Re: [oe-commits] [Bug 5437] New: packaged-staging freezes

Re: [oe-commits] [Bug 5437] New: packaged-staging freezes

[Andrea Adami]

I had same stall on:

NOTE: package stagemanager-native-0.0.1-r12: task do_package_write: Succeeded

There is one patch is shr repo removing one lockfile in
packaged-staging.bbclass which helped my rebuild.

http://gitorious.org/~jama/angstrom/jama-shr-experimental/commit/1ae868c5f4029f1862555e97a1aba37a8935221c

I think RP could say something about this lock.

Regards

Andrea


On Thu, May 20, 2010 at 1:45 PM,  <bugzilla-daemon@git.openembedded.org> wrote:
> http://bugs.openembedded.org/show_bug.cgi?id=5437
>
>           Summary: packaged-staging freezes
>           Product: Openembedded
>           Version: org.openembedded.dev
>          Platform: ARM
>        OS/Version: Linux
>            Status: UNCONFIRMED
>          Keywords: angstrom
>          Severity: normal
>          Priority: P3
>         Component: unspecified
>        AssignedTo: openembedded-commits@lists.openembedded.org
>        ReportedBy: henri.bragge@dcombus.com
>
>
> First I successfully built a custom Angstrom image for an armv5 platform with
> PSTAGE_DIR defined outside of the tmp-directory. Then I removed tmp and tried
> to build again. This time the building froze after the second pstage package
> (coreutils-native) was installed.
>
> Last few rows of the build output shows this:
>
> NOTE: Running task 20 of 2123 (ID: 169,
> .../oe/os/openembedded/recipes/coreutils/coreutils-native_7.2.bb, do_setscene)
> NOTE: Running task 21 of 2123 (ID: 64,
> .../oe/os/openembedded/recipes/quilt/quilt-native_0.48.bb, do_setscene)
> Unpacking staging-coreutils-native-i686-linux...Unpacking
> staging-quilt-native-i686-linux...Done.
> Configuring staging-quilt-native-i686-linux...Done.
> Done.
> Configuring staging-coreutils-native-i686-linux...Done.
> (never goes any further)
>
> If I remove coreutils-native package from PSTAGE_DIR and rebuild, same happens
> with a different package. I did not find any reason for this from logs (even
> with bitbake -DDD).
>
> I'm building on i686 Debian lenny with bitbake 1.8.18 and OE
> 06fff17991cf5105cb29d3cec2640995ace329b5 (5/19/2010).
>
> Previously, with OE from April (before packaged-staging.bbclass rewrite) I
> didn't have these problems. One difference now is that there used to be only
> one DEPLOY_PSTAGE_DIR, but now there is PSTAGE_DIR and PSTAGE_WORKDIR. Should I
> define them both outside TMPDIR to be able to reuse staging packages?
>
> --
> Configure bugmail: http://bugs.openembedded.org/userprefs.cgi?tab=email
> ------- You are receiving this mail because: -------
> You are the assignee for the bug.
>
> _______________________________________________
> Openembedded-commits mailing list
> Openembedded-commits@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-commits
>

Re: [oe-commits] [Bug 5437] New: packaged-staging freezes

[Tom Rini]

Andrea Adami wrote:
> I had same stall on:
> 
> NOTE: package stagemanager-native-0.0.1-r12: task do_package_write: Succeeded
> 
> There is one patch is shr repo removing one lockfile in
> packaged-staging.bbclass which helped my rebuild.
> 
> http://gitorious.org/~jama/angstrom/jama-shr-experimental/commit/1ae868c5f4029f1862555e97a1aba37a8935221c
> 
> I think RP could say something about this lock.

iirc, that's mine or Chris'.  We want to have a lock on installing 
pstaging otherwise gcc-cross{,-initial,-intermediate} and 
libc{,-initial} can get done in the wrong order.

-- 
Tom Rini
Mentor Graphics Corporation

Re: [oe-commits] [Bug 5437] New: packaged-staging freezes

[Martin Jansa]

On Thu, May 20, 2010 at 02:07:34PM +0200, Andrea Adami wrote:
> I had same stall on:
> 
> NOTE: package stagemanager-native-0.0.1-r12: task do_package_write: Succeeded
> 
> There is one patch is shr repo removing one lockfile in
> packaged-staging.bbclass which helped my rebuild.
> 
> http://gitorious.org/~jama/angstrom/jama-shr-experimental/commit/1ae868c5f4029f1862555e97a1aba37a8935221c
> 
> I think RP could say something about this lock.

I can reproduce it easier.

After image build I call ie bitbake -k task-shr-minimal (that task is
already built with same PV/PR and rebuilding the same also hangs in
do_setscene).

With Enrico's patch I can rebuild tasks without problems.

Regards,

-- 
uin:136542059                jid:Martin.Jansa@gmail.com
Jansa Martin                 sip:jamasip@voip.wengo.fr 
JaMa                         

Re: [oe-commits] [Bug 5437] New: packaged-staging freezes

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 20-05-10 14:07, Andrea Adami wrote:
> I had same stall on:
> 
> NOTE: package stagemanager-native-0.0.1-r12: task do_package_write: Succeeded
> 
> There is one patch is shr repo removing one lockfile in
> packaged-staging.bbclass which helped my rebuild.
> 
> http://gitorious.org/~jama/angstrom/jama-shr-experimental/commit/1ae868c5f4029f1862555e97a1aba37a8935221c

I can confirm I see stalls as well, although I haven't tried that patch yet.

regards,

Koen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFL9ZClMkyGM64RGpERAkAOAJwNP9s32uHESDNihZ9BCYCnn3WXAwCfVKBm
dl3oCWs+kns2EXpJxdZT4j4=
=/7Al
-----END PGP SIGNATURE-----

Re: [oe-commits] [Bug 5437] New: packaged-staging freezes

[Tom Rini]

Tom Rini wrote:
> Andrea Adami wrote:
>> I had same stall on:
>>
>> NOTE: package stagemanager-native-0.0.1-r12: task do_package_write: 
>> Succeeded
>>
>> There is one patch is shr repo removing one lockfile in
>> packaged-staging.bbclass which helped my rebuild.
>>
>> http://gitorious.org/~jama/angstrom/jama-shr-experimental/commit/1ae868c5f4029f1862555e97a1aba37a8935221c 
>>
>>
>> I think RP could say something about this lock.
> 
> iirc, that's mine or Chris'.  We want to have a lock on installing 
> pstaging otherwise gcc-cross{,-initial,-intermediate} and 
> libc{,-initial} can get done in the wrong order.

OK, reproducing the problem locally and digging again, I can say quite 
confidently that when 3440328ac1ccbd5ca6aeea4a4b6084ecafd6fdda put the 
lock in that the above link takes back out, that was a merge error 
between Poky and oe.dev since Poky is/was missing 
394d321684eec271bc55438e3c01f74825d80988 where that lock got moved for 
the reasons I stated above.  You can add:
Acked-by: Tom Rini <tom_rini@mentor.com>
and merge that gitorious commit into mainline, asap please :)

-- 
Tom Rini
Mentor Graphics Corporation