Bug 3479

Summary: bitbake hangs while building pseudo
Product: [Build System, Metadata & Runtime] BitBake Reporter: Mikhail Lodigin <mishal>
Component: bitbakeAssignee: Richard Purdie <richard.purdie>
Status: RESOLVED NOTABUG QA Contact:
Severity: normal    
Priority: Medium CC: poky.bs.watcher, poky.watcher, sgw
Version: unspecified   
Target Milestone: 1.4   
Hardware: x86   
OS: x86   
Whiteboard:
OS type for building Yocto: --- Type of Regression: ---
Verified: Documentation change: ---
Attachments:
Description Flags
cooker log from hinged danny none

Description Mikhail Lodigin 2012-11-22 18:12:27 UTC 
in Danny, right after start "bitbake core-image-minimal" while building pseudo bitbake hangs (in ps is clearly seen two zombie python processes).

in Denzil everything works smoothly. 

I'm trying to build meta-fri2 in meta-intel

in order to reproduce one needs to download fri2-noemgd-denny-8.0.tar.bz2 and try to build fri2-noemgd machine.
Comment 1 Mikhail Lodigin 2012-11-22 18:45:36 UTC 
tested on ubuntu 12.04 - bug reproducible. on ubuntu 10.04 - no.
Comment 2 Richard Purdie 2012-11-28 09:14:38 UTC 
I run Ubuntu 12.04 and we also run test builds in QA on it so this is odd. Where exactly in the build process does it hang? Could you include the cooker log (tmp/log/cooker/*) for the hung build please? Logs from the hing task (tmp/work/*/<recipe>/temp/log*<task>*) would also be useful. When you say denzil works, was this on the same machine/hardware or a different one?

If you break out the hung build and retry it, does it continue or just hang in the same place? Does a build from scratch always stop in the same place?
Comment 3 Mikhail Lodigin 2012-11-28 14:51:41 UTC 
it hangs right in the beginning

Pseudo is not present but is required, building this first before the main build
Parsing recipes:  99% |######################################## | ETA:  00:00:0

percentage could be different but the result is the same always.
Comment 4 Mikhail Lodigin 2012-11-28 14:56:11 UTC 
it hangs ALWAYS in the same place, ether from scratch or after restart.

No difference machine/version/hardware between denzil and danny. two directories on the same level.
Comment 5 Mikhail Lodigin 2012-11-28 15:10:10 UTC 
Created attachment 957 [details]
cooker log from hinged danny
Comment 6 Saul Wold 2012-11-30 00:11:07 UTC 
There might be a second cooker log, can you attach the other one, this is the output of bitbake -e which will be helpful also, but we could use the log file.
Comment 7 Mikhail Lodigin 2012-11-30 15:58:17 UTC 
i'm sorry but *.log files present only in this 
poky-danny-8.0/build-fri2/tmp/log/cooker/fri2-noemgd/
directory
Comment 8 Richard Purdie 2012-12-14 14:57:54 UTC 
I've tried to reproduce this on a few different machines and I simply cannot see the behaviour you describe :(

The cooker log you've shared is one of the generated files which has the contents of bitbake -e in it. This shows the system is getting to a certain point but its still hard to tell where in the process its crashing/hanging.

I assume it never gets as far as showing anything like "Parsing of 845 .bb files complete (565 cached, 280 parsed). 1150 targets, 29 skipped, 0 masked, 0 errors."?

Perhaps you could run "strace -f bitbake core-image-sato -DDD 1> logfile1 2> logfile2" and then compress and share logfiles somewhere? The files will be huge but that should give us a better idea of what it was doing before it hung...
Comment 9 Richard Purdie 2012-12-14 14:58:11 UTC 
(mark as NEEDINFO)
Comment 10 Mikhail Lodigin 2013-01-30 17:49:29 UTC 
Hello,
After latest Ubuntu updates described problem had never occurred.
Thank you everyone for attention and help.