do_fetch hangs on building linux-ti81xx-psp

[Kai Warncke]

Hello,
i got a problem while trying to build the linux-ti81xx-psp package.
The do_fetch started but nothing happens after that. No network IO
and just small process activity. The GIT-Repository seems to be ok.
I would appreciate any hints.

Kai


[kai@Zero2 ~/arbeit/hidav/HidaV/build]$ bitbake hydraip-image
Loading cache: 100% |
###################################################################################################################|
ETA: 00:00:00
Loaded 1740 entries from dependency cache.

OE Build Configuration:
BB_VERSION = "1.15.1"
TARGET_ARCH = "arm"
TARGET_OS = "linux-gnueabi"
MACHINE = "hidav"
DISTRO = "angstrom"
DISTRO_VERSION = "v2012.02-core"
TUNE_FEATURES = "armv7a vfp neon cortexa8"
TARGET_FPU = "vfp-neon"
meta-hidav = "master:905bf5aec68fa766590be060db79e3c1f88b3a80"
meta-java =
"(nobranch):ef2def27a34f906f8c5725ab91213ad70f7d5909"
meta-texasinstruments = "(nobranch):
294c589cd748ef627a39840bc216bfe7f11e32c7"
meta-angstrom =
"(nobranch):e1a39ca4822c3fde15f18a76cae1e49004014b79"
meta-oe = "(nobranch):
1fd309ff9c034c4ae796392901997a3522af03b4"
meta = "(nobranch):
29c71eeb98aa8ce0fb0e0a30483499525bf6305d"

NOTE: Resolving any missing task queue dependencies
NOTE: Preparing runqueue
NOTE: Executing SetScene Tasks
NOTE: Executing RunQueue Tasks
NOTE: Running task 748 of 2864 (ID: 253, /home/kai/arbeit/hidav/HidaV/
meta-texasinstruments/recipes-kernel/linux/linux-ti81xx-psp_2.6.37.bb,
do_fetch)
NOTE: package linux-ti81xx-psp-2.6.37-r24a
+gitr42b3d4eedd82ff7a1af974bdcd0fc29a100d8df5: task do_fetch: Started

[Thilo Fromm]

Hello Kai,

> i got a problem while trying to build the linux-ti81xx-psp package.
> The do_fetch started but nothing happens after that. No network IO
> and just small process activity. The GIT-Repository seems to be ok.
> I would appreciate any hints.

[...]

> NOTE: Running task 748 of 2864 (ID: 253, /home/kai/arbeit/hidav/HidaV/
> meta-texasinstruments/recipes-kernel/linux/linux-ti81xx-psp_2.6.37.bb,
> do_fetch)
> NOTE: package linux-ti81xx-psp-2.6.37-r24a
> +gitr42b3d4eedd82ff7a1af974bdcd0fc29a100d8df5: task do_fetch: Started

At this stage OpenEmbedded/bitbake would run a git clone on Texas
Instruments' kernel fork at
git://arago-project.org/git/projects/linux-omap3.git. This is an
initial step of processing the kernel recipe in TI's openembedded
layer; the corresponding recipe is at
"meta-texasinstruments/recipes-kernel/linux/linux-ti81xx-psp_2.6.37.bb".
Our experience is that cloning this git repo from TI's servers can
take a considerable amount of time; they don't seem to provide the
optimum of uplink and / or raw server power (e.g. github repos clone a
lot faster). So depending on TI's current server load, i.e. the amount
of clones from other TI users at the same time, this might take a
while.

So I guess it's time for a cup of tea :)

Hth,
Thilo

--
Dipl.-Ing (FH) Thilo Fromm, MSc., Embedded Systems Architect
DResearch Fahrzeugelektronik GmbH
Otto-Schmirgal-Str. 3, D-10319 Berlin, Germany
Tel: +49 (30) 515 932 228   mailto:fr...@dresearch-fe.de
Fax: +49 (30) 515 932 77    http://www.dresearch.de
Amtsgericht: Berlin Charlottenburg, HRB 130120 B
Ust.-IDNr. DE273952058
Geschäftsführer: Dr. M. Weber, W. Mögle

[Thilo Fromm]

Hello again,

> At this stage OpenEmbedded/bitbake would run a git clone on Texas
> Instruments' kernel fork

[...]

> So I guess it's time for a cup of tea :)
>

Heh, I just ran

git clone git://arago-project.org/git/projects/linux-omap3.git

at the command line to check what's going on at TI and it took about
20 minutes before the remote end started sending data. That's enough
time for an awful lot of tea.

Regards,

[Kai Warncke]

Hi,

yesterday I waited several hours, without any noticable progress, so
I'm
somehow sceptical. But I'll give it a try.
Thanks so far for the swift response.

Kai

[Kai Warncke]

Hi,

nearly 7 hours later I'm still stuck at the do_fork.
I guess the time for tea is expired by far.
When I run

git clone git://arago-project.org/git/projects/linux-omap3.git

at the command line everything starts fine.
The debug option of bitbake doesn't give me any hints either. Any
other Idea?

Kai

[Thilo Fromm]

Hi, Kai,

> nearly 7 hours later I'm still stuck at the do_fork.
> I guess the time for tea is expired by far.
> When I run
>
> git clone git://arago-project.org/git/projects/linux-omap3.git
>
> at the command line everything starts fine.
> The debug option of bitbake doesn't give me any hints either. Any
> other Idea?

This sure is weird. Maybe something went wrong when OE first checked
out the TI kernel and the whole thing got stuck. You might want to
clean the kernel package by issuing

bitbake -c cleanall virtual/kernel

and then start again. If this does not help then the only thing I can
come up with is to try from scratch, i.e. do a

rm -rf tmp-eglibc/ sstate-cache/

in build/, then start all over again.

Hth,

[Kai Warncke]

Hi Thilo,

> This sure is weird. Maybe something went wrong when OE first checked
> out the TI kernel and the whole thing got stuck. You might want to
> clean the kernel package by issuing
>
> bitbake -c cleanall virtual/kernel
>
> and then start again. If this does not help then the only thing I can
> come up with is to try from scratch, i.e. do a
>
> rm -rf tmp-eglibc/ sstate-cache/
>
> in build/, then start all over again.

unfortunately I did this a few times without any success.
This problem seems to be reproducable since a colleague has the same
problem. I'm curently trying to clone manually by running

git clone git://arago-project.org/git/projects/linux-omap3.git

and, if successfully finished, create a tar.gz so it can be
locally referenced by the .bb file. Cloning manually failed
once as well. So I'm guessing it's a problem of TI's server.

Kai

[Thilo Fromm]

Hello Kai,

>> bitbake -c cleanall virtual/kernel
>>
>> rm -rf tmp-eglibc/ sstate-cache/

>>
>
> unfortunately I did this a few times without any success.
> This problem seems to be reproducable since a colleague has the same
> problem. I'm curently trying to clone manually by running
>
> git clone git://arago-project.org/git/projects/linux-omap3.git
>
> and, if successfully finished, create a tar.gz so it can be
> locally referenced by the .bb file. Cloning manually failed
> once as well. So I'm guessing it's a problem of TI's server.

Don't forget to run

git checkout ti81xx-master

before creating the tarball - this is the branch HidaV is using.

We've had the occasional trouble connecting to the TI Linux fork git
tree in the past, and still have. Take today, for example - the
repository server was down completely for quite some time. Maybe TI
should switch to GitHub :)

Regards,