Bug 737278 - GPF during initial device scan
Summary: GPF during initial device scan
Keywords:
Status: CLOSED DUPLICATE of bug 736521
Alias: None
Product: Fedora
Classification: Fedora
Component: kernel
Version: 16
Hardware: x86_64
OS: Linux
unspecified
urgent
Target Milestone: ---
Assignee: Kernel Maintainer List
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks: F16Beta, F16BetaBlocker
TreeView+ depends on / blocked
 
Reported: 2011-09-10 14:42 UTC by Clyde E. Kunkel
Modified: 2011-09-12 15:36 UTC (History)
8 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2011-09-12 15:36:07 UTC


Attachments (Terms of Use)
screenshot of F16 beta TC2 gpf (1.77 MB, image/png)
2011-09-10 14:42 UTC, Clyde E. Kunkel
no flags Details
syslog of kernel bug (32.70 KB, text/plain)
2011-09-10 22:05 UTC, Clyde E. Kunkel
no flags Details

Description Clyde E. Kunkel 2011-09-10 14:42:49 UTC
Created attachment 522529 [details]
screenshot of F16 beta TC2 gpf

Description of problem:
GPF during initial device scan with F16 Beta TC2

Version-Release number of selected component (if applicable):
anaconda 16.17

How reproducible:
every time

Steps to Reproduce:
1. boot F16 Beta TC2 DVD
2. select defaults and basic devices
3.
  
Actual results:
during initial device scan, GPF failure occurs.  Screenshot attached.

Expected results:
Continue with installation

Additional info:  I recognize that this is probably not anaconda, and is probably mdadm, but I don't know how the two interact and since I did not see this error on any of the alpha dvds or live cds, then this is some kind of regression.  I also saw this with the test boot.iso created earlier this week, but couldn't follow up due to flooding.

I did try to start a netconsole to capture more info but was not successful.
If there is a way to start a netconsole before beginning the initial device scan, could someone, please, tell me how.  I followed the instructions in the Fedora docs, but I wonder if they apply to the anaconda world.

Comment 1 Clyde E. Kunkel 2011-09-10 14:54:44 UTC
From F15 running on same machine:


$ sudo mdadm --detail /dev/md127
/dev/md127:
        Version : 0.90
  Creation Time : Tue Aug 25 23:23:50 2009
     Raid Level : raid10
     Array Size : 81919872 (78.12 GiB 83.89 GB)
  Used Dev Size : 40959936 (39.06 GiB 41.94 GB)
   Raid Devices : 4
  Total Devices : 4
Preferred Minor : 127
    Persistence : Superblock is persistent

    Update Time : Sat Sep 10 10:52:42 2011
          State : active
 Active Devices : 4
Working Devices : 4
 Failed Devices : 0
  Spare Devices : 0

         Layout : near=2
     Chunk Size : 64K

           UUID : b9438b55:1d815c8b:bfe78010:bc810f04
         Events : 0.332344

    Number   Major   Minor   RaidDevice State
       0       8        1        0      active sync   /dev/sda1
       1       8       17        1      active sync   /dev/sdb1
       2       8       33        2      active sync   /dev/sdc1
       3       8       49        3      active sync   /dev/sdd1
$ sudo blkid /dev/sd* | grep raid
/dev/sda1: UUID="b9438b55-1d81-5c8b-bfe7-8010bc810f04" TYPE="linux_raid_member" 
/dev/sdb1: UUID="b9438b55-1d81-5c8b-bfe7-8010bc810f04" TYPE="linux_raid_member" 
/dev/sdc1: UUID="b9438b55-1d81-5c8b-bfe7-8010bc810f04" TYPE="linux_raid_member" 
/dev/sdd1: UUID="b9438b55-1d81-5c8b-bfe7-8010bc810f04" TYPE="linux_raid_member"

Comment 2 Adam Williamson 2011-09-10 18:43:17 UTC
Proposing as Beta blocker, under criterion "The installer must be able to create and install to software, hardware or BIOS RAID-0, RAID-1 or RAID-5 partitions for anything except /boot " (which we're interpreting as meaning the installer must be able to read existing RAID layouts).

Comment 3 Clyde E. Kunkel 2011-09-10 21:41:26 UTC
(In reply to comment #0)
> 
> I did try to start a netconsole to capture more info but was not successful.
> If there is a way to start a netconsole before beginning the initial device
> scan, could someone, please, tell me how.  I followed the instructions in the
> Fedora docs, but I wonder if they apply to the anaconda world.

Nevermind.  Got syslog working and am attaching a syslog of similar failure.

Comment 4 Clyde E. Kunkel 2011-09-10 22:05:10 UTC
Created attachment 522551 [details]
syslog of kernel bug

syslog of a crash that was not complete in that tty2 was still available.

FWIW: retrying and specifying devices and not selecting the raid devices allowed an install to a standard ext4 partition.

Comment 5 Clyde E. Kunkel 2011-09-10 22:11:04 UTC
this is probably a duplicate of 736521.

Comment 6 Tom H 2011-09-12 15:01:10 UTC
I can't install F16betaTC2 with mdadm. The installation's OK but at first reboot, I'm dropped to a "grub rescue" prompt (similar to what happened previously with an alpha but this time with the latest grub2).

I tried to boot into Anaconda's rescue mode twice but I got a GPF after hitting "continue" and my computer froze completely. I'm using VirtualBox on OS X.

I re-installed and hit the "grub rescue" prompt again. I've unfortunately deleted the two vistual disks but if I have the time later today, I'll do a re-install and both look at the array before Anaconda shuts down and try to bring up it through another install.

Comment 7 Adam Williamson 2011-09-12 15:36:07 UTC

*** This bug has been marked as a duplicate of bug 736521 ***


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