From: Koen Kooi <koen@dominion.thruhere.net>
To: openembedded-devel@lists.openembedded.org
Subject: Re: [meta-oe] do_rootfs fails with networkmanager (was network-manager-applet) included
Date: Thu, 13 Oct 2011 15:15:25 +0200 [thread overview]
Message-ID: <j76o9h$504$1@dough.gmane.org> (raw)
In-Reply-To: <201110102141.56124.schnitzeltony@gmx.de>
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Op 10-10-11 21:41, Andreas Müller schreef:
> On Monday, October 10, 2011 12:19:53 PM Koen Kooi wrote:
>> Op 25-09-11 15:06, Andreas Müller schreef:
>>> On Thursday, September 22, 2011 11:37:30 PM Andreas Müller wrote:
>>>> Hi
>>>>
>>>> For my xfce-image ( lives in meta-angstrom/recipes-angstrom/images
>>>> ) I want to make use of network-manager-applet which came in meta
>>>> oe few days ago.
>>>>
>>>> It seems that as soon as I add network-manager-applet do_rootfs
>>>> fails with something like:
>>>>
>>>>
>>>> | Configuring openssh-sftp. | Configuring xf86-input-evdev. |
>>>> Configuring cpufrequtils. | Configuring systemd-compat-units. |
>>>> Configuring gvfsd-ftp. | Collected errors: | * extract_archive:
>>>> Cannot create symlink from ./var/run to 'volatile/run': File
>>>> exists. | + '[' '!' -z '' ']' | +
>>>> package_tryout_install_multilib_ipk | + multilib_tryout_dirs= | +
>>>> '[' '!' -z
>>>> /home/Superandy/tmp/oe-core-eglibc/work/overo-angstrom-linux-gnueabi/xfce-base-image-1.0-r0/multilib_check.py
>>>>
>>>>
']' | | ERROR: Function 'do_rootfs' failed (see
>>>> /home/Superandy/tmp/oe-core-eglibc/work/overo-angstrom-linux-gnueabi/xfce-base-image-1.0-r0/temp/log.do_rootfs.28313
>>>>
>>>>
for further information) NOTE: package xfce-base-image-1.0-r0: task
>>>> do_rootfs: Failed ERROR: Task 8
>>>> (/home/Superandy/data/oe-core/sources/meta-angstrom/recipes-angstrom/images/xfce-base-image.bb,
>>>>
>>>>
do_rootfs) failed with exit code '1' ERROR:
>>>> '/home/Superandy/data/oe-core/sources/meta-angstrom/recipes-angstrom/images/xfce-base-image.bb'
>>>>
>>>>
failed
>>>>
>>> Looked around a bit for this issue. As soon as there is
>>> networkmanager included, do_rootfs fails with with something like
>>>
>>> | Collected errors: | * extract_archive: Cannot create symlink from
>>> ./var/run to 'volatile/run': File exists.
>>>
>>> In 'image'-folder of networkmanager I find a folder
>>> 'var/run/NetworkManager'. At do_rootfs this folder has moved to
>>> '/var/volatile/run/NetworkManager' ( see
>>> <build>/<machine>-angstrom-linux-gnueabi/<image>-1.0-r0/rootfs/var/volatile).
>>>
>>>
>>>
Seems as do_rootfs tries to setup a link 'var/run' where still a
>>> (remaining) folder is present - or the transition from var/run ->
>>> var/volatile/run is somehow incomplete.
>>>
>>> Any ideas how - or especially where to fix ( not only for
>>> networkmanager I guess ). Maybe this is an oe-core issue but there no
>>> networkmanager is available...
>>
>> I've finally managed to reproduce this! I vaguely recall you finding
>> out some more details, could you please give a summary of your
>> finding?
>>
>> regards,
>>
>> koen
>>
>>
> Hi Koen,
>
> In historical order trying not to cause confusion this time :-)
>
> 1. I found a conflict between OE-Core creating /var/run-symlink by
> 'package.bbclass' together with 'meta/files/fs-perms.txt' on the one side
> and angstrom 'base-files_3.0.14.bbappend' trying to create
> /var/run-directory on the other side. I sent a patch [1] which fixed the
> issue. Unfortunately I was a bit over-enthusiastic and sent the patch
> together with [2] in a series. 2. On the first run, I recognized, that
> the second part of the series caused missing directories and wrote that
> in [3]. Here I meant to skip only the _second_ part of the series. 3.
> You committed a workaround for /var/run issue [4]. 4. I asked if this is
> neccessary in [5]. Your comment in the source ('we do not want that')
> made me think it is mandatory to have /var/run as dir and I hoped to
> learn the reason. 5. The broken do_rootfs still scared me in [6]. 6.
> After none of the mails were responded, few days ago I decided to give
> your workaround a try, reverted my patch and - surprise - did not get an
> error on do_rootfs! So I thought I missed some magic and decided not to
> cause further noise on that.
So I did some more digging and dumped the contents of all the ipks in deploy
and here are the packages (except basefiles) that touch /var/volatile/
bind_9.7.2-P3-r0_armv7a.ipk
drwxr-xr-x root/root 0 2011-10-13 00:35 ./var/volatile/
drwxr-xr-x root/root 0 2011-10-13 00:35 ./var/volatile/run/
lrwxrwxrwx root/root 0 2011-10-13 00:35 ./var/run -> volatile/run
consolekit_0.4.5-r5_armv7a.ipk
lrwxrwxrwx root/root 0 2011-10-13 00:06 ./var/log -> volatile/log
drwxr-xr-x root/root 0 2011-10-13 00:06 ./var/volatile/
drwxr-xr-x root/root 0 2011-10-13 00:06 ./var/volatile/log/
drwxr-xr-x root/root 0 2011-10-13 00:06 ./var/volatile/log/ConsoleKit/
drwxr-xr-x root/root 0 2011-10-13 00:06 ./var/volatile/run/
drwxr-xr-x root/root 0 2011-10-13 00:06 ./var/volatile/run/ConsoleKit/
lrwxrwxrwx root/root 0 2011-10-13 00:06 ./var/run -> volatile/run
samba_3.5.6-r0_armv7a.ipk
drwxr-xr-x root/root 0 2011-10-13 00:18 ./etc/default/volatiles/
- -rw-r--r-- root/root 89 2011-10-13 00:17
./etc/default/volatiles/volatiles.03_samba
lrwxrwxrwx root/root 0 2011-10-13 00:18 ./var/log -> volatile/log
lrwxrwxrwx root/root 0 2011-10-13 00:18 ./var/lock -> volatile/lock
drwxr-xr-x root/root 0 2011-10-13 00:18 ./var/volatile/
drwxr-xr-x root/root 0 2011-10-13 00:18 ./var/volatile/log/
drwxr-xr-x root/root 0 2011-10-13 00:18 ./var/volatile/log/samba/
drwxrwxrwt root/root 0 2011-10-13 00:18 ./var/volatile/lock/
drwxr-xr-x root/root 0 2011-10-13 00:18 ./var/volatile/run/
lrwxrwxrwx root/root 0 2011-10-13 00:18 ./var/run -> volatile/run
samba-dbg_3.5.6-r0_armv7a.ipk
resolvconf_1.59-r1_all.ipk
drwxr-xr-x root/root 0 2011-10-13 00:39 ./var/volatile/
drwxr-xr-x root/root 0 2011-10-13 00:39 ./var/volatile/run/
drwxr-xr-x root/root 0 2011-10-13 00:39 ./var/volatile/run/resolvconf/
drwxr-xr-x root/root 0 2011-10-13 00:39
./var/volatile/run/resolvconf/interface/
Which makes me think the error message we get is reversed:
| * extract_archive: Cannot create symlink from ./var/run to
'volatile/run': File exists.
I read that as: 'ln -s /var/run /var/volatile/run', but the packages above
have 'ln -s /var/volatile/run /var/run'.
Can anyone please tripplecheck my interpretation of the error message?
regards,
Koen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)
Comment: GPGTools - http://gpgtools.org
iD8DBQFOluRtMkyGM64RGpERAtBuAJ4+TFhhrkNjVL7W6orZ6+10EIgifgCfZOEx
axGVAQixLwAzcMTDhrLNB+w=
=hr3N
-----END PGP SIGNATURE-----
next prev parent reply other threads:[~2011-10-13 13:21 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-09-22 21:37 do_rootfs fails with network-manager-applet included Andreas Müller
[not found] ` <201109251506.30431.schnitzeltony@gmx.de>
2011-10-10 10:19 ` [meta-oe] do_rootfs fails with networkmanager (was network-manager-applet) included Koen Kooi
2011-10-10 19:41 ` Andreas Müller
2011-10-13 13:15 ` Koen Kooi [this message]
2011-10-18 9:47 ` Sergey Lapin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='j76o9h$504$1@dough.gmane.org' \
--to=koen@dominion.thruhere.net \
--cc=openembedded-devel@lists.openembedded.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox