun, Apr 01, 2012 at 04:53:32PM -0500, Michael Hennebry wrote:
> >>Is there an equivalent in grub 1?
OK, so what do you really mean by "grub 1"? You either have
grub-0.97-... or grub2-1.99-... a.k.a. grub2. I thought that by
"grub 1" you meant the former.
The former, aka grub legacy, aka grub never known as grub2.
> I should have been more specific.
> I suspect that module is the grub1 equivalent of insmod,
> but load_video, set and even linux stump me.
Forget about stuff you do not want to bother to check in documentation.
Eventually I did manage to find them.
Sometimes this brain isn't at its best.
Write a boot entry in configuration file of something you are using
and that will be about it. If you are using grub-0.97 then this
will look somewhat of this sort (only adjust _all_ settings to your
circumstances) and it will be in a file to which /etc/grub.conf
is a symlink:
title Anaconda
kernel (hd0,1)/boota/vmlinuz
repo=hd:/dev/sda<X>:/<path_to_a_directory_with_an_iso_image>/
initrd (hd0,1)/boota/initrd.img
This approximately what I'd been trying to do.
This is what I'd been using to try to boot F17-alpha:
title Install Fedora17 from iso
root (hd0,2)
find /isolinux17/vmlinuz
kernel /isolinux17/vmlinuz askmethod
find /isolinux17/initrd.img
initrd /isolinux17/initrd.img
It didn't work very well trying to boot a F17 nightly live CD either:
title Install Fedora17-3-28 from iso
root (hd0,2)
find /isolinux17-3-28/vmlinuz0
kernel /isolinux17-3-28/vmlinuz0 askmethod
find /isolinux17-3-28/initrd0.img
initrd /isolinux17-3-28/initrd0.img
In neither case did adding noacpi to the kernel line help.
In both cases, the iso was a file.
I was wondering if it could be done with a live CD iso as a partition.
If you are using some other bootloader then change all of that
accordingly.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
11:03 p.m.
On Sun, Apr 01, 2012 at 10:05:50PM -0500, Michael Hennebry wrote:
This approximately what I'd been trying to do.
This is what I'd been using to try to boot F17-alpha:
title Install Fedora17 from iso
root (hd0,2)
find /isolinux17/vmlinuz
kernel /isolinux17/vmlinuz askmethod
AFAIK 'askmethod' is no longer valid. Use repo=...
And why on earth you need "find"? Either those files are there or
they are missing and you will not boot.
Again. This "/isolinux17/" is relative to (hd0,2). Right?
If not then locations above are wrong.
Michal
11:30 a.m.
On Sun, 1 Apr 2012, Michal Jaegermann wrote:
On Sun, Apr 01, 2012 at 10:05:50PM -0500, Michael Hennebry wrote:
>
> This approximately what I'd been trying to do.
> This is what I'd been using to try to boot F17-alpha:
> title Install Fedora17 from iso
> root (hd0,2)
> find /isolinux17/vmlinuz
> kernel /isolinux17/vmlinuz askmethod
AFAIK 'askmethod' is no longer valid. Use repo=...
Arrrg. Syntax!!!!!!!!!!!!!!!
kernel /isolinux17/vmlinuz repo=hd:LABEL=/home1:/iso17/
good?
And why on earth you need "find"? Either those files are
there or
they are missing and you will not boot.
I'd used it for debugging. I'd forgotten to delete the following pause.
Again. This "/isolinux17/" is relative to (hd0,2). Right?
So right.
If not then locations above are wrong.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
12:55 p.m.
On Mon, 2 Apr 2012, Michael Hennebry wrote:
On Sun, 1 Apr 2012, Michal Jaegermann wrote:
> On Sun, Apr 01, 2012 at 10:05:50PM -0500, Michael Hennebry wrote:
>>
>> This approximately what I'd been trying to do.
>> This is what I'd been using to try to boot F17-alpha:
>> title Install Fedora17 from iso
>> root (hd0,2)
>> find /isolinux17/vmlinuz
>> kernel /isolinux17/vmlinuz askmethod
>
> AFAIK 'askmethod' is no longer valid. Use repo=...
Arrrg. Syntax!!!!!!!!!!!!!!!
kernel /isolinux17/vmlinuz repo=hd:LABEL=/home1:/iso17/
good?
Here is the grub stanza I used this time:
title Install Fedora17 from iso
root (hd0,2)
pause have rooted (hd0,2)
find /isolinux17/vmlinuz
pause have finded
kernel /isolinux17/vmlinuz kernel /isolinux17/vmlinuz repo=hd:LABEL=/home1:/iso17/
pause kernel /isolinux17/vmlinuz kernel /isolinux17/vmlinuz
repo=hd:LABEL=/home1:/iso17/
find /isolinux17/initrd.img
pause have finded
initrd /isolinux17/initrd.img
pause have initrded
All the finds gave the expected results. All the pauses happened.
The last thing I saw before the screen went blurry was
... unpacking initramfs
I took a picture of the final screen.
It is attached.
The original color JPEG was over 2MB, so I whittled.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
1:10 p.m.
On Mon, 2 Apr 2012, Michael Hennebry wrote:
Here is the grub stanza I used this time:
title Install Fedora17 from iso
root (hd0,2)
pause have rooted (hd0,2)
find /isolinux17/vmlinuz
pause have finded
kernel /isolinux17/vmlinuz kernel /isolinux17/vmlinuz
repo=hd:LABEL=/home1:/iso17/
pause kernel /isolinux17/vmlinuz kernel /isolinux17/vmlinuz
repo=hd:LABEL=/home1:/iso17/
find /isolinux17/initrd.img
pause have finded
initrd /isolinux17/initrd.img
pause have initrded
All the finds gave the expected results. All the pauses happened.
The last thing I saw before the screen went blurry was
... unpacking initramfs
I took a picture of the final screen.
It is attached.
The original color JPEG was over 2MB, so I whittled.
Adding noacpi to the kernel line didn't allow booting,
but it did change the final screen.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
1:40 p.m.
I just installed 32 bit to 2.8GHz 1024 cache family 15 model 3 stepping 4
without NX on i915 using
http://mirrors.kernel.org/fedora/development/17/i386/os/images/boot.iso
Since yesterday I'd been trying to start install from HD using this Grub stanza:
title Install Rawhide (hd0,13)
kernel (hd0,2)/fedora/vmlinuz
repo=http://mirrors.us.kernel.org/fedora/development/17/i386/os
gateway=192.168.1.1 dns=192.168.1.1 graphical ip=192.168.0.2
netmask=255.255.255.0 hostname=myhost noselinux noipv6 vga=0x318
resolution=1024x768
initrd (hd0,2)/fedora/initrd.img
That got me along _much_ farther than you, halting as described in today's
thread "Unable to read package metadata".
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
1:44 p.m.
[corrected typo]
I just installed 32 bit to 2.8GHz 1024 cache family 15 model 3 stepping 4
without NX on i915 using
http://mirrors.kernel.org/fedora/development/17/i386/os/images/boot.iso
Since yesterday I'd been trying to start install from HD using this Grub stanza:
title Install Rawhide (hd0,13)
kernel (hd0,2)/fedora/vmlinuz
repo=http://mirrors.us.kernel.org/fedora/development/17/i386/os
gateway=192.168.1.1 dns=192.168.1.1 graphical ip=192.168.1.2
netmask=255.255.255.0 hostname=myhost noselinux noipv6 vga=0x318
resolution=1024x768
initrd (hd0,2)/fedora/initrd.img
That got me along _much_ farther than you, halting as described in today's
thread "Unable to read package metadata".
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
1:34 p.m.
On Mon, Apr 02, 2012 at 11:30:47AM -0500, Michael Hennebry wrote:
On Sun, 1 Apr 2012, Michal Jaegermann wrote:
>
>AFAIK 'askmethod' is no longer valid. Use repo=...
Arrrg. Syntax!!!!!!!!!!!!!!!
kernel /isolinux17/vmlinuz repo=hd:LABEL=/home1:/iso17/
good?
I really do not know if "LABEL=/home1" device specification will
work in this context. It might but that would depend on what really
happens in initrd.img you are using. Try and see.
I prefer to be explicit in such situations, i.e.
kernel (hd0,2)/isolinux17/vmlinuz ...
initrd (hd0,2)/....
but if 'root' is set correctly this should not matter.
It is easy to experiment with all these details from a 'grub>' prompt
while booting. Your interactive adjustments are not permanent but
this is not that important here.
In any case I do not have a corresponding iso image on hand so I cannot
go through the whole thing but using vmlinuz and initrd.img from F17 and
booting like the above I am getting at least a dracut debug shell prompt
(after expected here "no suitable images" complaints).
Michal
4:58 p.m.
On Mon, 2 Apr 2012, Michal Jaegermann wrote:
On Mon, Apr 02, 2012 at 11:30:47AM -0500, Michael Hennebry wrote:
> On Sun, 1 Apr 2012, Michal Jaegermann wrote:
> >
> >AFAIK 'askmethod' is no longer valid. Use repo=...
>
> Arrrg. Syntax!!!!!!!!!!!!!!!
>
> kernel /isolinux17/vmlinuz repo=hd:LABEL=/home1:/iso17/
> good?
I really do not know if "LABEL=/home1" device specification will
work in this context. It might but that would depend on what really
happens in initrd.img you are using. Try and see.
I prefer to be explicit in such situations, i.e.
kernel (hd0,2)/isolinux17/vmlinuz ...
initrd (hd0,2)/....
but if 'root' is set correctly this should not matter.
It is easy to experiment with all these details from a 'grub>' prompt
while booting. Your interactive adjustments are not permanent but
this is not that important here.
It would be if I were that good a typist.
If I were that good a typist this thread
would have been named F16 vs Pentium 4.
'Twas suggested to me that I make an F16 install image with a new kernel,
but that would involve a "forward version".
In context, 'twas fairly obvious what that meant,
but I couldn't find a clue how to do it.
In any case I do not have a corresponding iso image on hand so I
cannot
go through the whole thing but using vmlinuz and initrd.img from F17 and
booting like the above I am getting at least a dracut debug shell prompt
(after expected here "no suitable images" complaints).
I tried it with /dev and still got crap.
I think it's time for me to give up.
Eventually, I'll replace my DVD burner.
From now on, I'll just have to settle for
letting installers
slog through DVDs.
That assumes I don't zap my video card or something again.
This install from hd crap is just too hard.
I've never had an easy time installing, but this is the worst ever.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
4:46 p.m.
My most recent attempt to install F17, used the following grub stanza:
title Felix17
find /felix17/vmlinuz
pause find /felix17/vmlinuz
kernel (hd0,2)/felix17/vmlinuz
repo=http://mirrors.us.kernel.org/fedora/development/17/i386/os noapic noacpi nolapic
pause kernel
initrd (hd0,2)/felix17/initrd.img
pause initrd (hd0,2)/felix17/initrd.img
The kernel and initrd images come from
http://mirrors.us.kernel.org/fedora/development/17/i386/os/isolinux .
After the last pause, it reboots as if I had hit the reset button.
Arrrrg.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
5:17 p.m.
On 2012/04/05 16:46 (GMT-0500) Michael Hennebry composed:
My most recent attempt to install F17, used the following grub
stanza:
title Felix17
find /felix17/vmlinuz
pause find /felix17/vmlinuz
kernel (hd0,2)/felix17/vmlinuz
repo=http://mirrors.us.kernel.org/fedora/development/17/i386/os noapic noacpi nolapic
pause kernel
initrd (hd0,2)/felix17/initrd.img
pause initrd (hd0,2)/felix17/initrd.img
The kernel and initrd images come from
http://mirrors.us.kernel.org/fedora/development/17/i386/os/isolinux .
After the last pause, it reboots as if I had hit the reset button.
Seems the expected result of there being no /felix17/initrd.img or
/felix17/vmlinuz on the 3rd partition on the first BIOS HD. Do these in fact
exist? The find command should tell you where it found files by those names
in the following form:
(hd0,1)
(hd0,2)
(hd0,5)
without any need for a pause command.
Similar for the kernel command (and the initrd command):
[Linux-bzImage, setup=0x####, size=0x######]
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
8:17 p.m.
On Thu, 5 Apr 2012, Felix Miata wrote:
On 2012/04/05 16:46 (GMT-0500) Michael Hennebry composed:
> My most recent attempt to install F17, used the following grub stanza:
> title Felix17
> find /felix17/vmlinuz
> pause find /felix17/vmlinuz
> kernel (hd0,2)/felix17/vmlinuz
repo=http://mirrors.us.kernel.org/fedora/development/17/i386/os noapic noacpi
nolapic
> pause kernel
> initrd (hd0,2)/felix17/initrd.img
> pause initrd (hd0,2)/felix17/initrd.img
> The kernel and initrd images come from
> http://mirrors.us.kernel.org/fedora/development/17/i386/os/isolinux .
> After the last pause, it reboots as if I had hit the reset button.
Seems the expected result of there being no /felix17/initrd.img or
/felix17/vmlinuz on the 3rd partition on the first BIOS HD. Do these in fact
exist? The find command should tell you where it found files by those names
in the following form:
(hd0,1)
(hd0,2)
(hd0,5)
It did, (hd0,2).
without any need for a pause command.
The pause commands are so I can tell how
far booting gets before it craps out on me.
'Tis annoying to watch what little I might
understand disappear before I can read it
or even tell where it came from.
Similar for the kernel command (and the initrd command):
[Linux-bzImage, setup=0x####, size=0x######]
If this was there, I didn't see it.
What is the significance? Google has not helped.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
9:02 p.m.
On 2012/04/05 20:17 (GMT-0500) Michael Hennebry composed:
On Thu, 5 Apr 2012, Felix Miata wrote:
>> After the last pause, it reboots as if I had hit the reset
button.
> Seems the expected result of there being no /felix17/initrd.img
or
> /felix17/vmlinuz on the 3rd partition on the first BIOS HD. Do these in fact
> exist? The find command should tell you where it found files by those names
> in the following form:
> (hd0,1)
> (hd0,2)
> (hd0,5)
It did, (hd0,2).
> without any need for a pause command.
The pause commands are so I can tell how
far booting gets before it craps out on me.
'Tis annoying to watch what little I might
understand disappear before I can read it
or even tell where it came from.
> Similar for the kernel command (and the initrd command):
> [Linux-bzImage, setup=0x####, size=0x######]
If this was there, I didn't see it.
I just checked with F16. It's Grub Legacy produces no such response,
surprisingly (or maybe not), considering openSUSE's and (IIRC) Mandriva's and
all (AFAIK) distros' Grub Legacys (which support filename extensions no
longer than 3 characters for FAT filesystem compatibility) not derived from
RedHat do.
What is the significance? Google has not helped.
Maybe it's time to dispense with the script (menu.lst/grub.conf). When the
Grub menu appears, exit to the grub> prompt and execute the individual
commands [help,find,root,kernel,initrd,boot,reboot] as you're inclined to see
messages without proceeding to another step until you're ready.
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
12:35 p.m.
On Thu, 5 Apr 2012, Felix Miata wrote:
On 2012/04/05 20:17 (GMT-0500) Michael Hennebry composed:
> The pause commands are so I can tell how
> far booting gets before it craps out on me.
> 'Tis annoying to watch what little I might
> understand disappear before I can read it
> or even tell where it came from.
Maybe it's time to dispense with the script (menu.lst/grub.conf).
When the
Grub menu appears, exit to the grub> prompt and execute the individual
commands [help,find,root,kernel,initrd,boot,reboot] as you're inclined to see
messages without proceeding to another step until you're ready.
I don't understand how that would help.
It seems to me that that would leave me with even less documentation.
I installed openSUSE just to see if I could.
I could, but not without problems.
The first one was that it hung after its automatic reboot.
Pressing reset got the job done.
The other is that it took me a while to discover that to change
my boot options I needed to edit menu.lst , not grub.conf .
I was really panicking for a while.
Still no go on F16 or F17.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
1:15 p.m.
On 2012/04/06 12:35 (GMT-0500) Michael Hennebry composed:
Felix Miata wrote:
> Maybe it's time to dispense with the script
(menu.lst/grub.conf)
CLUE
^^^^^^^^^^^^^^^^^^ CLUE :-)
I don't understand how that would help.
It seems to me that that would leave me with even less documentation.
Like running an app in an xterm, otherwise invisible messages may provide a
necessary clue to solving a problem.
I installed openSUSE just to see if I could.
I could, but not without problems.
The first one was that it hung after its automatic reboot.
By default, first reboot is not a reboot but a kexec. Kexec doesn't always
work as intended.
Pressing reset got the job done.
Meaning openSUSE 12.1 is successfully installed and bootable?
The other is that it took me a while to discover that to change
my boot options I needed to edit menu.lst , not grub.conf .
See above. :-)
In F17 neither will work. It uses Grub2.
I was really panicking for a while.
Still no go on F16 or F17.
:-(
How about "cloning" your F15, and upgrading that to F16 with latest fixes
with Yum?
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
3:24 p.m.
On Fri, 6 Apr 2012, Felix Miata wrote:
On 2012/04/06 12:35 (GMT-0500) Michael Hennebry composed:
> Felix Miata wrote:
>> Maybe it's time to dispense with the script (menu.lst/grub.conf)
CLUE ^^^^^^^^^^^^^^^^^^ CLUE
:-)
> I don't understand how that would help.
> It seems to me that that would leave me with even less documentation.
Like running an app in an xterm, otherwise invisible messages may provide a
necessary clue to solving a problem.
That means that from the command line,
grub will emit messages that I would not see using pause?
> I installed openSUSE just to see if I could.
> I could, but not without problems.
> The first one was that it hung after its automatic reboot.
By default, first reboot is not a reboot but a kexec. Kexec doesn't always
work as intended.
> Pressing reset got the job done.
Meaning openSUSE 12.1 is successfully installed and bootable?
Yes.
How about "cloning" your F15, and upgrading that to F16
with latest fixes
with Yum?
I suppose I could clone my F15 by copying its partition,
but I thought that yum could only upgrade package versions,
not fedora versions.
man yum seems to concur.
Doing an upgrade with the installer would require running the installer,
something that hasn't worked well for me.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
4:35 p.m.
On 2012/04/06 15:24 (GMT-0500) Michael Hennebry composed:
On Fri, 6 Apr 2012, Felix Miata wrote:
> Like running an app in an xterm, otherwise invisible messages
may provide a
> necessary clue to solving a problem.
That means that from the command line,
grub will emit messages that I would not see using pause?
Presumably.
>> I installed openSUSE just to see if I could.
>> I could, but not without problems.
>> The first one was that it hung after its automatic reboot.
> By default, first reboot is not a reboot but a kexec. Kexec
doesn't always
> work as intended.
>> Pressing reset got the job done.
> Meaning openSUSE 12.1 is successfully installed and bootable?
Yes.
Which means something peculiar about the combination of your hardware and F16/17.
> How about "cloning" your F15, and upgrading that to
F16 with latest fixes
> with Yum?
I suppose I could clone my F15 by copying its partition,
but I thought that yum could only upgrade package versions,
not fedora versions.
man yum seems to concur.
If it does, it's not something I ever noticed. It's not uncommon for man
pages to be in need of update. Updating versions via package manager I've
done more than a few times on Fedora, routinely on openSUSE.
http://fedoraproject.org/wiki/How_to_use_PreUpgrade
Just be sure that the first thing done after cloning is ensuring UUID and
volume label on copy and clone do not match, and that on the clone fstab and
Grub's menu are matched to the clone's location. Also, a typical cloning
process does not clone Grub. That invariably must be done separately. Failing
the fstab & Grub menu post-cloning steps would almost surely result in damage
to the F15 source, if not complete destruction, once is upgrade is begun.
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
7:05 p.m.
On Fri, 6 Apr 2012, Felix Miata wrote:
On 2012/04/06 15:24 (GMT-0500) Michael Hennebry composed:
> On Fri, 6 Apr 2012, Felix Miata wrote:
>> How about "cloning" your F15, and upgrading that
to F16 with latest fixes
>> with Yum?
> I suppose I could clone my F15 by copying its partition,
> but I thought that yum could only upgrade package versions,
> not fedora versions.
> man yum seems to concur.
If it does, it's not something I ever noticed. It's not uncommon for man
pages to be in need of update. Updating versions via package manager I've
done more than a few times on Fedora, routinely on openSUSE.
http://fedoraproject.org/wiki/How_to_use_PreUpgrade
Verrrry interesting.
It seems that yum doesn't know about it either.
So far as yum is concerned, it's just installing another boring package.
I just hope that this mechanism doesn't hit the same wall.
Just be sure that the first thing done after cloning is ensuring UUID
and
volume label on copy and clone do not match, and that on the clone fstab and
Grub's menu are matched to the clone's location. Also, a typical cloning
process does not clone Grub. That invariably must be done separately. Failing
the fstab & Grub menu post-cloning steps would almost surely result in damage
to the F15 source, if not complete destruction, once is upgrade is begun.
cp /dev/sdb17 /dev/sdb27
wouldn't get the first sector of the partition?
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
8:36 p.m.
On 2012/04/06 19:05 (GMT-0500) Michael Hennebry composed:
Felix Miata wrote:
Verrrry interesting.
It seems that yum doesn't know about it either.
So far as yum is concerned, it's just installing another boring package.
I just hope that this mechanism doesn't hit the same wall.
> Just be sure that the first thing done after cloning is ensuring
UUID and
> volume label on copy and clone do not match, and that on the clone fstab and
> Grub's menu are matched to the clone's location. Also, a typical cloning
> process does not clone Grub. That invariably must be done separately. Failing
> the fstab& Grub menu post-cloning steps would almost surely result in damage
> to the F15 source, if not complete destruction, once is upgrade is begun.
cp /dev/sdb17 /dev/sdb27
wouldn't get the first sector of the partition?
You can't clone a partition with cp, only files. Better to use something like
"rsync -av --exclude 'lost+found'", and still you wouldn't have a
partition
clone, but only a complete filesystem copy, which _is_ the kind of clone you
want for this purpose, as long as it isn't performed using a currently booted
OS as source. Also there is software specifically made for cloning like you
need. I do all my cloning with DFSee, which isn't free, but is multi-platform
executable, and compatible, meaning among other things I can run it any time
I want without a reboot first.
The part of Grub that initiates boot isn't in the filesystem though. Grub's
files will get copied, but it still needs to be installed to the clone if it
is to be used by the clone for booting.
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
5:08 p.m.
It's now been moe than a month sicne I started trying to install F16.
Grrrrrrr.
On Fri, 6 Apr 2012, Felix Miata wrote:
On 2012/04/06 15:24 (GMT-0500) Michael Hennebry composed:
> On Fri, 6 Apr 2012, Felix Miata wrote:
>> How about "cloning" your F15, and upgrading that
to F16 with latest fixes
>> with Yum?
I made a filesystem in a new partition and copied to it as follows:
from /media as root:
cp -a sata400-3-slash/* sata-400-17
I edited the target's fstab.
LABEL=sata400-17 / ext3 defaults 1 1
Just be sure that the first thing done after cloning is ensuring UUID
and
volume label on copy and clone do not match, and that on the clone fstab and
Grub's menu are matched to the clone's location. Also, a typical cloning
process does not clone Grub. That invariably must be done separately. Failing
the fstab & Grub menu post-cloning steps would almost surely result in damage
to the F15 source, if not complete destruction, once is upgrade is begun.
My first F15 install includes four partitions,
one for / , one for /boot and two for swap.
The copy uses the same /boot and swap partitions.
Here are my grub stanzas:
title Fedora 15 (2.6.42.12-1.fc15.i686.PAE) sdb17
root (hd1,1)
kernel /vmlinuz-2.6.42.12-1.fc15.i686.PAE ro root=/dev/sdb17
initrd /initramfs-2.6.42.12-1.fc15.i686.PAE.img
title Fedora 15 (2.6.42.12-1.fc15.i686.PAE) sdb3
root (hd1,1)
kernel /vmlinuz-2.6.42.12-1.fc15.i686.PAE ro
root=UUID=ce978949-d044-4020-8001-02454648a64e rd_NO_LUKS rd_NO_LVM rd_NO_MD rd_NO_DM
LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYTABLE=us rhgb quiet
initrd /initramfs-2.6.42.12-1.fc15.i686.PAE.img
I can boot sdb3.
sdb17 fails.
The last line before catatonia included:
systemd-readahead-collect[406]: Failed to open pack file Permission denied
There were other failures, but I mostly don't remember what they wer.
One mentioned systemctl states and NetworkManager.service .
Any ideas?
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
1:01 a.m.
On 2012/04/08 17:08 (GMT-0500) Michael Hennebry composed:
Felix Miata wrote:
> On 2012/04/06 15:24 (GMT-0500) Michael Hennebry composed:
>> On Fri, 6 Apr 2012, Felix Miata wrote:
>>> How about "cloning" your F15, and upgrading
that to F16 with latest fixes
>>> with Yum?
I made a filesystem in a new partition and copied to it as follows:
Booted to what? File copy can't be expected to work satisfactorily if source
is running.
from /media as root:
cp -a sata400-3-slash/* sata-400-17
You have far more confidence in the competency of cp than I. Until last week,
I'd never cloned a partition except by one of two ways: 1-DFSee (sector by
sector); 2-MC, however it manages. Last time I tried MC I ran into trouble
that probably was little different from what you just did, so last week I
switched to doing file copies via rsync.
I edited the target's fstab.
LABEL=sata400-17 / ext3 defaults 1 1
> Just be sure that the first thing done after cloning is ensuring
UUID and
> volume label on copy and clone do not match, and that on the clone fstab and
> Grub's menu are matched to the clone's location. Also, a typical cloning
> process does not clone Grub. That invariably must be done separately. Failing
> the fstab& Grub menu post-cloning steps would almost surely result in damage
> to the F15 source, if not complete destruction, once is upgrade is begun.
My first F15 install includes four partitions,
one for / , one for /boot and two for swap.
The copy uses the same /boot and swap partitions.
I suppose someone with a lot of experience could get away with sharing a
/boot partition, but based upon your experiences of the past month I wouldn't
expect that group to include you. I would have merged the content of the
/boot partition into the /boot directory on the new /.
Here are my grub stanzas:
title Fedora 15 (2.6.42.12-1.fc15.i686.PAE) sdb17
root (hd1,1)
kernel /vmlinuz-2.6.42.12-1.fc15.i686.PAE ro root=/dev/sdb17
initrd /initramfs-2.6.42.12-1.fc15.i686.PAE.img
Leaving off some of those cmdline options could be a problem. I'm not
familiar with all of them, and so would have included all I don't understand.
title Fedora 15 (2.6.42.12-1.fc15.i686.PAE) sdb3
root (hd1,1)
kernel /vmlinuz-2.6.42.12-1.fc15.i686.PAE ro
root=UUID=ce978949-d044-4020-8001-02454648a64e rd_NO_LUKS rd_NO_LVM rd_NO_MD rd_NO_DM
LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYTABLE=us rhgb quiet
initrd /initramfs-2.6.42.12-1.fc15.i686.PAE.img
I can boot sdb3.
sdb17 fails.
The last line before catatonia included:
systemd-readahead-collect[406]: Failed to open pack file Permission denied
There were other failures, but I mostly don't remember what they wer.
One mentioned systemctl states and NetworkManager.service .
Any ideas?
1-Do it again with rsync or a program designed specifically for cloning.
2-Merge /boot partition into the new /.
3-Either find out if any of those cmdline options can play a part in the
problem and put back the ones you don't know you don't need, or add them all
back except rhgb & quiet
4-Add option enforcing=0 to cmdline to ensure SELinux isn't in the way
5-Use the Grub on hd1,1 to start both, but for the clone include a hd1,16
configfile stanza until you get it booted, then install Grub on sdb17 and
afterward chainload from hd1,1 to hd1,16, where updates will ultimately be
applied to grub.conf only for the clone on account of new kernels.
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
4:29 p.m.
On Mon, 9 Apr 2012, cornel panceac wrote:
this may be of help, too.
https://fedoraproject.org/wiki/Upgrading_Fedora_using_yum
Indeed it might.
It looks hairier than preupgrade, but it avoids anaconda.
That might avoid the land mine that causes
crashes when I try to do an install.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
11:59 p.m.
On Mon, 9 Apr 2012, cornel panceac wrote:
this may be of help, too.
https://fedoraproject.org/wiki/Upgrading_Fedora_using_yum
I'm trying it, but I'm feeling fuzzy at
Install the new fedora gpg key for the version you are updating to
Would this command do it:
rpm --import https://fedorproject.org/static/A82BA4B7.txt
yum --releasever=16 --disableplugin=presto distro-sync
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
4:22 p.m.
On Mon, 9 Apr 2012, Felix Miata wrote:
On 2012/04/08 17:08 (GMT-0500) Michael Hennebry composed:
> Felix Miata wrote:
>> On 2012/04/06 15:24 (GMT-0500) Michael Hennebry composed:
>>> On Fri, 6 Apr 2012, Felix Miata wrote:
>>>> How about "cloning" your F15, and upgrading that to F16 with
latest
fixes
>>>> with Yum?
> I made a filesystem in a new partition and copied to it as follows:
Booted to what? File copy can't be expected to work satisfactorily if source
is running.
Source wasn't running.
I'm still running F14.
F15 was installed primarily to establish that I could.
> from /media as root:
> cp -a sata400-3-slash/* sata-400-17
You have far more confidence in the competency of cp than I. Until last week,
I'd never cloned a partition except by one of two ways: 1-DFSee (sector by
sector); 2-MC, however it manages. Last time I tried MC I ran into trouble
that probably was little different from what you just did, so last week I
switched to doing file copies via rsync.
I was under the impression that rsync was just a convenient version of cp.
Is there something that rsync gets right that cp gets wrong?
I've read cp vs. rsync theads.
From those, rsync would have been much better if
the copy had
been interrupted for some reason.
Fortunately it wasn't.
> I edited the target's fstab.
> LABEL=sata400-17 / ext3 defaults 1 1
Here I made a big booboo.
I'd forgotten I'd had a separate /var .
The result was that the two shared a /var .
>> the fstab& Grub menu post-cloning steps would almost
surely result in
damage
>> to the F15 source, if not complete destruction, once is upgrade is begun.
> My first F15 install includes four partitions,
> one for / , one for /boot and two for swap.
> The copy uses the same /boot and swap partitions.
Oops. See above.
I suppose someone with a lot of experience could get away with
sharing a
/boot partition, but based upon your experiences of the past month I wouldn't
expect that group to include you. I would have merged the content of the
/boot partition into the /boot directory on the new /.
> Here are my grub stanzas:
> title Fedora 15 (2.6.42.12-1.fc15.i686.PAE) sdb17
> root (hd1,1)
> kernel /vmlinuz-2.6.42.12-1.fc15.i686.PAE ro root=/dev/sdb17
> initrd /initramfs-2.6.42.12-1.fc15.i686.PAE.img
Leaving off some of those cmdline options could be a problem. I'm not
familiar with all of them, and so would have included all I don't understand.
Me too. I just edited poorly.
> title Fedora 15 (2.6.42.12-1.fc15.i686.PAE) sdb3
> root (hd1,1)
> kernel /vmlinuz-2.6.42.12-1.fc15.i686.PAE ro
root=UUID=ce978949-d044-4020-8001-02454648a64e rd_NO_LUKS rd_NO_LVM rd_NO_MD
rd_NO_DM LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYTABLE=us rhgb quiet
> initrd /initramfs-2.6.42.12-1.fc15.i686.PAE.img
> I can boot sdb3.
> sdb17 fails.
Here is what I did:
Established that I could still boot my first F15, sdb3.
Added the options mistakenly removed from kernel line in menu.lst .
Made a filesystem in new partition, sdb18, for the new /var .
Booted F15 under sdb17, yaaaay.
Tried and failed to login.
Established that I could still login to my first F15.
Tried for a gnome session under the new F15, still no go.
Trying to login from a text console gave me permission errors, even for root.
Added enforcing=0 to the kernel line.
Booted and logged in to the new F15, yaaaay.
My next trick should be to fix the need for enforcing=0.
Should triggering an automatic relabeling do that?
Why does SELinux need fixing?
Do its rules reference sectors, absolute or relative?
> Any ideas?
1-Do it again with rsync or a program designed specifically for cloning.
2-Merge /boot partition into the new /.
3-Either find out if any of those cmdline options can play a part in the
problem and put back the ones you don't know you don't need, or add them all
back except rhgb & quiet
4-Add option enforcing=0 to cmdline to ensure SELinux isn't in the way
5-Use the Grub on hd1,1 to start both, but for the clone include a hd1,16
configfile stanza until you get it booted, then install Grub on sdb17 and
afterward chainload from hd1,1 to hd1,16, where updates will ultimately be
applied to grub.conf only for the clone on account of new kernels.
I'll definitely need to fix /boot .
The best case scenario is that it gets really cluttered
when I start updating kernels on multiple OSs.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
5:08 p.m.
On 2012/04/09 16:22 (GMT-0500) Michael Hennebry composed:
My next trick should be to fix the need for enforcing=0.
Should triggering an automatic relabeling do that?
Why does SELinux need fixing?
Do its rules reference sectors, absolute or relative?
I suggest a thread on a users list on this subject with a matching subject
line. I never boot with SELinux enabled on purpose, and know little about
SELinux except ways to avoid it.
--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
6:46 p.m.
On Mon, Apr 09, 2012 at 04:22:38PM -0500, Michael Hennebry wrote:
My next trick should be to fix the need for enforcing=0.
Should triggering an automatic relabeling do that?
One of reasons why it is needed.
Why does SELinux need fixing?
Do its rules reference sectors, absolute or relative?
No. SELinux looks at labels on files. Your copy targets got labels
according to their locations on a system where you were doing a copy
(or do not have such labels at all if SELinux was not active at that
time). Now you need labels relative to your new root before you can
operate with 'enforcing=1'.
In general if you are messing with files using anything else than
your intended target, or if you were running even for a short moment
with SELinux off, then to turn it on you have to fix labels.
I'll definitely need to fix /boot .
The best case scenario is that it gets really cluttered
when I start updating kernels on multiple OSs.
With multiple installations on the same machine chainloading "private"
bootloaders is likely the cleanest option even if grub2 will raise a
hissy fit about putting it on a partition. Shrug! An "old grub" still
can be used wherever possible.
Michal
8:47 p.m.
On Sun, 2012-04-08 at 17:08 -0500, Michael Hennebry wrote:
It's now been moe than a month sicne I started trying to install
F16.
Grrrrrrr.
To be honest, I've mostly tuned out this thread because you seem to have
become intent on doing it in a very complicated way, which some people
have been trying to help you with out of the kindness of their hearts.
I'm pretty sure there must be a less strange and rarely-used way of
doing it that would be usable in your situation...
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
9:14 p.m.
On Mon, 9 Apr 2012, Adam Williamson wrote:
On Sun, 2012-04-08 at 17:08 -0500, Michael Hennebry wrote:
> It's now been moe than a month sicne I started trying to install F16.
> Grrrrrrr.
To be honest, I've mostly tuned out this thread because you seem to have
become intent on doing it in a very complicated way, which some people
have been trying to help you with out of the kindness of their hearts.
Please don't take the grrrrrr as an indication
that I am mad at you or blame you for something.
Take it as an indication that seemingly interminal failure is frustrating.
I'm pretty sure there must be a less strange and rarely-used way
of
doing it that would be usable in your situation...
If you know what it is, I'm all eyes.
I started by burning a minimal F16 install CD.
It didn't work.
So far anything involving F16 or F17 has crashed my system.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
1:06 a.m.
On Mon, 2012-04-09 at 21:14 -0500, Michael Hennebry wrote:
On Mon, 9 Apr 2012, Adam Williamson wrote:
> On Sun, 2012-04-08 at 17:08 -0500, Michael Hennebry wrote:
>> It's now been moe than a month sicne I started trying to install F16.
>> Grrrrrrr.
>
> To be honest, I've mostly tuned out this thread because you seem to have
> become intent on doing it in a very complicated way, which some people
> have been trying to help you with out of the kindness of their hearts.
Please don't take the grrrrrr as an indication
that I am mad at you or blame you for something.
Take it as an indication that seemingly interminal failure is frustrating.
> I'm pretty sure there must be a less strange and rarely-used way of
> doing it that would be usable in your situation...
If you know what it is, I'm all eyes.
I started by burning a minimal F16 install CD.
It didn't work.
So far anything involving F16 or F17 has crashed my system.
If it's that bad, then I'm not at all sure F16 or F17 would work if you
somehow bypassed the regular installation method...
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
12:37 p.m.
On Mon, 9 Apr 2012, Adam Williamson wrote:
On Mon, 2012-04-09 at 21:14 -0500, Michael Hennebry wrote:
> On Mon, 9 Apr 2012, Adam Williamson wrote:
> > I'm pretty sure there must be a less strange and
rarely-used way of
> > doing it that would be usable in your situation...
>
> If you know what it is, I'm all eyes.
> I started by burning a minimal F16 install CD.
> It didn't work.
> So far anything involving F16 or F17 has crashed my system.
If it's that bad, then I'm not at all sure F16 or F17 would work if you
somehow bypassed the regular installation method...
I've been told that a kernel bug broke my motherboard for the F16 install.
An install mechanism that uses a recent kernel should to the trick.
Preupdate might be one, but I need a disposable place from which to start.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
12:51 p.m.
On Tue, 2012-04-10 at 12:37 -0500, Michael Hennebry wrote:
On Mon, 9 Apr 2012, Adam Williamson wrote:
> On Mon, 2012-04-09 at 21:14 -0500, Michael Hennebry wrote:
>> On Mon, 9 Apr 2012, Adam Williamson wrote:
>> > I'm pretty sure there must be a less strange and rarely-used way of
>> > doing it that would be usable in your situation...
>>
>> If you know what it is, I'm all eyes.
>> I started by burning a minimal F16 install CD.
>> It didn't work.
>> So far anything involving F16 or F17 has crashed my system.
>
> If it's that bad, then I'm not at all sure F16 or F17 would work if you
> somehow bypassed the regular installation method...
I've been told that a kernel bug broke my motherboard for the F16 install.
An install mechanism that uses a recent kernel should to the trick.
Preupdate might be one, but I need a disposable place from which to start.
Oh, right, I forgot that. I think possibly a yum upgrade might be the
easiest option, really. Someone linked to the yum upgrade instructions
page already, right?
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
6:04 p.m.
On 04/10/2012 01:37 PM, Michael Hennebry wrote:
On Mon, 9 Apr 2012, Adam Williamson wrote:
> On Mon, 2012-04-09 at 21:14 -0500, Michael Hennebry wrote:
>> On Mon, 9 Apr 2012, Adam Williamson wrote:
>> > I'm pretty sure there must be a less strange and rarely-used way of
>> > doing it that would be usable in your situation...
>>
>> If you know what it is, I'm all eyes.
>> I started by burning a minimal F16 install CD.
>> It didn't work.
>> So far anything involving F16 or F17 has crashed my system.
>
> If it's that bad, then I'm not at all sure F16 or F17 would work if you
> somehow bypassed the regular installation method...
I've been told that a kernel bug broke my motherboard for the F16 install.
An install mechanism that uses a recent kernel should to the trick.
Preupdate might be one, but I need a disposable place from which to start.
If this is the kernel bug you are referring to, perhaps there is a
workaround you could use to get your minimal CD running in order to do
an install.
https://bugzilla.redhat.com/show_bug.cgi?id=730007
That issue appears to be related to the following ACPI patch.
http://marc.info/?l=linux-acpi&m=133002974918284&w=2
The description of that patch suggests that passing any one of these
kernel parameters will prevent the problem.
acpi=off
processor.nocst=1
maxcpus=1
If one of those works, you should then be able to install F16, boot it
the same way, and upgrade the kernel to one which contains the patch.
-Mike
4:59 p.m.
On Tue, 10 Apr 2012, mwesten wrote:
If this is the kernel bug you are referring to, perhaps there is a
workaround you could use to get your minimal CD running in order to do
an install.
The description of that patch suggests that passing any one of these
kernel parameters will prevent the problem.
acpi=off
processor.nocst=1
maxcpus=1
I'm getting to anaconda now.
When I tried the upgrade,
the only system it would let me upgrade was the F14 system I wanted to keep.
For that reason, I told it to do a fresh install and a custom layout.
It crapped out after downloading or installing the last package.
After that,
I tried again using the install from iso without media method again.
A debug message mentioned SELinux,
so I tried again after adding selinux=0 to the kernel line.
Here is the final grub stanza:
title Install Fedora16 from iso 11 15:34
root (hd0,2)
kernel /isolinux16/vmlinuz repo=hd:LABEL=/home1:/iso16 acpi=off processor.nocst=1
maxcpus=1 selinux=0
initrd /isolinux16/initrd.img
Still no go.
I took some pictures of the "Details" message,
but the three are >60KB a piece after processing, >2MB before.
Should I send them to the list?
After clicking on exit, I saw four copies of the following
message at the bottom before I had to push reset.
Error: cannot open Packages database in /mnt/sysimage/var/lib/rpm .
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
7:52 p.m.
On Wed, 11 Apr 2012, Michael Hennebry wrote:
On Tue, 10 Apr 2012, mwesten wrote:
> If this is the kernel bug you are referring to, perhaps there is a
> workaround you could use to get your minimal CD running in order to do
> an install.
> The description of that patch suggests that passing any one of these
> kernel parameters will prevent the problem.
>
> acpi=off
> processor.nocst=1
> maxcpus=1
Here is the final grub stanza:
title Install Fedora16 from iso 11 15:34
root (hd0,2)
kernel /isolinux16/vmlinuz repo=hd:LABEL=/home1:/iso16 acpi=off
processor.nocst=1 maxcpus=1 selinux=0
initrd /isolinux16/initrd.img
I tried to install Fedora 17-alpha part duh in similar fashion,
but it didn't even boot.
dracut: FATAL No or empty root= argument
Attempted to kill init
I tried root=live, but got
don't know how to handle root=live.
'Twasn't a live CD, so that was not a surprise.
Cut and pasted the update= option suggested in the known issues section,
but that had no effect.
Pardon me.
I need to go kill something.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
9:33 p.m.
On 04/11/2012 08:52 PM, Michael Hennebry wrote:
On Wed, 11 Apr 2012, Michael Hennebry wrote:
> Here is the final grub stanza:
> title Install Fedora16 from iso 11 15:34
> root (hd0,2)
> kernel /isolinux16/vmlinuz repo=hd:LABEL=/home1:/iso16 acpi=off
> processor.nocst=1 maxcpus=1 selinux=0
> initrd /isolinux16/initrd.img
I tried to install Fedora 17-alpha part duh in similar fashion,
but it didn't even boot.
dracut: FATAL No or empty root= argument
Attempted to kill init
I tried root=live, but got
don't know how to handle root=live.
'Twasn't a live CD, so that was not a surprise.
Cut and pasted the update= option suggested in the known issues section,
but that had no effect.
Pardon me.
I need to go kill something.
I wouldn't recommend that...
As far as "installing without media", I believe your repo= is incorrect.
LABEL= takes a volume label, not a mountpoint. The blkid command will
tell you if there is one and what it is. You can set it with GPARTED.
So, if blkid says the label of the /home1 partition is "F16DVD", use
repo=hd:LABEL=F16DVD:/iso16
I would stick with F16 for now. Due to the changes in Anaconda for F17,
this gets a bit trickier to do...
-Mike
10:52 a.m.
On Wed, 11 Apr 2012, mwesten wrote:
On 04/11/2012 08:52 PM, Michael Hennebry wrote:
> On Wed, 11 Apr 2012, Michael Hennebry wrote:
>
>> Here is the final grub stanza:
>> title Install Fedora16 from iso 11 15:34
>> root (hd0,2)
>> kernel /isolinux16/vmlinuz repo=hd:LABEL=/home1:/iso16 acpi=off
>> processor.nocst=1 maxcpus=1 selinux=0
>> initrd /isolinux16/initrd.img
As far as "installing without media", I believe your repo=
is incorrect.
LABEL= takes a volume label, not a mountpoint. The blkid command will
tell you if there is one and what it is. You can set it with GPARTED.
So, if blkid says the label of the /home1 partition is "F16DVD", use
/home1 is the label.
It's usual mountpoint is /homes .
repo=hd:LABEL=F16DVD:/iso16
I would stick with F16 for now. Due to the changes in Anaconda for F17,
this gets a bit trickier to do...
I'll use whichever I can get to work.
So far, neither.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
3:44 p.m.
On 04/12/2012 11:52 AM, Michael Hennebry wrote:
On Wed, 11 Apr 2012, mwesten wrote:
> On 04/11/2012 08:52 PM, Michael Hennebry wrote:
>> On Wed, 11 Apr 2012, Michael Hennebry wrote:
>>
>>> Here is the final grub stanza:
>>> title Install Fedora16 from iso 11 15:34
>>> root (hd0,2)
>>> kernel /isolinux16/vmlinuz repo=hd:LABEL=/home1:/iso16 acpi=off
>>> processor.nocst=1 maxcpus=1 selinux=0
>>> initrd /isolinux16/initrd.img
> As far as "installing without media", I believe your repo= is incorrect.
>
> LABEL= takes a volume label, not a mountpoint. The blkid command will
> tell you if there is one and what it is. You can set it with GPARTED.
>
> So, if blkid says the label of the /home1 partition is "F16DVD", use
/home1 is the label.
It's usual mountpoint is /homes .
> repo=hd:LABEL=F16DVD:/iso16
>
> I would stick with F16 for now. Due to the changes in Anaconda for
> F17, this gets a bit trickier to do...
I'll use whichever I can get to work.
So far, neither.
I get that it's been frustrating for you, but this is a bleeding-edge
distribution which will have some issues now and then. Also consider
that you are trying to use a non-standard means of installation.
In any case, there still seems to be a problem with accessing the repo
files. I'm not sure what would cause the selinux message you saw.
I hope this /home1 is just a standard partition, without encryption or
LVM. Is that right? Whenever I do this, I just stick the iso in the
root of a FAT32 partition.
-Mike
5:36 p.m.
On Thu, 12 Apr 2012, mwesten wrote:
On 04/12/2012 11:52 AM, Michael Hennebry wrote:
> On Wed, 11 Apr 2012, mwesten wrote:
>
>> On 04/11/2012 08:52 PM, Michael Hennebry wrote:
>>> On Wed, 11 Apr 2012, Michael Hennebry wrote:
>>>
>>>> Here is the final grub stanza:
>>>> title Install Fedora16 from iso 11 15:34
>>>> root (hd0,2)
>>>> kernel /isolinux16/vmlinuz repo=hd:LABEL=/home1:/iso16 acpi=off
>>>> processor.nocst=1 maxcpus=1 selinux=0
>>>> initrd /isolinux16/initrd.img
>
>> As far as "installing without media", I believe your repo= is
incorrect.
>>
>> LABEL= takes a volume label, not a mountpoint. The blkid command will
>> tell you if there is one and what it is. You can set it with GPARTED.
>>
>> So, if blkid says the label of the /home1 partition is "F16DVD", use
>
> /home1 is the label.
> It's usual mountpoint is /homes .
> I'll use whichever I can get to work.
> So far, neither.
I get that it's been frustrating for you, but this is a
bleeding-edge
distribution which will have some issues now and then. Also consider
that you are trying to use a non-standard means of installation.
In any case, there still seems to be a problem with accessing the repo
files. I'm not sure what would cause the selinux message you saw.
I hope this /home1 is just a standard partition, without encryption or
LVM. Is that right? Whenever I do this, I just stick the iso in the
root of a FAT32 partition.
Correct. ext3.
The F16 installer installs packages, then dies.
The F17 installer just dies.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
6:12 p.m.
On Wed, 11 Apr 2012, Michael Hennebry wrote:
On Tue, 10 Apr 2012, mwesten wrote:
> The description of that patch suggests that passing any one of
these
> kernel parameters will prevent the problem.
>
> acpi=off
> processor.nocst=1
> maxcpus=1
I'm getting to anaconda now.
When I tried the upgrade,
the only system it would let me upgrade was the F14 system I wanted to keep.
For that reason, I told it to do a fresh install and a custom layout.
It crapped out after downloading or installing the last package.
After that,
I tried again using the install from iso without media method again.
A debug message mentioned SELinux,
so I tried again after adding selinux=0 to the kernel line.
Here is the final grub stanza:
title Install Fedora16 from iso 11 15:34
root (hd0,2)
kernel /isolinux16/vmlinuz repo=hd:LABEL=/home1:/iso16 acpi=off
processor.nocst=1 maxcpus=1 selinux=0
initrd /isolinux16/initrd.img
No selinux message, but
Still no go.
I took some pictures of the "Details" message,
but the three are >60KB a piece after processing, >2MB before.
Should I send them to the list?
I put them on http://web.cs.ndsu.nodak.edu/~hennebry/new16-[123].png
![http://web.cs.ndsu.nodak.edu/~hennebry/new16-[123].png](http://web.cs.ndsu.nodak.edu/~hennebry/new16-[123].png){kind=link}
After clicking on exit, I saw four copies of the following
message at the bottom before I had to push reset.
Error: cannot open Packages database in /mnt/sysimage/var/lib/rpm .
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
8:44 p.m.
On 04/12/2012 07:12 PM, Michael Hennebry wrote:
No selinux message, but
> Still no go.
> I took some pictures of the "Details" message,
> but the three are >60KB a piece after processing, >2MB before.
> Should I send them to the list?
I put them on http://web.cs.ndsu.nodak.edu/~hennebry/new16-[123].png
> After clicking on exit, I saw four copies of the following
> message at the bottom before I had to push reset.
> Error: cannot open Packages database in /mnt/sysimage/var/lib/rpm .
I see, the installer is finding what it needs alright, but just can't
complete. Isn't that what the netinstall did to you too?
I'm not getting anything at that URL, and I probably won't have a way
around whatever it is, either.
Of course, you could give F17 Beta a shot. At this point, however, I
would really only recommend using optical media or a LiveUSB for that.
-Mike
9:51 p.m.
On Thu, 12 Apr 2012, mwesten wrote:
On 04/12/2012 07:12 PM, Michael Hennebry wrote:
> No selinux message, but
>> Still no go.
>> I took some pictures of the "Details" message,
>> but the three are >60KB a piece after processing, >2MB before.
>> Should I send them to the list?
>
> I put them on http://web.cs.ndsu.nodak.edu/~hennebry/new16-[123].png
>
>> After clicking on exit, I saw four copies of the following
>> message at the bottom before I had to push reset.
>> Error: cannot open Packages database in /mnt/sysimage/var/lib/rpm .
>
I see, the installer is finding what it needs alright, but just can't
complete. Isn't that what the netinstall did to you too?
I'm not getting anything at that URL, and I probably won't have a way
around whatever it is, either.
Oops.
http://web.cs.ndsu.nodak.edu/~hennebry/new16-1.png
http://web.cs.ndsu.nodak.edu/~hennebry/new16-2.png
http://web.cs.ndsu.nodak.edu/~hennebry/new16-3.png
Of course, you could give F17 Beta a shot. At this point, however, I
would really only recommend using optical media or a LiveUSB for that.
I'll download it.
I've got one more rather hairy thing to try before trying to use it.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
1:47 p.m.
On 04/12/2012 10:51 PM, Michael Hennebry wrote:
On Thu, 12 Apr 2012, mwesten wrote:
> On 04/12/2012 07:12 PM, Michael Hennebry wrote:
>> No selinux message, but
>>> Still no go.
>>> I took some pictures of the "Details" message,
>>> but the three are >60KB a piece after processing, >2MB before.
>>> Should I send them to the list?
>>
>> I put them on http://web.cs.ndsu.nodak.edu/~hennebry/new16-[123].png
>>
>>> After clicking on exit, I saw four copies of the following
>>> message at the bottom before I had to push reset.
>>> Error: cannot open Packages database in /mnt/sysimage/var/lib/rpm .
>>
>
> I see, the installer is finding what it needs alright, but just can't
> complete. Isn't that what the netinstall did to you too?
>
> I'm not getting anything at that URL, and I probably won't have a way
> around whatever it is, either.
Oops.
http://web.cs.ndsu.nodak.edu/~hennebry/new16-1.png
http://web.cs.ndsu.nodak.edu/~hennebry/new16-2.png
http://web.cs.ndsu.nodak.edu/~hennebry/new16-3.png
> Of course, you could give F17 Beta a shot. At this point, however, I
> would really only recommend using optical media or a LiveUSB for that.
I'll download it.
I've got one more rather hairy thing to try before trying to use it.
Alright, I looked at the pics you posted and it appears that you have
reproduced this anaconda crash during the bootloader installation:
https://bugzilla.redhat.com/show_bug.cgi?id=796472
I was looking at the included exception report and it seems this person
was installing F16 on the second of two drives with the bootloader going
on the first.
Is that what you were doing as well?
-Mike
3:36 p.m.
On Fri, 13 Apr 2012, mwesten wrote:
On 04/12/2012 10:51 PM, Michael Hennebry wrote:
> On Thu, 12 Apr 2012, mwesten wrote:
>
>> On 04/12/2012 07:12 PM, Michael Hennebry wrote:
>>> No selinux message, but
>>>> Still no go.
>>>> I took some pictures of the "Details" message,
>> I'm not getting anything at that URL, and I probably
won't have a way
>> around whatever it is, either.
>
> Oops.
> http://web.cs.ndsu.nodak.edu/~hennebry/new16-1.png
> http://web.cs.ndsu.nodak.edu/~hennebry/new16-2.png
> http://web.cs.ndsu.nodak.edu/~hennebry/new16-3.png
Alright, I looked at the pics you posted and it appears that you have
reproduced this anaconda crash during the bootloader installation:
https://bugzilla.redhat.com/show_bug.cgi?id=796472
I was looking at the included exception report and it seems this person
was installing F16 on the second of two drives with the bootloader going
on the first.
Is that what you were doing as well?
That is not what I recall, but I might be wrong.
Is there a way to tell which disk was used
for booting that does not involve rebooting?
I've been poking around /proc , but haven't found anything.
As mentioned earlier, when attempting upgrade,
the only allowed source was F14, i.e. the one I wanted to keep.
For that reason, I selected fresh custom install.
When listing available partitions, anaconda listed the sda partitions as
the target partitions and the sdb partitions as the data partitions.
I put everything except a swap file on sdb.
When it asked where to put the bootloader,
I think it specified sda as the default.
My recollection was that sdb was the boot disk, so that bothered me.
Given the troubles I'd had previously after rejecting a bootloader default,
I decided to accept this one.
All seemed to be going well until the crash.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
11:57 p.m.
Why can grub find one file and not the other?
[root@localhost homes]# pwd
/homes
[root@localhost homes]# df / /homes
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/sdb10 82582288 32223896 46163628 42% /
/dev/sdb1 38456308 35386332 1116476 97% /homes
[root@localhost homes]# ls vmlinuz-3.1.9-1.4-desktop
vmlinuz-2.6.35.14-106.fc14.i686.PAE
vmlinuz-2.6.35.14-106.fc14.i686.PAE vmlinuz-3.1.9-1.4-desktop
[root@localhost homes]# grub
Probing devices to guess BIOS drives. This may take a long time.
GNU GRUB version 0.97 (640K lower / 3072K upper memory)
[ Minimal BASH-like line editing is supported. For the first word, TAB
lists possible command completions. Anywhere else TAB lists the possible
completions of a device/filename.]
grub> find /vmlinuz-2.6.35.14-106.fc14.i686.PAE
find /vmlinuz-2.6.35.14-106.fc14.i686.PAE
Error 15: File not found
grub> find /vmlinuz-3.1.9-1.4-desktop
find /vmlinuz-3.1.9-1.4-desktop
(hd1,0)
(hd1,1)
grub>
[root@localhost homes]# mount | grep sdb
/dev/sdb10 on / type ext4 (rw)
/dev/sdb11 on /var type ext3 (rw)
/dev/sdb1 on /homes type ext3 (rw)
/homes is the mountpoint for sdb1==(hd1,0)
What gives?
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
3:29 a.m.
On Fri, 13 Apr 2012, mwesten wrote:
Alright, I looked at the pics you posted and it appears that you have
reproduced this anaconda crash during the bootloader installation:
https://bugzilla.redhat.com/show_bug.cgi?id=796472
I was looking at the included exception report and it seems this person
was installing F16 on the second of two drives with the bootloader going
on the first.
Is that what you were doing as well?
Yes.
So I tried again.
I use grub's setup to put the bootloader on sdb and point it at sdb1.
I updated BIOS settings to use sdb.
Again anaconda labeled sdb partitons storage and sda partitions install.
It had sda marked for the bootloader.
It offered me the choice of sda or sdb19 for the bootloader.
I picked sdb19.
It choked the same way again.
The next time, I made /boot sda2 and picked sda for the bootloader.
This time I got an install, but I'm not sure how to boot to it.
I'd rather not have to play with the BIOS
every time I want to boot a different install.
Is there a way to chainload to sda?
I'd have tried configfile,
but I don't think that works from old grub to grub2.
The grub2 directory has nearly 200 files in it.
Wow.
Also, I truly hate UUIDs.
For desktop systems, they are large, unintelligible overkill.
Labels are quite sufficient.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
8:41 a.m.
On 04/14/2012 04:29 AM, Michael Hennebry wrote:
So I tried again.
I use grub's setup to put the bootloader on sdb and point it at sdb1.
I updated BIOS settings to use sdb.
Again anaconda labeled sdb partitons storage and sda partitions install.
It had sda marked for the bootloader.
It offered me the choice of sda or sdb19 for the bootloader.
I picked sdb19.
It choked the same way again.
The next time, I made /boot sda2 and picked sda for the bootloader.
This time I got an install, but I'm not sure how to boot to it.
I'd rather not have to play with the BIOS
every time I want to boot a different install.
Is there a way to chainload to sda?
I'd have tried configfile,
but I don't think that works from old grub to grub2.
This is not quite clear to me. It sounds like you can now boot F16 or
whatever you had before by changing the BIOS boot drive. Is that where
we are?
GRUB Legacy can chainload the bootsector of the second drive with the
chainloader command.
chainloader (hd1)+1
-Mike
12:51 p.m.
On Sat, 14 Apr 2012, mwesten wrote:
On 04/14/2012 04:29 AM, Michael Hennebry wrote:
> The next time, I made /boot sda2 and picked sda for the
bootloader.
> This time I got an install, but I'm not sure how to boot to it.
> I'd rather not have to play with the BIOS
> every time I want to boot a different install.
> Is there a way to chainload to sda?
> I'd have tried configfile,
> but I don't think that works from old grub to grub2.
This is not quite clear to me. It sounds like you can now boot F16 or
whatever you had before by changing the BIOS boot drive. Is that where
we are?
GRUB Legacy can chainload the bootsector of the second drive with the
chainloader command.
chainloader (hd1)+1
Thanks.
Chainloading worked, Grub 2 didn't.
GRUB Loading stage1.5
GRUB loading please wait...
Error 18
Huh?
I thought that Grub 2 was supposed to be smarter than old decrepit Grub.
1023 cylinders or not, sda2 was my /boot from the first time I installed
fedoralinux to some considerable time after I put in the second drive.
Old decrepit Grub handled it.
Why not wave of the future Grub?
Also note that sda2 does not cross the 1023 cylinder boundary.
root@localhost ~]# fdisk /dev/sda
Command (m for help): p
Disk /dev/sda: 40.0 GB, 40020664320 bytes
255 heads, 63 sectors/track, 4865 cylinders, total 78165360 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0xece6ece6
Device Boot Start End Blocks Id System
/dev/sda1 * 63 28676024 14337981 c W95 FAT32 (LBA)
/dev/sda2 28676025 28871576 97776 83 Linux
/dev/sda3 28871577 45668888 8398656 83 Linux
/dev/sda4 45668889 78165359 16248235+ 5 Extended
/dev/sda5 45668952 49574888 1952968+ 82 Linux swap / Solaris
/dev/sda6 49574952 62856296 6640672+ 83 Linux
/dev/sda7 62856360 78165359 7654500 83 Linux
1023*78165360/4865=16436415< 28676025
I'm pretty sure it is Grub 2.
Not only do I have a brand new grub2 directory with nearly 200 files in it,
[root@localhost ~]# file - < /dev/sda
/dev/stdin: x86 boot sector; GRand Unified Bootloader, stage1 version 0x3, boot drive
0x80, stage2 address 0x2000, stage2 segment 0x200; partition 1: ID=0xc, active, starthead
1, startsector 63, 28675962 sectors; partition 2: ID=0x83, starthead 7, startsector
28676025, 195552 sectors; partition 3: ID=0x83, starthead 7, startsector 28871577,
16797312 sectors; partition 4: ID=0x5, starthead 7, startsector 45668889, 32496471
sectors, code offset 0x63
[root@localhost ~]# file - < /dev/sdb
/dev/stdin: x86 boot sector; GRand Unified Bootloader, stage1 version 0x3, stage2 address
0x2000, stage2 segment 0x200, GRUB version 0.97; partition 1: ID=0x83, starthead 1,
startsector 63, 78140097 sectors; partition 2: ID=0x83, active, starthead 254, startsector
78140160, 192780 sectors; partition 3: ID=0x83, starthead 254, startsector 78332940,
31262490 sectors; partition 4: ID=0x5, starthead 254, startsector 109595430, 671822235
sectors, code offset 0x48
[root@localhost ~]#
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
5:48 a.m.
On Sat, 2012-04-14 at 12:51 -0500, Michael Hennebry wrote:
On Sat, 14 Apr 2012, mwesten wrote:
> On 04/14/2012 04:29 AM, Michael Hennebry wrote:
>> The next time, I made /boot sda2 and picked sda for the bootloader.
>> This time I got an install, but I'm not sure how to boot to it.
>> I'd rather not have to play with the BIOS
>> every time I want to boot a different install.
>> Is there a way to chainload to sda?
>> I'd have tried configfile,
>> but I don't think that works from old grub to grub2.
>
> This is not quite clear to me. It sounds like you can now boot F16 or
> whatever you had before by changing the BIOS boot drive. Is that where
> we are?
>
> GRUB Legacy can chainload the bootsector of the second drive with the
> chainloader command.
>
> chainloader (hd1)+1
Thanks.
Chainloading worked, Grub 2 didn't.
GRUB Loading stage1.5
GRUB loading please wait...
Error 18
See https://bugzilla.redhat.com/show_bug.cgi?id=804835 .
As we understand it at present, you can't 'chainload' grub2 that way.
The correct way to 'chainload' grub2 is to load core.img as a 'kernel',
as described in https://bugzilla.redhat.com/show_bug.cgi?id=804835#c2 .
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
5:39 p.m.
I've booted F16.
On Tue, 17 Apr 2012, Adam Williamson wrote:
See https://bugzilla.redhat.com/show_bug.cgi?id=804835 .
As we understand it at present, you can't 'chainload' grub2 that way.
The correct way to 'chainload' grub2 is to load core.img as a 'kernel',
as described in https://bugzilla.redhat.com/show_bug.cgi?id=804835#c2 .
Thanks.
My old-grub stanza:
title chainload to sda
kernel (hd1,1)/grub2/core.img
As I didn't expect it to work, I only installed the minimal version.
There are a few things that concern me.
When I tried to login as hennebry,
my password showed on the console.
Also, I could not login as hennebry.
From catting passwd, there are no ordinary users at all.
Whether
I was allowed to specify an ordinary user, I do not recall.
None is listed in the generated kickstart file.
When using cntrl-alt-Fx to switch consoles,
I could not get back to my original.
When I got around to trying root, I was allowed to login.
Also, which of the following should concern me:
[root@localhost root]# grep -v Installing install.log
warning: setup-2.8.36-2.fc16.noarch: Header V3 RSA/SHA256 Signature, key ID a82ba4b7:
NOKEY
Need 'inst' function, try setting PLYMOUTH_POPULATE_SOURCE_FUNCTIONS to a file
that defines it
SELinux: Could not downgrade policy file /etc/selinux/targeted/policy/policy.26,
searching for an older version.
SELinux: Could not open policy file <= /etc/selinux/targeted/policy/policy.26: No
such file or directory
load_policy: Can't load policy: No such file or directory
grubby fatal error: unable to find a suitable template
grubby: doing this would leave no kernel entries. Not writing out new config.
*** FINISHED INSTALLING PACKAGES ***
[root@localhost root]#
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
10:57 a.m.
On Tue, 2012-04-17 at 17:39 -0500, Michael Hennebry wrote:
I've booted F16.
On Tue, 17 Apr 2012, Adam Williamson wrote:
> See https://bugzilla.redhat.com/show_bug.cgi?id=804835 .
>
> As we understand it at present, you can't 'chainload' grub2 that way.
> The correct way to 'chainload' grub2 is to load core.img as a
'kernel',
> as described in https://bugzilla.redhat.com/show_bug.cgi?id=804835#c2 .
Thanks.
My old-grub stanza:
title chainload to sda
kernel (hd1,1)/grub2/core.img
As I didn't expect it to work, I only installed the minimal version.
There are a few things that concern me.
When I tried to login as hennebry,
my password showed on the console.
Probably a plymouth issue.
Also, I could not login as hennebry.
From catting passwd, there are no ordinary users at all.
Whether I was allowed to specify an ordinary user, I do not recall.
You wouldn't, with a minimal install. User creation is done by firstboot
in a graphical install. In a minimal install, there is no firstboot any
more (since F16 IIRC); when there was, the tool it ran didn't handle
user creation. So when you do a minimal install, you have to log in as
root and do user creation yourself.
None is listed in the generated kickstart file.
When using cntrl-alt-Fx to switch consoles,
I could not get back to my original.
When I got around to trying root, I was allowed to login.
Also, which of the following should concern me:
[root@localhost root]# grep -v Installing install.log
warning: setup-2.8.36-2.fc16.noarch: Header V3 RSA/SHA256 Signature, key ID a82ba4b7:
NOKEY
Need 'inst' function, try setting PLYMOUTH_POPULATE_SOURCE_FUNCTIONS to a file
that defines it
SELinux: Could not downgrade policy file /etc/selinux/targeted/policy/policy.26,
searching for an older version.
SELinux: Could not open policy file <= /etc/selinux/targeted/policy/policy.26: No
such file or directory
load_policy: Can't load policy: No such file or directory
grubby fatal error: unable to find a suitable template
grubby: doing this would leave no kernel entries. Not writing out new config.
*** FINISHED INSTALLING PACKAGES ***
[root@localhost root]#
I wouldn't be hugely worried about any of them. Try doing a full install
and I suspect a lot of the above 'weirdness' would go away.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
4:39 p.m.
On Wed, 18 Apr 2012, Adam Williamson wrote:
On Tue, 2012-04-17 at 17:39 -0500, Michael Hennebry wrote:
> I've booted F16.
> My old-grub stanza:
> title chainload to sda
> kernel (hd1,1)/grub2/core.img
>
> As I didn't expect it to work, I only installed the minimal version.
>
> There are a few things that concern me.
> When I tried to login as hennebry,
> my password showed on the console.
Probably a plymouth issue.
I'd found these before your response:
http://en.wikipedia.org/wiki/Plymouth_%28software%29
http://en.wikipedia.org/wiki/Talk:Plymouth_(software)
Someone really doesn't like plymouth.
> Also, which of the following should concern me:
> [root@localhost root]# grep -v Installing install.log
> warning: setup-2.8.36-2.fc16.noarch: Header V3 RSA/SHA256 Signature, key ID a82ba4b7:
NOKEY
> Need 'inst' function, try setting PLYMOUTH_POPULATE_SOURCE_FUNCTIONS to a
file that defines it
> SELinux: Could not downgrade policy file /etc/selinux/targeted/policy/policy.26,
searching for an older version.
> SELinux: Could not open policy file <= /etc/selinux/targeted/policy/policy.26:
No such file or directory
> load_policy: Can't load policy: No such file or directory
> grubby fatal error: unable to find a suitable template
> grubby: doing this would leave no kernel entries. Not writing out new config.
> *** FINISHED INSTALLING PACKAGES ***
> [root@localhost root]#
I wouldn't be hugely worried about any of them. Try doing a full install
and I suspect a lot of the above 'weirdness' would go away.
Will do.
Thanks.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
6:10 p.m.
On Wed, 2012-04-18 at 16:39 -0500, Michael Hennebry wrote:
On Wed, 18 Apr 2012, Adam Williamson wrote:
> On Tue, 2012-04-17 at 17:39 -0500, Michael Hennebry wrote:
>> I've booted F16.
>> My old-grub stanza:
>> title chainload to sda
>> kernel (hd1,1)/grub2/core.img
>>
>> As I didn't expect it to work, I only installed the minimal version.
>>
>> There are a few things that concern me.
>> When I tried to login as hennebry,
>> my password showed on the console.
>
> Probably a plymouth issue.
I should be more precise, and say it's probably a _lack_ of plymouth
issue...
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
1:09 p.m.
On Wed, 18 Apr 2012, Michael Hennebry wrote:
On Wed, 18 Apr 2012, Adam Williamson wrote:
> I wouldn't be hugely worried about any of them. Try doing a
full install
> and I suspect a lot of the above 'weirdness' would go away.
Will do.
Done.
The problems that I am aware of so far:
I seem to be stuck with KDE though I specified at least two other desktops.
The login screen did not have an obvious option for another desktop.
man -k desktop showed me xfdesktop,
but I'm not at all sure that running it from within KDE is the thing to do.
I have a terminal window set for 80 characters across and an 11pt font.
When I sshed into another machine, I got a bigger font and fewer characters.
The preferences are still set for 11pt monospace.
Does this mean that the remote site sent some control characters?
Which of these should I worry about?
[root@localhost ~]# grep -v stall install.log
warning: fontpackages-filesystem-1.44-2.fc15.noarch: Header V3 RSA/SHA256 Signature, key
ID a82ba4b7: NOKEY
/var/tmp/rpm-tmp.70BdJf: line 1: /usr/sbin/update-alternatives: No such file or directory
warning: %post(tomcat-servlet-3.0-api-0:7.0.26-1.fc16.noarch) scriptlet failed, exit
status 127
/var/tmp/rpm-tmp.pKiL8w: line 1: /usr/sbin/update-alternatives: No such file or directory
warning: %post(tomcat-jsp-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed, exit status
127
/var/tmp/rpm-tmp.NaLh0X: line 3: find: command not found
/var/tmp/rpm-tmp.hawg7Q: line 1: /usr/sbin/update-alternatives: No such file or directory
warning: %post(tomcat-el-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed, exit status 127
Running in chroot, ignoring request.
SELinux: Could not downgrade policy file /etc/selinux/targeted/policy/policy.26,
searching for an older version.
SELinux: Could not open policy file <= /etc/selinux/targeted/policy/policy.26: No
such file or directory
load_policy: Can't load policy: No such file or directory
6874 blocks
*** FINISHED INSTALLING PACKAGES ***
The teminal situation is weirder than I had originally thought.
I can produce more terminal windows, but none like I started with.
Right-clicking on the terminal icon (grrrr) list the
first terminal window and new window as options.
Clicking on new window produces a window with
a different toolbar and a different color scheme.
Probably this is all KDE funkiness and not an install issue.
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
3:25 p.m.
On Thu, 2012-04-19 at 13:09 -0500, Michael Hennebry wrote:
On Wed, 18 Apr 2012, Michael Hennebry wrote:
> On Wed, 18 Apr 2012, Adam Williamson wrote:
>> I wouldn't be hugely worried about any of them. Try doing a full install
>> and I suspect a lot of the above 'weirdness' would go away.
>
> Will do.
Done.
The problems that I am aware of so far:
I seem to be stuck with KDE though I specified at least two other desktops.
The login screen did not have an obvious option for another desktop.
Do you know if you have KDM or GDM? If GDM, you get to pick a desktop
*after* you pick a username - there's a drop-down selector below the
password entry box.
I have a terminal window set for 80 characters across and an 11pt
font.
When I sshed into another machine, I got a bigger font and fewer characters.
The preferences are still set for 11pt monospace.
Does this mean that the remote site sent some control characters?
Don't know anything about this.
Which of these should I worry about?
[root@localhost ~]# grep -v stall install.log
warning: fontpackages-filesystem-1.44-2.fc15.noarch: Header V3 RSA/SHA256 Signature, key
ID a82ba4b7: NOKEY
/var/tmp/rpm-tmp.70BdJf: line 1: /usr/sbin/update-alternatives: No such file or directory
warning: %post(tomcat-servlet-3.0-api-0:7.0.26-1.fc16.noarch) scriptlet failed, exit
status 127
/var/tmp/rpm-tmp.pKiL8w: line 1: /usr/sbin/update-alternatives: No such file or directory
warning: %post(tomcat-jsp-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed, exit status
127
/var/tmp/rpm-tmp.NaLh0X: line 3: find: command not found
/var/tmp/rpm-tmp.hawg7Q: line 1: /usr/sbin/update-alternatives: No such file or directory
warning: %post(tomcat-el-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed, exit status
127
Not quite sure what's going on there; alternatives is a mechanism for
allowing multiple providers for a given file or set of files, basically.
update-alternatives is the tool to specify which one should be used, and
some packages use it in scripts, obviously. I'm not sure why it wouldn't
be present when they're trying to run it; badly stated deps in the
packages or some kind of issue with usrmove maybe. I _doubt_ the
consequences would be critical, but it's hard to be sure.
Running in chroot, ignoring request.
SELinux: Could not downgrade policy file /etc/selinux/targeted/policy/policy.26,
searching for an older version.
SELinux: Could not open policy file <= /etc/selinux/targeted/policy/policy.26: No
such file or directory
load_policy: Can't load policy: No such file or directory
What does 'getenforce' say now?
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
4:17 p.m.
On Thu, 19 Apr 2012, Adam Williamson wrote:
On Thu, 2012-04-19 at 13:09 -0500, Michael Hennebry wrote:
> On Wed, 18 Apr 2012, Michael Hennebry wrote:
> The problems that I am aware of so far:
>
> I seem to be stuck with KDE though I specified at least two other desktops.
> The login screen did not have an obvious option for another desktop.
Do you know if you have KDM or GDM? If GDM, you get to pick a desktop
*after* you pick a username - there's a drop-down selector below the
password entry box.
gdm.
Thanks.
> Which of these should I worry about?
> [root@localhost ~]# grep -v stall install.log
> warning: fontpackages-filesystem-1.44-2.fc15.noarch: Header V3 RSA/SHA256 Signature,
key ID a82ba4b7: NOKEY
> /var/tmp/rpm-tmp.70BdJf: line 1: /usr/sbin/update-alternatives: No such file or
directory
> warning: %post(tomcat-servlet-3.0-api-0:7.0.26-1.fc16.noarch) scriptlet failed, exit
status 127
> /var/tmp/rpm-tmp.pKiL8w: line 1: /usr/sbin/update-alternatives: No such file or
directory
> warning: %post(tomcat-jsp-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed, exit
status 127
> /var/tmp/rpm-tmp.NaLh0X: line 3: find: command not found
> /var/tmp/rpm-tmp.hawg7Q: line 1: /usr/sbin/update-alternatives: No such file or
directory
> warning: %post(tomcat-el-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed, exit
status 127
Not quite sure what's going on there; alternatives is a mechanism for
allowing multiple providers for a given file or set of files, basically.
update-alternatives is the tool to specify which one should be used, and
some packages use it in scripts, obviously. I'm not sure why it wouldn't
be present when they're trying to run it; badly stated deps in the
packages or some kind of issue with usrmove maybe. I _doubt_ the
consequences would be critical, but it's hard to be sure.
IIRC alternatives works by manipulating symlinks.
The one time I tried to use it was for java.
Didn't like it very much, but don't remember details.
> Running in chroot, ignoring request.
> SELinux: Could not downgrade policy file /etc/selinux/targeted/policy/policy.26,
searching for an older version.
> SELinux: Could not open policy file <= /etc/selinux/targeted/policy/policy.26:
No such file or directory
> load_policy: Can't load policy: No such file or directory
What does 'getenforce' say now?
Disabled
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
9:55 a.m.
On Wed, 18 Apr 2012, Michael Hennebry wrote:
On Wed, 18 Apr 2012, Adam Williamson wrote:
> I wouldn't be hugely worried about any of them. Try doing a full install
> and I suspect a lot of the above 'weirdness' would go away.
Will do.
Done.
Which of these should I worry about?
[root@localhost ~]# grep -v stall install.log
warning: fontpackages-filesystem-1.44-2.fc15.noarch: Header V3 RSA/SHA256
Signature, key ID a82ba4b7: NOKEY
/var/tmp/rpm-tmp.70BdJf: line 1: /usr/sbin/update-alternatives: No such file
or directory
warning: %post(tomcat-servlet-3.0-api-0:7.0.26-1.fc16.noarch) scriptlet
failed, exit status 127
/var/tmp/rpm-tmp.pKiL8w: line 1: /usr/sbin/update-alternatives: No such file
or directory
warning: %post(tomcat-jsp-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed,
exit status 127
/var/tmp/rpm-tmp.NaLh0X: line 3: find: command not found
/var/tmp/rpm-tmp.hawg7Q: line 1: /usr/sbin/update-alternatives: No such file
or directory
warning: %post(tomcat-el-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed,
exit status 127
Running in chroot, ignoring request.
SELinux: Could not downgrade policy file
/etc/selinux/targeted/policy/policy.26, searching for an older version.
SELinux: Could not open policy file <=
/etc/selinux/targeted/policy/policy.26: No such file or directory
load_policy: Can't load policy: No such file or directory
6874 blocks
*** FINISHED INSTALLING PACKAGES ***
getenforce says Disabled, I expect because of the above Could nots.
Is this something that wold be fixed with an auto-relabel?
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
4:35 a.m.
On Tue, 2012-04-24 at 09:55 -0500, Michael Hennebry wrote:
On Wed, 18 Apr 2012, Michael Hennebry wrote:
> On Wed, 18 Apr 2012, Adam Williamson wrote:
>
>> I wouldn't be hugely worried about any of them. Try doing a full install
>> and I suspect a lot of the above 'weirdness' would go away.
>
> Will do.
Done.
Which of these should I worry about?
[root@localhost ~]# grep -v stall install.log
warning: fontpackages-filesystem-1.44-2.fc15.noarch: Header V3 RSA/SHA256
Signature, key ID a82ba4b7: NOKEY
/var/tmp/rpm-tmp.70BdJf: line 1: /usr/sbin/update-alternatives: No such file
or directory
warning: %post(tomcat-servlet-3.0-api-0:7.0.26-1.fc16.noarch) scriptlet
failed, exit status 127
/var/tmp/rpm-tmp.pKiL8w: line 1: /usr/sbin/update-alternatives: No such file
or directory
warning: %post(tomcat-jsp-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed,
exit status 127
/var/tmp/rpm-tmp.NaLh0X: line 3: find: command not found
/var/tmp/rpm-tmp.hawg7Q: line 1: /usr/sbin/update-alternatives: No such file
or directory
warning: %post(tomcat-el-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed,
exit status 127
Running in chroot, ignoring request.
SELinux: Could not downgrade policy file
/etc/selinux/targeted/policy/policy.26, searching for an older version.
SELinux: Could not open policy file <=
/etc/selinux/targeted/policy/policy.26: No such file or directory
load_policy: Can't load policy: No such file or directory
6874 blocks
*** FINISHED INSTALLING PACKAGES ***
getenforce says Disabled, I expect because of the above Could nots.
Is this something that wold be fixed with an auto-relabel?
Are you somehow missing the selinux-policy-targeted package? That should
have the file in question.
The RPM post errors could be reported as bugs in those packages, though
they may be 'false errors', I guess. It _looks_ like those packages
aren't properly requiring tools they execute in their %post scripts.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
11:46 p.m.
On Wed, 25 Apr 2012, Adam Williamson wrote:
On Tue, 2012-04-24 at 09:55 -0500, Michael Hennebry wrote:
> On Wed, 18 Apr 2012, Michael Hennebry wrote:
> Which of these should I worry about?
> [root@localhost ~]# grep -v stall install.log
> warning: fontpackages-filesystem-1.44-2.fc15.noarch: Header V3 RSA/SHA256
> Signature, key ID a82ba4b7: NOKEY
> /var/tmp/rpm-tmp.70BdJf: line 1: /usr/sbin/update-alternatives: No such file
> or directory
> warning: %post(tomcat-servlet-3.0-api-0:7.0.26-1.fc16.noarch) scriptlet
> failed, exit status 127
> /var/tmp/rpm-tmp.pKiL8w: line 1: /usr/sbin/update-alternatives: No such file
> or directory
> warning: %post(tomcat-jsp-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed,
> exit status 127
> /var/tmp/rpm-tmp.NaLh0X: line 3: find: command not found
> /var/tmp/rpm-tmp.hawg7Q: line 1: /usr/sbin/update-alternatives: No such file
> or directory
> warning: %post(tomcat-el-2.2-api-0:7.0.26-1.fc16.noarch) scriptlet failed,
> exit status 127
> Running in chroot, ignoring request.
> SELinux: Could not downgrade policy file
> /etc/selinux/targeted/policy/policy.26, searching for an older version.
> SELinux: Could not open policy file <=
> /etc/selinux/targeted/policy/policy.26: No such file or directory
> load_policy: Can't load policy: No such file or directory
> 6874 blocks
> *** FINISHED INSTALLING PACKAGES ***
>
>
> getenforce says Disabled, I expect because of the above Could nots.
> Is this something that wold be fixed with an auto-relabel?
Are you somehow missing the selinux-policy-targeted package? That should
have the file in question.
I think this says that I have it, but not the most recent:
[root@localhost ~]# yum info selinux-policy-targeted
Loaded plugins: langpacks, presto, refresh-packagekit
fedora/metalink | 21 kB 00:00
updates/metalink | 7.4 kB 00:00
updates | 4.5 kB 00:00
updates/primary_db | 4.7 MB 00:07
updates/group_gz | 435 kB 00:00
Installed Packages
Name : selinux-policy-targeted
Arch : noarch
Version : 3.10.0
Release : 80.fc16
Size : 12 M
Repo : installed
From repo : updates
Summary : SELinux targeted base
policy
URL : http://oss.tresys.com/repos/refpolicy/
License : GPLv2+
Description : SELinux Reference policy targeted base module.
Available Packages
Name : selinux-policy-targeted
Arch : noarch
Version : 3.10.0
Release : 84.fc16
Size : 3.5 M
Repo : updates
Summary : SELinux targeted base policy
URL : http://oss.tresys.com/repos/refpolicy/
License : GPLv2+
Description : SELinux Reference policy targeted base module.
Also:
[root@localhost ~]# locate update-alternatives
/usr/sbin/update-alternatives
/usr/share/man/man8/update-alternatives.8.gz
[root@localhost ~]# locate tomcat-el-2.2-api-0:7.0.26-1.fc16.noarch
[root@localhost ~]# which find
/bin/find
[root@localhost ~]# locate policy.26
/etc/selinux/targeted/policy/policy.26
[root@localhost ~]# ls -l /bin/find /etc/selinux/targeted/policy/policy.26
-rwxr-xr-x 1 root root 189736 May 11 2011 /bin/find
-rw-r--r-- 1 root root 4238308 Mar 13 06:21 /etc/selinux/targeted/policy/policy.26
[root@localhost ~]#
The RPM post errors could be reported as bugs in those packages,
though
they may be 'false errors', I guess. It _looks_ like those packages
aren't properly requiring tools they execute in their %post scripts.
What should I do about the messages?
Would yum reinstalling some packages, e.g. selinux-policy-targeted, help?
What are the consequences of just ignoring them?
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
3:26 p.m.
On Sat, 2012-04-28 at 23:46 -0500, Michael Hennebry wrote:
What should I do about the messages?
Would yum reinstalling some packages, e.g. selinux-policy-targeted, help?
What are the consequences of just ignoring them?
At this point I'd say file a bug against selinux-policy for that issue,
at least. I'm not sure what's going on there.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
3:44 p.m.
On Wed, 2 May 2012, Adam Williamson wrote:
On Sat, 2012-04-28 at 23:46 -0500, Michael Hennebry wrote:
> What should I do about the messages?
> Would yum reinstalling some packages, e.g. selinux-policy-targeted, help?
> What are the consequences of just ignoring them?
At this point I'd say file a bug against selinux-policy for that issue,
at least. I'm not sure what's going on there.
'Tis done: https://bugzilla.redhat.com/show_bug.cgi?id=818748
I expect that it will be a while before the bug is either fixed or abandoned.
What to do in the mean time?
Should I use it as-is?
Will something blow up if I try to enable selinux?
Should I stick with F14?
Should I yum reinstall packages?
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
4:16 p.m.
I am not sure what caused this because i have fedora 17 beta installed
on my old p4 pc and its working fine earlier i had fedora 16 installed
i didn't faced any issue with selinux
On Fri, May 4, 2012 at 2:14 AM, Michael Hennebry
<hennebry(a)web.cs.ndsu.nodak.edu> wrote:
On Wed, 2 May 2012, Adam Williamson wrote:
> On Sat, 2012-04-28 at 23:46 -0500, Michael Hennebry wrote:
>
>> What should I do about the messages?
>> Would yum reinstalling some packages, e.g. selinux-policy-targeted, help?
>> What are the consequences of just ignoring them?
>
>
> At this point I'd say file a bug against selinux-policy for that issue,
> at least. I'm not sure what's going on there.
'Tis done: https://bugzilla.redhat.com/show_bug.cgi?id=818748
I expect that it will be a while before the bug is either fixed or
abandoned.
What to do in the mean time?
Should I use it as-is?
Will something blow up if I try to enable selinux?
Should I stick with F14?
Should I yum reinstall packages?
--
Michael hennebry(a)web.cs.ndsu.NoDak.edu
"On Monday, I'm gonna have to tell my kindergarten class,
whom I teach not to run with scissors,
that my fiance ran me through with a broadsword." -- Lily
--
test mailing list
test(a)lists.fedoraproject.org
To unsubscribe:
https://admin.fedoraproject.org/mailman/listinfo/test
--
Akshay vyas
2:40 p.m.
Michael Hennebry wrote:
I'm not looking for NX.
I only mentioned it because in another thread someone suggested that
NX might be the reason I couldn't get the F16 installer to boot.
Actually, I think you’ll find someone questioned whether PAE was
necessary (it isn’t, but a Pentium 4s has PAE anyway). The only reason I
mentioned NX was that PAE doesn’t buy you much unless you have NX or
more than 2 GB of memory (which I think you said you do).
Hope this helps,
James.
--
E-mail: james@ | As Tim switched off the engine, opened the driver’s door,
aprilcottage.co.uk | undid his seat belt and fell out of the tree, he decided
| that perhaps he didn’t have the makings of the best
| minicab driver. -- “I’m Sorry, I Haven’t A Clue”, BBC R4
3:20 p.m.
On Thu, Mar 29, 2012 at 12:25:46 -0500,
Michael Hennebry <hennebry(a)web.cs.ndsu.nodak.edu> wrote:
Has F17 been tested to see whether it will install on a Pentium 4?
Recently I installed F15 because I could not get the F16 installer to boot.
I can get the F13, F14 and F15 installers to boot without media.
The F16 installer was a no go, both without media and with a minimal CD.
In both cases, I got crashes that required pushing the reset button.
I never got to tell them about the iso file.
I have done live installs to a pentium M which should be a wimpier
processor. With P4s being dogs, I would expect P3 systems to work
OK as well.
Currently there is a check for 768 MiB of ram. This limit is too conservative
now that some changes have been made to the install process. I just did
a test install this morning on a machine with 512 MiB of ram and it
worked just fine. I used the --no-memcheck option on liveinst. I am not
sure where to put that option when using an install image instead of
a live image.
4376
days inactive
4411
days old
91 comments
13 participants
participants (13)
-
Adam Williamson -
Akshay Vyas -
Bruno Wolff III -
cornel panceac -
Felix Miata -
James Wilkinson -
John Reiser -
Josh Boyer -
Michael Cronenworth -
Michael Hennebry -
Michal Jaegermann -
mwesten -
Stephen John Smoogen