 | | | bdflush general protection fault | bdflush general protection fault 2004-04-01 - By List manager
Back
I have a serious problem with a samba/mail server running
RHEL 3ES kernel 2.4.21-9.0.1.EL with all packages updated
On an irregular basis the server gets into a state where
files are not written to disk anymore.
Users can still create or modify files.
kupdated and kjournald get in DW state and sometimes kswapd too
a sync command will hang and stay in D state
reboot/shutdown does not work
The load starts to increase at about 1 unit/hour
only thing that can be done is a reset
All files created and/or modified on the /home partition after
the bdflush crash are lost.
Mail however, that lives in /var, is not lost
/home is ext3 mounted acl,defaults
/var is ext3 mounted defaults
Hardware is a Fujitsu-Siemens Primergy TX200
with Mylex RAID controller and 3 disks in RAID5
Adaptec aic79xx SCSI controller with HP DLT tape
single Xeon 2.66G (Hyperthreading disabled in BIOS) and 1.5G of RAM
Only non RHEL3 modules are the aic79xx.o and bcm5700.o which
are provided and recommended by Fujitsu-Siemens Computers.
(Primergy 's are RH certified)
I have four other machines that are almost identical (smp/up memory)
and they do not experience that problem.
This is what I found with dmesg:
-- ---- ---- -----snip-- ---- ---- ---- ------
general protection fault: 0000
sg smbfs autofs bcm5700 iptable_filter iptable_mangle iptable_nat
ip_conntrack
ip_tables microcode st nls_iso8859-1 nls_cp437 vfat fat keybdev mousedev
hid in
CPU: 0
EIP: 0060:[ <c011d25b >] Not tainted
EFLAGS: 00010086
EIP is at __wake_up [kernel] 0x1b (2.4.21-9.0.1.EL/i686)
eax: c61b09d8 ebx: ffffffff ecx: 00000001 edx: 00000003
esi: c61b09d8 edi: 00000001 ebp: f7f91f00 esp: f7f91ed8
ds: 0068 es: 0068 ss: 0068
Process bdflush (pid: 6, stackpage�f91000)
Stack: 067b9ac8 c01b54f8 c0408a60 00000001 c61b0988 00000286 00000003
c61b0988
00000008 00000001 00000000 c01b55d7 00000001 c61b0988 00000008
f7f91fa8
0001366e c0153c1b 00000001 c61b0988 c61b06b0 c61b0648 c0153d80
f7f91f44
Call Trace: [ <c01b54f8 >] generic_make_request [kernel] 0xe8 (0xf7f91edc)
[ <c01b55d7 >] submit_bh_rsector [kernel] 0x87 (0xf7f91f04)
[ <c0153c1b >] write_locked_buffers [kernel] 0x3b (0xf7f91f1c)
[ <c0153d80 >] write_some_buffers [kernel] 0x150 (0xf7f91f30)
[ <c0157a4c >] bdflush [kernel] 0x9c (0xf7f91fd4)
[ <c01579b0 >] bdflush [kernel] 0x0 (0xf7f91fe4)
[ <c010945d >] kernel_thread_helper [kernel] 0x5 (0xf7f91ff0)
Code: 8b 03 39 f3 89 45 e8 74 32 8d b6 00 00 00 00 8d bf 00 00 00
Kernel panic: Fatal exception
-- ---- ---- -----snip-- ---- ---- ---- ------
I have no idea what triggers this, or how I could reproduce it.
The server is never under heavy load. Only 8 users and 500 mails/day
Any help/pointers would be greatly appreciated.
Yves Bruggeman.
--
Taroon-list mailing list
Taroon-list@(protected)
http://www.redhat.com/mailman/listinfo/taroon-list
|
|
|