F10 post installation kernel issue?

[Rick Stevens]

Daniel B. Thurman wrote:
> Rick Stevens wrote:
>> Daniel B. Thurman wrote:
>>> Rick Stevens wrote:
>>>> Daniel B. Thurman wrote:
>>>>> Daniel B. Thurman wrote:
>>>>>>
>>>>>> First time F10 install went well.  One thing I did
>>>>>> differently in installing F10 was to:
>>>>>>
>>>>>> 1) Use the Volume based filesystems
>>>>>> 2) Enabled disk encryption
>>>>>>
>>>>>> I noticed that on every reboot, one must enter the password
>>>>>> long before seeing a grub display.  Hmm...  maybe for a server
>>>>>> this is not the way to go, but for a workstation, it's probably ok.
>>>>>>
>>>>>> Anyway, the initial kernel I started with is:
>>>>>> (1) kernel-2.6.27.5-117.fc10.i686
>>>>>>
>>>>>> I proceeded to get the latest updates and this was approx. 1 week 
>>>>>> ago.
>>>>>>
>>>>>> I later added programs I wanted installed, configured the services 
>>>>>> I wanted,
>>>>>> etc., etc., and everything went well.  I was able to reboot, no 
>>>>>> problems.
>>>>>>
>>>>>> But then a few days later, more updates came through, but 
>>>>>> specifically
>>>>>> a new kernel was added:
>>>>>> (2) kernel-2.6.27.9-159.fc10.i686
>>>>>>
>>>>>> Rebooting, I got the messages:
>>>>>> ======================
>>>>>> ata1: ACPI get timing mode failed (AE 0x300d)
>>>>>> Loading /lib/kdb/Keymaps/i386/qwerty/us.map
>>>>
>>>> Eh?  Sure that's not "/lib/kbd/keymaps/i386/qwerty/us.map" (/lib/kbd 
>>>> NOT
>>>> /lib/kdb and no capital K)?
>>>>
>>>> If what you posted is what's really being displayed, then we have
>>>> serious problems.  The correct directory is provided by kbd RPM.
>>>>
>>>>
>>>>
>>>>>> [hang]
>>>>>>
>>>>>> So, I never got to the point where I needed to enter
>>>>>> the encrypted disk password for continuance.
>>>>>>
>>>>>> To be sure, I rebooted back to the original kernel (1),
>>>>>> and it booted just fine.  Leaving it there, I continued using
>>>>>> the system, but got yet another kernel update:
>>>>>> (3) kernel-2.6.27.12-170.2.5.fc10.i686
>>>>>>
>>>>>> Same problem reported in (2) above.  So I am still
>>>>>> stuck at using my initial kernel at (1).
>>>>>>
>>>>>> Is there anything I can do or to check to understand why
>>>>>> I am not able to use the latest kernels?
>>>>
>>>> If the system is looking for the keymap you've shown, it won't find it,
>>>> the console won't be set up and things will come to a screeching halt.
>>>> I run 64-bit kernels so I can't test it and I don't know where it's
>>>> getting that path from.  I have run all the kernels you show and they
>>>> run fine here.  None ask for that funky keymap path.
>>> I double-checked and got that path wrong initially.
>>> The correct path shown on boot up (but appears ONLY
>>> with the later newer kernels) are:
>>>
>>> /lib/kbd/keymaps/i386/qwerty/us.map
>>>
>>> It still hangs.  The interesting thing is, as I said, I can
>>> boot with the first kernel (1) installed but not the ones
>>> following.  Still scratching my head...
>>
>> Ok, hmmm.  It looks like the initrd images didn't get built right.  Boot
>> up under the kernel that works, then as root:
>>
>>     # cd /boot
>>     # mkinitrd -f -v initrd-2.6.27.12-170.2.5.fc10.i686.img
>>         2.6.27.12-170.2.5.fc10.i686
>>
>> (the second and third lines should be ONE line...my mailer is wrapping
>> them)
>>
>> Then try rebooting using that -170 kernel again.  The keymaps and things
>> actually are in the initrd image as well as the main system.  See if
>> that does the job.
> Did what you suggested and it does not change anything.  Still hangs.
> I tried autorelabel for SeLinux just in case, no change.  It is subjective,
> but could having the filesystem encrypted be a problem?

I've heard of issues regarding encrypted filesystems.  I'm pretty sure
there's a thread on it in this forum somewhere.  It may be that the
initrd didn't get built with the cryptographic stuff.  I don't think
mkinitrd is smart enough to realize it needs the crypto stuff from the
/etc/modprobe.conf or /etc/fstab and you have to force feed it that
stuff with "--with=module" on the command line.

 >                                                            Did you try
> to see if you can run all the kernel version under an encrypted filesystem?

I don't use encrypted filesystems myself, so I'm going to have to bow
out on any of that stuff.  However, when you built the initrd, I 
recommended you use the "-v" flag.  You should have seen it include the
crypto modules along with the "dm-" stuff.  If not...well, that may be
your problem.

> I am now testing to see if removing any packages (one by one) has any 
> effect,
> a shot in the dark, but I do not know what else to do.

Don't think that's it.  It's quitting long before any other stuff is
loaded up...it's definitely having issues with the root filesystem,
and I'm willing to bet it is this crypto stuff.  Check the archives of
this list to find the thread(s) regarding it.
----------------------------------------------------------------------
- Rick Stevens, Systems Engineer                      ricks at nerd.com -
- AIM/Skype: therps2        ICQ: 22643734            Yahoo: origrps2 -
-                                                                    -
-  Tempt not the dragons of fate, since thou art crunchy and taste   -
-                         good with ketchup.                         -
----------------------------------------------------------------------