Red Hat Bugzilla – Bug 207474
Dynamically linked lvm utilities seg faults leaving broken initrd images
Last modified: 2007-11-30 17:11:43 EST
Description of problem:
Dynamically linked tools from the lvm2 packages segfaults, possibly due to
linked in libraries that has been updated. lvm.static works
Version-Release number of selected component (if applicable):
Steps to Reproduce:
segfault from all the normal utilities and lvm.static works as normal.
# rpm -q device-mapper libselinux ncurses
I'm attachming my lvm.conf and a backtrace from a segfaulting binary, is there
any other information that would be helpful?
Starting program: /usr/sbin/lvdisplay
Reading symbols from shared object read from target memory...done.
Loaded system supplied DSO at 0x721000
File descriptor 3 left open
File descriptor 4 left open
File descriptor 5 left open
Program received signal SIGSEGV, Segmentation fault.
0x00000000 in ?? ()
#0 0x00000000 in ?? ()
#1 0x0807304e in dev_create_file (filename=0x8cddfd0 "/etc/lvm/lvm.conf",
dev=0x0, alias=0x8ce2010, use_malloc=1) at device/dev-cache.c:70
#2 0x080728ce in read_config_file (cft=0x8cddfb8) at config/config.c:237
#3 0x0806f7a3 in _load_config_file (cmd=0x8cd8e90, tag=0x80b55d3 "") at
#4 0x0806fa3f in _init_lvm_conf (cmd=0x8cd8e90) at commands/toolcontext.c:377
#5 0x08070474 in create_toolcontext (the_args=0x80c5760) at
#6 0x0805734c in lvm2_main (argc=1, argv=0xbf8a2bc4, is_static=0) at
#7 0x0806b3ea in main (argc=0, argv=0x11) at lvm.c:20
Created attachment 136846 [details]
tarred down /etc/lvm directory
selinux is not enabled on this server btw
I think you've got fc4 and fc5 packages mixed up.
fc5 contains device-mapper-1.02.02-3.2 and lvm2-2.02.01-1.2.1
fc4 contains device-mapper-1.02.07-2.0 and lvm2-2.02.06-1.0.fc4
(There's evidently a missing 'conflicts' line in the fc4 dm package which should
have stopped you updating dm without also updating lvm2.)
Seems to be the culprit, i downgraded to the version from FC5 - and it works
agains :) I will try to reinstall the newer kernels to see if those boots now.
So I guess the fix is to update the packages in FC5 and push them out?
It seems this manifested itself as a problem with some laters updates, so I
guess there is a lot of installs out there which has gone the yum upgrade path
from FC4 to FC5 with this situation.
the FC5 packages generates the correct initrd
[root@mainty ~]# rpm -aq | grep device-mapper
[root@mainty ~]# rpm -aq | grep lvm
I stil hang with a "soft lockup detected on CPU#0".
I have to revert to older kernel (2.6.15) to have a working system.
new fc5 package updates on their way that should fix the numbering