Bug 1110322 - The installer hangs on detecting BTRFS partitions installed by RHEL7
Summary: The installer hangs on detecting BTRFS partitions installed by RHEL7
Keywords:
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: Fedora
Classification: Fedora
Component: coreutils
Version: rawhide
Hardware: x86_64
OS: Linux
unspecified
high
Target Milestone: ---
Assignee: Ondrej Vasik
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2014-06-17 12:50 UTC by Jiri Eischmann
Modified: 2014-10-03 13:00 UTC (History)
14 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2014-10-03 13:00:08 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)

Description Jiri Eischmann 2014-06-17 12:50:53 UTC
I tried to installed Fedora Rawhide (2014-06-15) on a computer with RHEL 7 (installed on a BRTFS partition). The installer hangs recognizing partitions and the only thing I can do is to quit the installation.

Comment 1 David Shea 2014-06-17 15:59:54 UTC
Please attach the log files from /tmp to this bug as individual, text/plain attachments.

Comment 2 Vratislav Podzimek 2014-06-18 13:00:48 UTC
I have the reproducer-laptop on my desk.

Comment 3 Vratislav Podzimek 2014-06-18 13:31:01 UTC
The issue is that Anaconda tries to determine the architecture of the installed RHEL 7 system by running the 'arch' utility in the /mnt/sysimage chroot (the installed RHEL 7 system root). However, that hangs which hangs the whole Anaconda's storage-probing thread. Further investigation reveals that 'chroot /mnt/sysimage' doesn't work at all. Running 'strace chroot /mnt/sysimage' ends with access("/etc/ld.so.preload", R_OK) returning -1.

coreutils maintainers, any idea? Reassigning the bug cause it doesn't seem like anything anaconda could do something about.

Comment 4 Ondrej Vasik 2014-06-18 21:58:09 UTC
So ENOENT? Or which errno? Is ltrace more helpful there? Adding Lukas Czerner to cc - as this might be btrfs filesystem issue - I haven't seen such report before.

Comment 5 Vratislav Podzimek 2014-06-19 11:25:11 UTC
(In reply to Ondrej Vasik from comment #4)
> So ENOENT? Or which errno? Is ltrace more helpful there? Adding Lukas
> Czerner to cc - as this might be btrfs filesystem issue - I haven't seen
> such report before.
Don't have time now to run the whole process again to find out the errno, but listing the /etc/ dir didn't show the ld.so.preload file so I'm quite sure it was ENOENT.

Probably the more important information is that an older compose (~2 weeks old) works without any issues on the same laptop. So it seems like some problematic change in 'chroot' or something related.

Comment 6 Ondrej Vasik 2014-06-19 14:37:26 UTC
No, there was no change in chroot or arch of these utilities in the past several months. I would suspect kernel or glibc more...

Comment 7 Ondrej Vasik 2014-10-03 12:38:48 UTC
Any update here? As I said, no change in chroot/arch in these days and I don't have reproducer here. I tend to close it insufficient data - as without further analysis it doesn't have sense to reassign to glibc/kernel - which was likely the culprit.

Comment 8 Vratislav Podzimek 2014-10-03 12:49:49 UTC
(In reply to Ondrej Vasik from comment #7)
> Any update here? As I said, no change in chroot/arch in these days and I
> don't have reproducer here. I tend to close it insufficient data - as
> without further analysis it doesn't have sense to reassign to glibc/kernel -
> which was likely the culprit.
I'm okay with CLOSED INSUFFICIENT_DATA.

Comment 9 Ondrej Vasik 2014-10-03 13:00:08 UTC
Ok, thanks, closing.


Note You need to log in before you can comment on or make changes to this bug.