You have been subscribed to a public bug:

---Problem Description---
Not able to load the kdump kernel and save the vmcore in /var/crash/
  
---uname output---
Linux lep8d 3.16.0-5-generic #10-Ubuntu SMP Mon Jul 21 16:17:25 UTC 2014 
ppc64le ppc64le ppc64le GNU/Linux
 
Machine Type = P8 
  
---Steps to Reproduce---
Install a P8 machine with Ubuntu 14.10 in Non virtualised environment.
Installed all the kexec-tools and kdump-tools packages.

Then trying to start the kdump service and loading.

root@lep8d:~# /etc/init.d/kdump-tools start
Starting kdump-tools: Cannot open `/boot/vmlinuz-3.16.0-5-generic': No such 
file or directory
 * failed to load kdump kernel
root@lep8d:~# echo $?
0

root@lep8d:~# kdump-config load
Cannot open `/boot/vmlinuz-3.16.0-5-generic': No such file or directory
 * failed to load kdump kernel

root@lep8d:~# ls -l /boot/vmlinux-3.16.0-5-generic
-rw------- 1 root root 20712936 Jul 21 22:02 /boot/vmlinux-3.16.0-5-generic
root@lep8d:~#

root@lep8d:~# dpkg -l | grep kexec
ii  kexec-tools                                            1:2.0.6-0ubuntu2     
                      ppc64el      tools to support fast kexec reboots
ii  pxe-kexec                                              0.2.4-3              
                      ppc64el      Fetch PXE configuration file and netboot 
using kexec
root@lep8d:~# dpkg -l | grep kdump
ii  kdump-tools                                            1.5.6-2              
                      all          scripts and tools for automating kdump 
(Linux crash dumps)

root@lep8d:~# cat /sys/kernel/kexec_crash_loaded
0

root@lep8d:~# kdump-config show
USE_KDUMP:        1
KDUMP_SYSCTL:     kernel.panic_on_oops=1
KDUMP_COREDIR:    /var/crash
crashkernel addr:
current state:    Not ready to kdump

kexec command:
  no kexec command recorded
root@lep8d:~# kdump-config status
current state   : Not ready to kdump

Tried to manually trigger a crash as below:

root@lep8d:~# sysctl -w kernel.sysrq=1
kernel.sysrq = 1
root@lep8d:~# cat /proc/sys/kernel/sysrq
1
root@lep8d:~# echo c > /proc/sysrq-trigger

[ 4252.703681] SysRq : Trigger a crash
[ 4252.703773] Unable to handle kernel paging request for data at address 
0x00000000
[ 4252.703779] Faulting instruction address: 0xc0000000005b88f4
[ 4252.703807] Oops: Kernel access of bad area, sig: 11 [#1]
[ 4252.703852] SMP NR_CPUS=2048 NUMA PowerNV
[ 4252.703899] Modules linked in: dm_multipath scsi_dh shpchp powernv_rng 
uio_pdrv_genirq uio rtc_generic binfmt_misc parport_pc ppdev lp parport ses 
enclosure lpfc scsi_transport_fc ipr scsi_tgt
[ 4252.704162] CPU: 76 PID: 4635 Comm: bash Not tainted 3.16.0-5-generic 
#10-Ubuntu
[ 4252.704230] task: c000001fdf7cbeb0 ti: c000001fdf8e4000 task.ti: 
c000001fdf8e4000
[ 4252.704298] NIP: c0000000005b88f4 LR: c0000000005b997c CTR: c0000000005b88c0
[ 4252.704365] REGS: c000001fdf8e79d0 TRAP: 0300   Not tainted  
(3.16.0-5-generic)
[ 4252.704432] MSR: 9000000000009033 <SF,HV,EE,ME,IR,DR,RI,LE>  CR: 28422824  
XER: 20000000
[ 4252.704602] CFAR: c000000000009358 DAR: 0000000000000000 DSISR: 42000000 
SOFTE: 1
GPR00: c0000000005b997c c000001fdf8e7c50 c000000001346498 0000000000000063
GPR04: c000000014305db0 c000000014316618 0000000000018010 c0000000014ff2d8
GPR08: c000000000dd6498 0000000000000001 0000000000000000 0000000000018010
GPR12: c0000000005b88c0 c000000007e50a00 0000010016f94818 000000001016e008
GPR16: 000000001013ad70 0000010016f9c958 000000001016fed0 000000001016e008
GPR20: 00000000100c31e0 0000000000000000 0000000010171fc8 000000001016f840
GPR24: 000000001014d9b0 000000001014d0b0 0000000000000004 0000000000000000
GPR28: c00000000127eee8 0000000000000063 c00000000125d6a0 c00000000127f2a8
[ 4252.705502] NIP [c0000000005b88f4] sysrq_handle_crash+0x34/0x50
[ 4252.705558] LR [c0000000005b997c] __handle_sysrq+0xec/0x270
[ 4252.705604] Call Trace:
[ 4252.705631] [c000001fdf8e7c50] [c00000000018e3f0] 
__acct_update_integrals+0x80/0x170 (unreliable)
[ 4252.705722] [c000001fdf8e7c70] [c0000000005b997c] __handle_sysrq+0xec/0x270
[ 4252.705790] [c000001fdf8e7d10] [c0000000005ba138] 
write_sysrq_trigger+0x78/0xa0
[ 4252.705871] [c000001fdf8e7d40] [c0000000003141f0] proc_reg_write+0xb0/0x110
[ 4252.705940] [c000001fdf8e7d90] [c00000000028c07c] vfs_write+0xdc/0x260
[ 4252.706007] [c000001fdf8e7de0] [c00000000028ce1c] SyS_write+0x6c/0x110
[ 4252.706076] [c000001fdf8e7e30] [c00000000000a0fc] syscall_exit+0x0/0x7c
[ 4252.706143] Instruction dump:
[ 4252.706177] 3842dbd8 7c0802a6 f8010010 f821ffe1 60000000 60000000 3d22001b 
3949a954
[ 4252.706290] 39200001 912a0000 7c0004ac 39400000 <992a0000> 38210020 e8010010 
7c0803a6
[ 4252.706406] ---[ end trace 4473f520a462be3e ]---
[ 4252.706452]
[ 4254.712133] Kernel panic - not syncing: Fatal exception
[ 4254.712297] Rebooting in 1 seconds..

root@lep8d:~# dmesg | grep -i crash
[    0.000000] Reserving 128MB of memory at 128MB for crashkernel (System RAM: 
196608MB)
[    0.000000] Kernel command line: 
root=UUID=6c354e14-fa77-4e99-899b-dd21a3627c62 ro splash quiet 
crashkernel=384M-:128M

After triggering the crash, the machine rebooted and got the info in the
ipmi console.

 Petitboot Option Editor
 ------------------------------------------------------------------------------

 Device:         (*) sdb2 [6c354e14-fa77-4e99-899b-dd21a3627c62]
                 ( ) Specify paths/URLs manually

 Kernel:         /boot/vmlinux-3.16.0-5-generic
 Initrd:         /boot/initrd.img-3.16.0-5-generic
 Device tree:
 Boot arguments: 4e99-899b-dd21a3627c62 ro splash quiet crashkernel=384M-:128M

         [  OK  ]  [ Help ]  [Cancel]

The crashkernel parameter is set in the boot arguments as can be seen in
menu.

But once the machine boots to ubuntu 14.10 kernel, no vmcore is
generated in /var/crash folder.

Looks like kdump-config takes /boot/vmlinuz-`uname -r` as the kdump kernel by 
default.
One way to deal with this is to set KDUMP_KERNEL & KDUMP_INITRD variables
in /etc/default/kdump-tools configuration file.
Alternatively, there may be a need to add a provision in kdump configuration 
file
(/etc/default/kdump-tools) to add the type of kernel image (vmlinux/vmlinuz)
and make kdump-config script load the appropriate kernel image based on
the configuration.
Once this is sorted out, before testing kdump/kexec make sure the following 
patch
http://lists.infradead.org/pipermail/kexec/2014-July/012247.html is part of 
kexec-tools
as kdump is bound to fail otherwise as mentioned in LP bug 1349994

Regarding kdump-config issue mentioned by Hari, the patch mentioned in this 
launchpad's bug should be applied:
https://bugs.launchpad.net/ubuntu/+source/makedumpfile/+bug/1324544

Need to ask Ubuntu to include the patches provided in this bugzilla to be 
incorporated for Ubuntu 14.10.
We are currently testing Ubuntu 14.10 and the bug is occurring in this release.

** Affects: makedumpfile (Ubuntu)
     Importance: Undecided
         Status: New


** Tags: architecture-ppc64le bugnameltc-113879 severity-high 
targetmilestone-inin1410
-- 
Not able to load the kdump kernel and generate the dump in Ubuntu14.10 on Non 
virtualised system
https://bugs.launchpad.net/bugs/1350443
You received this bug notification because you are a member of Kernel Packages, 
which is subscribed to makedumpfile in Ubuntu.

-- 
Mailing list: https://launchpad.net/~kernel-packages
Post to     : kernel-packages@lists.launchpad.net
Unsubscribe : https://launchpad.net/~kernel-packages
More help   : https://help.launchpad.net/ListHelp

Reply via email to