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

Re: Against what do I file this bug?



2009/7/16 Stefan Hartsuiker <sahartsu xs4all nl>:
> On Thu, July 16, 2009 16:33, John5342 wrote:
>> 2009/7/16 Stefan Hartsuiker <sahartsu xs4all nl>:
>>> On Wed, July 15, 2009 17:31, Matthew Woehlke wrote:
>>>> S.A. Hartsuiker wrote:
>>>>> On 07/15/2009 01:24 AM, Matthew Woehlke wrote:
>>>>>> This sounds like an initrd problem to me, but I am no expert.
>>>>>
>>>>>  Uhm, no. I am in the ''Fedora Interactive'' bit, the
>>>>> /etc/rc.d/rc.sysinit at that moment.
>>>>
>>>> What I mean is, if you aren't getting something mounted, my first guess
>>>> would be that the contents of the initrd are wrong. That might be a
>>>> mkinitrd problem, but it might be something else. I am not a boot
>>>> process expert... ergo I probably will not be of further help; sorry.
>>>>
>>>
>>> Thanks for the effort anyway :)
>>>
>>>> I can think of one other question that might help someone else help you,
>>>> however. Is it only /boot that doesn't mount? What about /, swap, and
>>>> any other disk mounts (i.e. /proc, /sys are likely not interesting) you
>>>> have configured?
>>>
>>> The fakeraid mirrored device only contains two partitions, namely p2 ( to be
>>> mounted on /)  and p1 (to be mounted on /boot).
>>> All other things that are to be mounted can be mounted (at the point of the
>>> error not everything is yet moounted though)
>>>
>>> One other bit of an odd thing that may not have been made completely clear
>>> is that for the fakeraid mirrored device no device files exist. This
>>> includes
>>> the root partition/mount even though it is actually mounted ro at the point
>>> of
>>> the error.
>>>
>>> Another thing I just noticed is that the devicenode numbers (major/minor)
>>> that
>>> are used for the fakeraid mirrored device during a livecd run are 253/0,
>>> 253/1
>>> and 253/2. These numbers are in a ''normal'' boot sequence (the one that
>>> fails
>>> to startup correctly) used by the second fakeraid device.
>>>
>>> Does devicemapper or somesuch remove pre-existing devices if they don't
>>> match
>>> it's idea of waht is should be or does devicemapper just start at the
>>> ''beginning'' when creating devicenodes?
>>>
>>> Kind regards,
>>> Stefan Hartsuiker
>>
>> This may be completely un-related but ever since i first started using
>> FC3 (up to and including F11) i have had the same issue with nvidia
>> fake raid. I wasn't installing to the raid but the symptoms were
>> almost identical. My problem was that the dmraid command (which also
>> creates the devicemapper entries in /dev/) wasn't making it into the
>> initrd image. All i did was run the install dvd in rescue mode, chroot
>> to the install, run 'dmraid -ay' to make sure the fakeraid was enabled
>> and then recreate the initrd using mkinitrd.
>>
>
> Hmm, dmraid refuses to activate the fakeraid mirrored device. See the output
> below:
> [root sparhawk ~]# dmraid -d -ay -v
> DEBUG: _find_set: searching nvidia_ddcacjcd
> DEBUG: _find_set: not found nvidia_ddcacjcd
> DEBUG: _find_set: searching nvidia_ddcacjcd
> DEBUG: _find_set: not found nvidia_ddcacjcd
> DEBUG: _find_set: searching nvidia_bcfahchh
> DEBUG: _find_set: searching nvidia_bcfahchh
> DEBUG: _find_set: not found nvidia_bcfahchh
> DEBUG: _find_set: not found nvidia_bcfahchh
> DEBUG: _find_set: searching nvidia_bcfahchh
> DEBUG: _find_set: not found nvidia_bcfahchh
> DEBUG: _find_set: searching nvidia_bcfahchh
> DEBUG: _find_set: found nvidia_bcfahchh
> DEBUG: _find_set: searching nvidia_bcfahchh
> DEBUG: _find_set: found nvidia_bcfahchh
> DEBUG: checking nvidia device "/dev/sdc"
> DEBUG: set status of set "nvidia_ddcacjcd" to 16
> DEBUG: checking nvidia device "/dev/sda"
> DEBUG: checking nvidia device "/dev/sdb"
> DEBUG: set status of set "nvidia_bcfahchh" to 16
> RAID set "nvidia_ddcacjcd" already active
> INFO: Activating linear raid set "nvidia_ddcacjcd"
> RAID set "nvidia_bcfahchh" was not activated             <-- why?
> DEBUG: _find_set: searching nvidia_ddcacjcdp1
> DEBUG: _find_set: not found nvidia_ddcacjcdp1
> RAID set "nvidia_ddcacjcdp1" already active
> INFO: Activating partition raid set "nvidia_ddcacjcdp1"
> DEBUG: freeing devices of RAID set "nvidia_ddcacjcd"
> DEBUG: freeing device "nvidia_ddcacjcd", path "/dev/sdc"
> DEBUG: freeing devices of RAID set "nvidia_bcfahchh"
> DEBUG: freeing device "nvidia_bcfahchh", path "/dev/sda"
> DEBUG: freeing device "nvidia_bcfahchh", path "/dev/sdb"
> DEBUG: freeing devices of RAID set "nvidia_ddcacjcdp1"
> DEBUG: freeing device "nvidia_ddcacjcdp1", path "/dev/mapper/nvidia_ddcacjcd"
>
> Here the nvidia_bcfahchh is the fakeraid mirrored device and nvidia_ddcacjcd is
> the other one. nvidia_bcfahchh consists of /dev/sda and /dev/sdb.
>
> Does anyone have an idea as to why dmraid does not activate nvidia_bcfahchh?
>
>> Since you installed to the disk i can't imagine why dmraid didn't make
>> it into initrd but you're situation did sound familiar. Hope it helps.
>>
> I think dmraid is actually in the initrd (not checked yet, will do later
> tonight), since it can find nvidia_ddcacjcd.

By dmraid being in initrd i actually meant the call to the command but
given the error your getting i would say what i suggested is not the
same problem you are facing.

> But al least your tip gives me a tool to produce more debugging output :)

Glad it was at least of some use.

A google for RAID set nvidia was not activated turned up a few
results. No solid answers from any of the links i found but most of
them seem to have one of 3 causes. Bad connections on sata drives,
hardware issues or a bug in dmraid so to answer your original question
check connections, then try filing a bug against dmraid and see if the
experts have any better ideas. In the meantime see if you have any
more luck with google.

Good luck.

-- 
There are 10 kinds of people in the world: Those who understand binary
and those who don't...


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