Bug 10766

Summary:	patchtest: patchtest deletes the content of .git/config in openembedded-core
Product:	[Yocto Project Subprojects] Patchwork/Patchtest	Reporter:	Daniela Plascencia <daniela.plascencia>
Component:	Patchtest	Assignee:	Leonardo Sandoval Gonzalez <leonardo.sandoval.gonzalez>
Status:	RESOLVED NOTABUG	QA Contact:
Severity:	normal
Priority:	High	CC:	bluelightning, daniela.plascencia, leonardo.sandoval.gonzalez
Version:	2.3
Target Milestone:	2.3 M2
Hardware:	x86
OS:	Multiple
Whiteboard:
OS type for building Yocto:	---	Type of Regression:	---
Verified:		Documentation change:	No (bug/feature does not impact docs)

Description Daniela Plascencia 2016-12-05 17:16:11 UTC

It seems the patchtest execution in guest is corrupting the content of openembedded-core's .git/config, thus making further executions impossible due to dependencies on such repo. If ran for a first time, guest will do its job and test all mboxes, but when executed for the second time, the machine will stop and halt with the following message:

fatal: Not a git repository (or any parent up to mount point /patchtest)
Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).
INIT: Switching to runlevel: 0
INIT: Sending processes the TERM signal

Comment 1 Leonardo Sandoval Gonzalez 2016-12-06 17:40:49 UTC

Moving to M1

Comment 2 Leonardo Sandoval Gonzalez 2016-12-16 23:38:18 UTC

Daniela,

can you run the 'create-guest' script and follow directions to launch the new guest machine? The script has not being tested (neither the final indications) so please test it and let me know the result. Use the development branch [1] on the patchtest repo. This script will make it much easier to create the guest machine, so following it, 'identical' machines are produced no matter the host being used.

[1] http://git.yoctoproject.org/cgit/cgit.cgi/patchtest/log/?h=create-guest

Comment 3 Daniela Plascencia 2016-12-19 22:13:28 UTC

(In reply to comment #2)
> Daniela,
> 
> can you run the 'create-guest' script and follow directions to launch the
> new guest machine? The script has not being tested (neither the final
> indications) so please test it and let me know the result. 

The script builds and launches the guest machine as expected. I just had to change poky's branch a few commits before master as it had issues in build-time.

> Use the development branch [1] on the patchtest repo. This script will make it much easier to create the guest machine, so following it, 'identical' machines
> are produced no matter the host being used.

Guest seems to be working fine and the patchtest execution doesn't corrupt openembedded-core's .git/config anymore, but the the machine still halts before testing all mboxes [YOCTO 10765]. I haven't found a pattern here, but out of 61 mboxes, around 45 are tested correctly.

> [1] http://git.yoctoproject.org/cgit/cgit.cgi/patchtest/log/?h=create-guest

Comment 4 Leonardo Sandoval Gonzalez 2016-12-20 18:29:33 UTC

With the  reference script to create the guest machine, the guest machine does not cause the reported problem. Resolving it as not-a-bug.