教你如何捕捉內核崩潰訊息

apt-get -y install aptitude kdump-tools crash kexec-tools makedumpfile linux-image-`uname -r`-dbg 
aptitude full-upgrade # 避免运行的内核版本与调试的版本不一致导致无法调试

USE_KDUMP=1
KDUMP_SYSCTL="kernel.panic_on_oops=1"
KDUMP_KERNEL=/boot/vmlinuz-3.16.0-4-amd64
KDUMP_INITRD=/boot/initrd.img-3.16.0-4-amd64
KDUMP_COREDIR="/data/crash"
KDUMP_FAIL_CMD="reboot -f"
DEBUG_KERNEL=/usr/lib/debug/vmlinux-3.16.0-4-amd64
MAKEDUMP_ARGS="-c -d 31"
KDUMP_CMDLINE="crashkernel=512M"

内存大小	crashkernel=
0 - 12G	128M
13 - 48G	256M
49 - 128G	512M
129 - 256G	1G *(896M, 768M o或 512M)

kernel.sysrq = 1
kernel.watchdog = 1
kernel.nmi_watchdog = 1
kernel.panic_on_oops = 1
kernel.softlockup_panic = 1
kernel.watchdog_thresh = 10

update-grub
reboot -f
kdump-config load 
kdump-config show 

# cat /proc/cmdline
BOOT_IMAGE=/boot/vmlinuz-3.16.0-4-amd64 root=UUID=a58ab901-00aa-4f8b-b3eb-d352fc72233 ro net.ifnames=0 thash_entries=1048576 rhash_entries=1048576 biosdevname=0 nohz=off enforcing=0 ipv6.disable_ipv6=1 nmi_watchdog=1 selinux=0 transparent_hugepage=never cgroup_enable=memory swapaccount=1 vga=771 crashkernel=512M
# kdump-config test
USE_KDUMP:         1
KDUMP_SYSCTL:      kernel.panic_on_oops=1
KDUMP_COREDIR:     /data/crash
crashkernel addr:  0x2e000000
kdump kernel addr:
kdump kernel:
   /boot/vmlinuz-3.16.0-4-amd64
kdump initrd:
  /boot/initrd.img-3.16.0-4-amd64
debug kernel:
  /usr/lib/debug/vmlinux-3.16.0-4-amd64
kexec command to be used:
  /sbin/kexec -p --command-line="BOOT_IMAGE=/boot/vmlinuz-3.16.0-4-amd64 root=UUID=a58ab901-00aa-4f8b-b3eb-d352fc7f6acb ro net.ifnames=0 thash_entries=1048576 rhash_entries=1048576 biosdevname=0 nohz=off enforcing=0 ipv6.disable_ipv6=1 nmi_watchdog=1 selinux=0 transparent_hugepage=never cgroup_enable=memory swapaccount=1 vga=771 irqpoll maxcpus=1 nousb systemd.unit=kdump-tools.service crashkernel=512M" --initrd=/boot/initrd.img-3.16.0-4-amd64 /boot/vmlinuz-3.16.0-4-amd64

echo c > /proc/sysrq-trigger

crash /data/crash/201609010252/dump.201609012233 /usr/lib/debug/lib/modules/3.16.0-4-amd64/vmlinux

WARNING: kernel version inconsistency between vmlinux and dumpfile # 版本不一致，full-upgrade 一下。
崩溃后系统不重启。 # 请检查 sysctl 、 crashkernel 及 nmi_watchdog 的设定！