Bug 12569

Summary:	Rare non-determinism in gdb means it doesn't build gdbserver
Product:	[Build System, Metadata & Runtime] OE-Core	Reporter:	Juro Bystricky <juro.bystricky>
Component:	core	Assignee:	Khem Raj <raj.khem>
Status:	RESOLVED OBSOLETE	QA Contact:
Severity:	normal
Priority:	Medium	CC:	meta.mr.watcher, meta.watcher, stephano
Version:	2.5
Target Milestone:	4.99
Hardware:	x86
OS:	Multiple
Whiteboard:
OS type for building Yocto:	---	Type of Regression:	---
Verified:		Documentation change:	Don't know

Description Juro Bystricky 2018-02-23 17:52:13 UTC

I failed to build core-image-sato-sdk with the current master for qemumis:

Build Configuration:
BB_VERSION           = "1.37.0"
BUILD_SYS            = "x86_64-linux"
NATIVELSBSTRING      = "universal"
TARGET_SYS           = "mips-poky-linux"
MACHINE              = "qemumips"
DISTRO               = "poky"
DISTRO_VERSION       = "2.4+snapshot-20180223"
TUNE_FEATURES        = "o32 bigendian fpu-hard mips32r2"
TARGET_FPU           = ""
meta                 
meta-poky            
meta-yocto-bsp       = "master:2f030dfc082d6d8143c712cd9e18a0da13ec66e1"

local.conf:
MACHINE = "qemumips"


I get the error:
ERROR: core-image-sato-sdk-1.0-r0 do_rootfs: Could not invoke dnf. Command '/data/master/poky/build-sato-testimage/tmp/work/qemumips-poky-linux/core-image-sato-sdk/1.0-r0/recipe-sysroot-native/usr/bin/dnf -y -c /data/master/poky/build-sato-testimage/tmp/work/qemumips-poky-linux/core-image-sato-sdk/1.0-r0/rootfs/etc/dnf/dnf.conf --setopt=reposdir=/data/master/poky/build-sato-testimage/tmp/work/qemumips-poky-linux/core-image-sato-sdk/1.0-r0/rootfs/etc/yum.repos.d --repofrompath=oe-repo,/data/master/poky/build-sato-testimage/tmp/work/qemumips-poky-linux/core-image-sato-sdk/1.0-r0/oe-rootfs-repo --installroot=/data/master/poky/build-sato-testimage/tmp/work/qemumips-poky-linux/core-image-sato-sdk/1.0-r0/rootfs --setopt=logdir=/data/master/poky/build-sato-testimage/tmp/work/qemumips-poky-linux/core-image-sato-sdk/1.0-r0/temp --nogpgcheck install kernel-devsrc packagegroup-core-sdk packagegroup-core-x11-base packagegroup-core-x11-sato packagegroup-core-standalone-sdk-target packagegroup-core-ssh-openssh packagegroup-core-boot packagegroup-core-eclipse-debug packagegroup-core-tools-profile packagegroup-base-extended rpm dnf packagegroup-core-tools-debug packagegroup-core-tools-testapps run-postinsts psplash locale-base-en-us locale-base-en-gb' returned 1:
Added oe-repo repo from /data/master/poky/build-sato-testimage/tmp/work/qemumips-poky-linux/core-image-sato-sdk/1.0-r0/oe-rootfs-repo
Last metadata expiration check: 0:00:04 ago on Fri 23 Feb 2018 05:45:32 PM UTC.
Error: 
 Problem 1: conflicting requests
  - nothing provides gdbserver needed by packagegroup-core-eclipse-debug-1.0-r0.noarch
 Problem 2: conflicting requests
  - nothing provides gdbserver needed by packagegroup-core-tools-debug-1.0-r3.noarch

ERROR: core-image-sato-sdk-1.0-r0 do_rootfs: Function failed: do_rootfs


And indeed, gdbserver is not built (for no obvious reason, however it does get built for qemumips64)

Comment 1 Ross Burton 2018-02-23 18:20:01 UTC

I thought we had a bug for this already.

Welcome to an incredibly rare race where sometimes gdbserver doesn't get build.

$10 says if you force gdb to rebuild, it appears.

Comment 2 Juro Bystricky 2018-02-23 18:34:26 UTC

(In reply to comment #1)
> I thought we had a bug for this already.

we do now

> 
> Welcome to an incredibly rare race where sometimes gdbserver doesn't get
> build.
> 
> $10 says if you force gdb to rebuild, it appears.

indeed, I did

$ bitbake gdb -ccleanall
$ bitbake gdb
$ bitbake core-image-sato-sdk

and no problems this time.

Comment 3 Ross Burton 2018-02-23 18:55:12 UTC

I should have asked you to rebuild it using -C instead of cleanall to preserve the logs of the good and bad builds :/

Comment 4 Ross Burton 2020-05-20 09:29:44 UTC

Reassigning to Khem in case he has any thoughts. Or, maybe it's not a bug anymore.  I've been meaning to throw check that the file exists into the recipe and make it error out if not.

Comment 5 Khem Raj 2020-05-20 13:30:38 UTC

(In reply to comment #4)
> Reassigning to Khem in case he has any thoughts. Or, maybe it's not a bug
> anymore.  I've been meaning to throw check that the file exists into the
> recipe and make it error out if not.

I have never encountered this error myself so have no idea, We have also upgraded
gdb bunch of times since this bug so perhaps its gone too, next time when it happens
I would like to take a look at the build and config.log files of gdb. Without this
its hard to tell what could be going on.

Comment 6 Khem Raj 2020-08-20 22:10:49 UTC

I can not reproduce this in master for few  many months now. Please re-open if it can still be reproduced.