Bug 189572 - System hangs at boot because volume group for uuid not found
Summary: System hangs at boot because volume group for uuid not found
Keywords:
Status: CLOSED RAWHIDE
Alias: None
Product: Fedora
Classification: Fedora
Component: lvm2
Version: rawhide
Hardware: i686
OS: Linux
medium
high
Target Milestone: ---
Assignee: Alasdair Kergon
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2006-04-21 08:45 UTC by Joachim Frieben
Modified: 2007-11-30 22:11 UTC (History)
2 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2006-06-14 14:20:07 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)

Description Joachim Frieben 2006-04-21 08:45:10 UTC
Description of problem:
Volume group for uuid not found at boot. Consequently, the file
system cannot be mounted leading the kernel to panic.

Version-Release number of selected component (if applicable):
2.6.16-1.2147_FC6

How reproducible:
Always.

Steps to Reproduce:
1. Boot system.
  
Actual results:
System hangs when trying to mount the file system. First message
issued is "Volume group for uuid not found". Further error messages
are spawned until the kernel panics and the system hangs.

Expected results:
System should boot cleanly.

Additional info:
Happened first for version 2.6.16-1.2141_FC6 but still present
for current rawhide kernel. Booting kernel 2.6.16-1.2136_FC6
lets the system boot cleanly.

Comment 1 Dave Jones 2006-04-21 23:32:04 UTC
downgrading to the devicemapper & lvm2 from FC5 and rebuilding the initrd (or
reinstalling the kernel) makes this go away.


Comment 2 Joachim Frieben 2006-04-22 13:37:29 UTC
Maybe downgrading both packages in rawhide would be a suitable measure,
right? As a matter of fact, "Volume group for uuid not found" is even
issued for 2.6.16-1.2136_FC6 and possibly earlier kernels, but at least,
the system does not hang afterwards.

Comment 3 Alasdair Kergon 2006-04-22 17:34:51 UTC
I fixed this yesterday - only the lvm2 package is affected though you may need to regenerate initrd if it 
picked up a problematic lvm.static.  Unfortunately the build system had a long queue and it wasn't 
possible to get a replacement package built yesterday:-(

Comment 5 Alasdair Kergon 2006-04-22 18:22:32 UTC
OK, the package is now rebuilt except for s390/s390x.

lvm2-2.02.05-1.1


Comment 6 Joseph Sacco 2006-04-23 16:07:10 UTC
System: PowerMac G4 with dual 533MHZ CPU's, SCSI drives
=========================================================================
The problem persists after updating both lvm2 and device-mapper [22apr06]. 

As reported earlier the problem began when 2.6.16-1.2141_FC6 was released. I am
wondering if some other package that used when building a kernel is the real
culprit. Consider the following:

(1) 09Apr06: build a kernel locally
[2.6.15.6 9 (Tumbler-based sound works)]
==> all is well

(2) 19Apr06: build a different version kernel locally
[2.6.17-rc1-git13 (Tumbler-based sound does not work)]
==> all is well

(3) 20-21apr06: update kernels 2.6.16-1.2141_FC6, 2.6.16-1.2147_FC6
==> fail to boot

     * Volume group for uuid not found...

(4) 20-21Apr06: rebuild the 2.6.15.6 kernel using up to date rawhide env
==> kernel fails to boot

   * unable to find device-mapper major/minor
   * reading all pyhsical volumes. this may take some time
   * found volume group "VolGroup00" using metadata ...
   * /proc/misc: no entry for device-mapper found
                  ...

(4) 23apr06: rebuild the 2.6.15.6 kernel locally using up to date rawhide env
==> kernel fails to boot

-Joseph

Comment 7 Joseph Sacco 2006-04-23 16:15:37 UTC
Forgot to mention the relevant point...

The kernels built locally on 09Apr and 19Apr *still* work. I am running one now.

So... Whatever changed after 19Apr effects the static contents of vmlinux or
some of its modules.

-Joseph

Comment 8 Joseph Sacco 2006-04-24 12:41:22 UTC
Updated rawhide this morning bumping the kernel to version 2.6.16-1.2153_FC6. 
The problem has "gone away".

-Joseph


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