Re: Fedora 17’s unified filesystem (/usr-move)

Tom H tomh0665 at gmail.com
Thu Feb 2 00:03:20 UTC 2012


On Wed, Feb 1, 2012 at 6:32 PM, Adam Williamson <awilliam at redhat.com> wrote:
> On Wed, 2012-02-01 at 18:04 -0500, Tom H wrote:
>> On Tue, Jan 31, 2012 at 5:23 PM, Adam Williamson <awilliam at redhat.com> wrote:
>> > On Fri, 2012-01-27 at 14:10 +0100, Harald Hoyer wrote:
>> >>
>> >> Fedora 17 will locate the entire base operating system in /usr. The directories
>> >> /bin, /sbin, /lib, /lib64 will only be symlinks:
>> >>  /bin → /usr/bin
>> >>  /sbin → /usr/sbin
>> >>  /lib → /usr/lib
>> >>  /lib64 → /usr/lib64
>>
>> Regarding bug 786261.
>>
>> Frank Murphy's found that creating a "/etc/ld.so.conf.d/usrmove.conf"
>> containing "/usr/lib" (as well as "/usr/lib64" for amd64) before
>> running ldconfig allows dracut to create a bootable initramfs after
>> the /usr move. Would doing so before running ldconfig in
>> "usrmove-convert.sh" be enough of a fix?
>
> That seems strange. /usr/lib and /usr/lib64 should be in the linker path
> already anyway...

I would've thought so too. (Is there something special about libc in
respect of its linker path?)

[root at localhost ~]# mv /etc/ld.so.conf.d/usrmove.conf .
[root at localhost ~]# ldconfig
[root at localhost ~]# ldconfig -p | grep 'libc.so'
	libc.so.6 (libc6, OS ABI: Linux 2.6.32) => /lib/libc.so.6
[root at localhost ~]# mv usrmove.conf /etc/ld.so.conf.d/
[root at localhost ~]# ldconfig
[root at localhost ~]# ldconfig -p | grep 'libc.so'
	libc.so.6 (libc6, OS ABI: Linux 2.6.32) => /usr/lib/libc.so.6
[root at localhost ~]#

Also, Frank pointed out, either here or in the bug report, that the
problem might come from "/lib/libc.so.6" being a double symlink.

[root at localhost ~]# ls -l /lib/libc.so.6
lrwxrwxrwx. 1 root root 12 Jan 29 20:08 /lib/libc.so.6 -> libc-2.15.so
[root at localhost ~]# ls -l /usr/lib/libc.so.6
lrwxrwxrwx. 1 root root 12 Jan 29 20:08 /usr/lib/libc.so.6 -> libc-2.15.so
[root at localhost ~]#


More information about the test mailing list