On 5/4/2014 11:10 PM, Koen Kooi wrote:
Op 5 mei 2014, om 00:04 heeft Bob Feretich <bob.feret...@rafresearch.com> het
volgende geschreven:
On 5/4/2014 7:09 AM, Koen Kooi wrote:
Op 4 mei 2014, om 11:07 heeft Bob Feretich <bob.feret...@rafresearch.com> het
volgende geschreven:
I am trying to build the Angstrom system that ships with the BeagleBone Black
from source.
bitbake cloud9-gnome-image
...
ERROR: Function failed: Fetcher failure for URL:
'git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git;branch=linux-3.8.y'.
Unable to fetch URL from any source.
It uses the EOLed 3.8 kernel, which seems to be no longer available at
kernel.org.
It's still available:
http://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/log/?h=linux-3.8.y
maybe kernel.org had a small hickup in their server?
When I bitbake, the do_fetch starts for the kernel... there is very little
network traffic, then the fetch times out.
After repeating this several times over a couple days (thinking that kernel.org
had a server problem) and receiving your response that the kernel EOLed 3.8
kernel was still available, I manually did the kernel git.
I can manually git clone or git fetch the kernel successfully.
bitbake -c cleanall virtual/kernel
...does not fix the condition.
Attempts to fetch the kernel via bitbake still fail.
The problems seems to be the state that OE has the current build is in.
Is there a way to reset this condition without starting he whole build over?
In the sources/downloads dir remove the dirs with 'kernel.org' in them in the
git2/ subdir (sources/downloads/git2/*kernel.org*).
Now I'm very confused. I deleted everything (files and directories) at
/git2 referencing kernel.org.
Performed a: bitbake -c cleanall virtual/kernel and restarted...
bitbake cloud9-gnome-image
20 minutes after the kernel fetch started there was still very very
little network traffic. (Even though parallel make was set to 8, the
kernel mainline fetch was the only task scheduled. I guess all other
steps were now dependent on it.)
So I interrupted the bitbake, and repeated the above rm and cleanall
commands.
I then changed the kernel recipe to access Google's kernel mirror rather
than the main kernel.org repository.
I got the same results... very very little network traffic.
So I created a fresh Angstrom directory (starting with cloning the
Angstrom setup scripts) and performed...
bitbake virtual/kernel
Everything worked fine until the kernel mainline fetch started,
eventually it became the only task running.
Still there was very very little network traffic.
The fetch eventually timed out "unable to fetch url from any source".
I tried a manual git clone of the kernel (copying the src repository url
from the recipe) and the kernel clone progresses smoothly.
I don't know why fetching the kernel during bitbake doesn't work and
cloning it manually does.
But, I have been stuck on this problem too long and since I plan to do
some kernel driver development, I should have a local kernel repository
anyway, I will proceed to modify the kernel recipe to access the
manually cloned kernel repository.
One thing that would help in the future would be to better document the
steps "do_fetch" goes through and provide a debug or trace option to
find out what it's doing when it appears to be hung. the -v and -DDD
flags produced no additional information once the "do_fetch" task started.
Internet searches produced many instances of bitbake "do_fetch" of known
working recipes hanging, but the only solution seemed to be try again or
create a local repository.
Regards,
Bob
regards,
Koen
_______________________________________________
Angstrom-distro-devel mailing list
Angstrom-distro-devel@linuxtogo.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
_______________________________________________
Angstrom-distro-devel mailing list
Angstrom-distro-devel@linuxtogo.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
