Public bug reported:
After install linux-crashdump, everything seems successfully. But when I
trigger an panic over:
echo c > /proc/sysrq-trigger
, kernel panics and hungs, and doesn't reboot. even if i do the following:
echo 10 > /proc/sys/kernel/panic
And it is reproducible on my system.
Here is my system information:
rs@ubuntuDev$ uname -a
Linux ubuntuDev 3.19.0-80-generic #88~14.04.1-Ubuntu SMP Fri Jan 13 14:54:07
UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
And it is the lsb info:
rs@ubuntuDev$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04.2 LTS
Release: 14.04
Codename: trusty
rs@ubuntuDev$
Here is result of kdump-config show before i trigger an panic:
rs@ubuntuDev$ kdump-config show
DUMP_MODE: kdump
USE_KDUMP: 1
KDUMP_SYSCTL: kernel.panic_on_oops=1
KDUMP_COREDIR: /var/crash
crashkernel addr: 0x2d000000
current state: ready to kdump
kexec command:
/sbin/kexec -p --command-line="BOOT_IMAGE=/boot/vmlinuz-3.19.0-80-generic
root=UUID=5f248d51-af51-4178-bd4d-763d4a6648b3 ro console=ttyS0 console=tty0
irqpoll maxcpus=1 nousb" --initrd=/boot/initrd.img-3.19.0-80-generic
/boot/vmlinuz-3.19.0-80-generic
rs@ubuntuDev$
And I have successfully solved this problem by these steps after the above
kdump-config show:
1: /sbin/kexec -p
--command-line="root=UUID=5f248d51-af51-4178-bd4d-763d4a6648b3 ro console=ttyS0
console=tty0 irqpoll maxcpus=1 nousb"
--initrd=/boot/initrd.img-3.19.0-80-generic /boot/vmlinuz-3.19.0-80-generic
2: echo c > /proc/sysrq-trigger
And this time, system successfully rebooted into the secondary kernel whiout
crashkernel param, and the vmcore is dumpped successfully.
So, It seems we should strip out the BOOT_IMAGE string from result of `cat
/proc/cmdline`.
As a result, we need change kdump-config script from:
# Assemble the --commmand-line:
if [ -z "$KDUMP_CMDLINE" ] ; then
KDUMP_CMDLINE=`cat /proc/cmdline | \
sed -r -e 's/(^| )crashkernel=[^ ]*//g' \
-e 's/(^| )hugepages=[^ ]*//g' \
-e 's/(^| )hugepagesz=[^ ]*//g' \
-e 's/(^| )abm=[^ ]*//g'`
fi
into:
# Assemble the --commmand-line:
if [ -z "$KDUMP_CMDLINE" ] ; then
KDUMP_CMDLINE=`cat /proc/cmdline | \
sed -r -e 's/(^| )crashkernel=[^ ]*//g' \
-e 's/(^| )hugepages=[^ ]*//g' \
-e 's/(^| )hugepagesz=[^ ]*//g' \
-e 's/(^| )abm=[^ ]*//g'` \
-e 's/(^| )BOOT_IMAGE=[^ ]*//g'
fi
Could you guys re-check this bug?
I can provide other information with pleasure, if needed.
** Affects: kexec-tools (Ubuntu)
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to kexec-tools in Ubuntu.
https://bugs.launchpad.net/bugs/1771537
Title:
Kexec -p Seems load kernel failed, and doesn't reboot after a panic
Status in kexec-tools package in Ubuntu:
New
Bug description:
After install linux-crashdump, everything seems successfully. But when I
trigger an panic over:
echo c > /proc/sysrq-trigger
, kernel panics and hungs, and doesn't reboot. even if i do the following:
echo 10 > /proc/sys/kernel/panic
And it is reproducible on my system.
Here is my system information:
rs@ubuntuDev$ uname -a
Linux ubuntuDev 3.19.0-80-generic #88~14.04.1-Ubuntu SMP Fri Jan 13 14:54:07
UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
And it is the lsb info:
rs@ubuntuDev$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04.2 LTS
Release: 14.04
Codename: trusty
rs@ubuntuDev$
Here is result of kdump-config show before i trigger an panic:
rs@ubuntuDev$ kdump-config show
DUMP_MODE: kdump
USE_KDUMP: 1
KDUMP_SYSCTL: kernel.panic_on_oops=1
KDUMP_COREDIR: /var/crash
crashkernel addr: 0x2d000000
current state: ready to kdump
kexec command:
/sbin/kexec -p --command-line="BOOT_IMAGE=/boot/vmlinuz-3.19.0-80-generic
root=UUID=5f248d51-af51-4178-bd4d-763d4a6648b3 ro console=ttyS0 console=tty0
irqpoll maxcpus=1 nousb" --initrd=/boot/initrd.img-3.19.0-80-generic
/boot/vmlinuz-3.19.0-80-generic
rs@ubuntuDev$
And I have successfully solved this problem by these steps after the above
kdump-config show:
1: /sbin/kexec -p
--command-line="root=UUID=5f248d51-af51-4178-bd4d-763d4a6648b3 ro console=ttyS0
console=tty0 irqpoll maxcpus=1 nousb"
--initrd=/boot/initrd.img-3.19.0-80-generic /boot/vmlinuz-3.19.0-80-generic
2: echo c > /proc/sysrq-trigger
And this time, system successfully rebooted into the secondary kernel whiout
crashkernel param, and the vmcore is dumpped successfully.
So, It seems we should strip out the BOOT_IMAGE string from result of `cat
/proc/cmdline`.
As a result, we need change kdump-config script from:
# Assemble the --commmand-line:
if [ -z "$KDUMP_CMDLINE" ] ; then
KDUMP_CMDLINE=`cat /proc/cmdline | \
sed -r -e 's/(^| )crashkernel=[^ ]*//g' \
-e 's/(^| )hugepages=[^ ]*//g' \
-e 's/(^| )hugepagesz=[^ ]*//g' \
-e 's/(^| )abm=[^ ]*//g'`
fi
into:
# Assemble the --commmand-line:
if [ -z "$KDUMP_CMDLINE" ] ; then
KDUMP_CMDLINE=`cat /proc/cmdline | \
sed -r -e 's/(^| )crashkernel=[^ ]*//g' \
-e 's/(^| )hugepages=[^ ]*//g' \
-e 's/(^| )hugepagesz=[^ ]*//g' \
-e 's/(^| )abm=[^ ]*//g'` \
-e 's/(^| )BOOT_IMAGE=[^ ]*//g'
fi
Could you guys re-check this bug?
I can provide other information with pleasure, if needed.
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/kexec-tools/+bug/1771537/+subscriptions
--
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
