[OE-core] Feedback on building openembedded-core for qemuarm. Excerpts from buildlog

Ulf Samuelsson openembedded-core at emagii.com
Thu Dec 1 14:49:06 UTC 2011


On 2011-11-30 18:30, Scott Garman wrote:
> On 11/29/2011 01:18 PM, Ulf Samuelsson wrote:
>>
>>>> Seen a couple of errors as well.
>>>>
>>>> 1. ERROR: Function 'useradd_sysroot' failed
>>>> Tried to access "/etc/group" but this was locked.
>>>> Problem disappeared the next time I rebuilt.
>>> Can you file a bug about this problem please. I think we need to go
>>> through the code paths in shadow and ensure its locking is sane. I took
>>> a quick look at the code and was left wondering what lckpwdf() does for
>>> example. Scott, could you take a look at this?
>>
>> Bugzilla is out of service, so bugs cannot be filed.
>
> I have filed a bug for this, and have been trying to reproduce it 
> without success:
>
> http://bugzilla.pokylinux.org/show_bug.cgi?id=1794
>
>> Haven't seen the issue since then so it is hard to give any details.
>> I used very high BB_NUMBER_THREADS = 24.
>
> I do have access to a system with a lot of cores, and have been doing 
> builds on it.
>
> I'd just like to confirm that when you saw this, it was a build from 
> scratch (e.g, not using sstate cache from a previous build). And also, 
> you *weren't* building this from a remote filesystem (NFS), correct?
>
> Thanks,
>
> Scott
>
No NFS for sure.

I think this was the first build, but only 99% sure.
It disappered when I restarted the build.

Have not seen this problem repeated, but then, the build on this
machine has other problems

Packages tend to fail since the c++ compiler seems to ignore "--sysroot"

Machine = Ubuntu-11.10 x64.

This is with BB_NUMBER_THREADS = "4"
If I change this, then the same problem occur, but in a different package.
(Tried with "4" AND "24")

BR
Ulf



-- 
Best Regards
Ulf Samuelsson
eMagii





More information about the Openembedded-core mailing list