[troubleshoot][daily][redhat] 设备反复重启故障排查

一台服务器设备，反复重启，每天重启数次。

一：原因分析及初步排异。

1. 硬件，内存主板，一一更换，甚至除了硬盘将整台机器都换掉了，依然重启。

2. 排除电源问题，换了电源线，换了插座，还是重启。

3. 那么接下来，还有三种可能：

　　A。内核问题，内核crash。（redhat的稳定性还是十分让人信赖的，这种可能性不高）

　　B。硬盘或文件系统故障。本质上，这样会导致内核crash。

　　C。程序自主reboot。（我们自己的程序reboot，或进了黑客放了reboot脚本。好无聊的黑客。。。。）

二：最好排除的，就是先解决内核的问题。

　　内核在crash那一刻是会发现，自己即将crash的，于是他会在临死前留下一些信息。告诉用户我发生了什么。可是问题在于：文件系统的复杂性，会导致内核临死之前文件系统也随之崩溃了。

　　通过重启之后查看日志，确实没有留下有用的信息。

　　这是时候我们还有另一种手段，netcosole，他的功能是吧内核日志从socket以udp的方式，自组IP包而不走协议栈，讲包推出网卡端口。包的格式为syslog格式。

netcosole使用：

1. 修改配置文件

[root@S205 ~]# cat /etc/sysconfig/netconsole

# This is the configuration file for the netconsole service.  By starting

# this service you allow a remote syslog daemon to record console output

# from this system.

# The local port number that the netconsole module will use

LOCALPORT=

# The ethernet device to send console messages out of (only set this if it

# can't be automatically determined)

DEV=enp3s0

# The IP address of the remote syslog server to send messages to

SYSLOGADDR=192.168.10.214

# The listening port of the remote syslog daemon

SYSLOGPORT=

# The MAC address of the remote syslog server (only set this if it can't

# be automatically determined)

SYSLOGMACADDR=:8d:5c:::

[root@S205 ~]#

cat /etc/sysconfig/netconsole

2. 启动服务

[root@S205 ~]# systemctl start netconsole

[root@S205 ~]# systemctl enable netconsole

当前系统及内核版本：

[root@S205 ~]# cat /etc/redhat-release

CentOS Linux release 7.3. (Core)

[root@S205 ~]# uname -a

Linux S205 3.10.-.el7.x86_64 # SMP Tue Nov  :: UTC  x86_64 x86_64 x86_64 GNU/Linux

[root@S205 ~]#

成功收到内核crash日志：

Jul  :: 192.168.10.205 [20239.422386] NMI watchdog: Watchdog detected hard LOCKUP on cpu

Jul  :: 192.168.10.205

Jul  :: 192.168.10.205 [20239.422529] Kernel panic - not syncing: Hard LOCKUP

Jul  :: 192.168.10.205 [20239.422543] CPU:  PID:  Comm: swapper/ Not tainted 3.10.-.el7.x86_64 #

Jul  :: 192.168.10.205 [20239.422561] Hardware name: LENOVO 10C0A038CD/ , BIOS FCKT73AUS //

Jul  :: 192.168.10.205 [20239.422579]  ffffffff818d9784

Jul  :: 192.168.10.205  90a5d9572fc8872b

Jul  :: 192.168.10.205  ffff88041edc5b18

Jul  :: 192.168.10.205  ffffffff81685fac

Jul  :: 192.168.10.205

Jul  :: 192.168.10.205 [20239.422603]  ffff88041edc5b98

Jul  :: 192.168.10.205  ffffffff8167f3b3

Jul  :: 192.168.10.205

Jul  :: 192.168.10.205  ffff88041edc5ba8

Jul  :: 192.168.10.205

Jul  :: 192.168.10.205 [20239.422627]  ffff88041edc5b48

Jul  :: 192.168.10.205  90a5d9572fc8872b

Jul  :: 192.168.10.205  ffff88041edc5ba8

Jul  :: 192.168.10.205  ffffffff818d948a

Jul  :: 192.168.10.205

Jul  :: 192.168.10.205 [20239.422651] Call Trace:

Jul  :: 192.168.10.205 [20239.422658]  <NMI>

Jul  :: 192.168.10.205  [<ffffffff81685fac>] dump_stack+0x19/0x1b

Jul  :: 192.168.10.205 [20239.422678]  [<ffffffff8167f3b3>] panic+0xe3/0x1f2

Jul  :: 192.168.10.205 [20239.422692]  [<ffffffff8108562f>] nmi_panic+0x3f/0x40

Jul  :: 192.168.10.205 [20239.422706]  [<ffffffff8112f0e6>] watchdog_overflow_callback+0xf6/0x100

Jul  :: 192.168.10.205 [20239.422725]  [<ffffffff8117465e>] __perf_event_overflow+0x8e/0x1f0

Jul  :: 192.168.10.205 [20239.422741]  [<ffffffff811752a4>] perf_event_overflow+0x14/0x20

Jul  :: 192.168.10.205 [20239.422759]  [<ffffffff81009d88>] intel_pmu_handle_irq+0x1f8/0x4e0

Jul  :: 192.168.10.205 [20239.422776]  [<ffffffff8168dbeb>] perf_event_nmi_handler+0x2b/0x50

Jul  :: 192.168.10.205 [20239.422793]  [<ffffffff8168f019>] nmi_handle.isra.+0x69/0xb0

Jul  :: 192.168.10.205 [20239.422808]  [<ffffffff8168f193>] do_nmi+0x133/0x410

Jul  :: 192.168.10.205 [20239.422822]  [<ffffffff8168e453>] end_repeat_nmi+0x1e/0x2e

Jul  :: 192.168.10.205 [20239.422838]  [<ffffffff8168d9c7>] ? _raw_spin_lock_irqsave+0x47/0x60

Jul  :: 192.168.10.205 [20239.422855]  [<ffffffff8168d9c7>] ? _raw_spin_lock_irqsave+0x47/0x60

Jul  :: 192.168.10.205 [20239.422871]  [<ffffffff8168d9c7>] ? _raw_spin_lock_irqsave+0x47/0x60

Jul  :: 192.168.10.205 [20239.422887]  <<EOE>>

Jul  :: 192.168.10.205  <IRQ>

Jul  :: 192.168.10.205  [<ffffffffa01fae23>] nvkm_fantog_update+0x43/0x110 [nouveau]

Jul  :: 192.168.10.205 [20239.422947]  [<ffffffffa01faf48>] nvkm_fantog_set+0x38/0x40 [nouveau]

Jul  :: 192.168.10.205 [20239.422976]  [<ffffffffa01fa378>] nvkm_fan_update+0xc8/0x210 [nouveau]

Jul  :: 192.168.10.205 [20239.423005]  [<ffffffffa01fa519>] nvkm_therm_fan_set+0x19/0x20 [nouveau]

Jul  :: 192.168.10.205 [20239.423035]  [<ffffffffa01f9b87>] nvkm_therm_update+0x97/0x310 [nouveau]

Jul  :: 192.168.10.205 [20239.423064]  [<ffffffffa01f9e17>] nvkm_therm_alarm+0x17/0x20 [nouveau]

Jul  :: 192.168.10.205 [20239.423106]  [<ffffffffa01fd2d3>] nvkm_timer_alarm_trigger+0x103/0x150 [nouveau]

Jul  :: 192.168.10.205 [20239.423147]  [<ffffffffa01fd3d0>] nvkm_timer_alarm+0x60/0xb0 [nouveau]

Jul  :: 192.168.10.205 [20239.423176]  [<ffffffffa01fb651>] alarm_timer_callback+0xd1/0xe0 [nouveau]

Jul  :: 192.168.10.205 [20239.423207]  [<ffffffffa01fd2d3>] nvkm_timer_alarm_trigger+0x103/0x150 [nouveau]

Jul  :: 192.168.10.205 [20239.423238]  [<ffffffffa01fd3d0>] nvkm_timer_alarm+0x60/0xb0 [nouveau]

Jul  :: 192.168.10.205 [20239.423266]  [<ffffffffa01faeea>] nvkm_fantog_update+0x10a/0x110 [nouveau]

Jul  :: 192.168.10.205 [20239.423295]  [<ffffffffa01faf0a>] nvkm_fantog_alarm+0x1a/0x20 [nouveau]

Jul  :: 192.168.10.205 [20239.423324]  [<ffffffffa01fd2d3>] nvkm_timer_alarm_trigger+0x103/0x150 [nouveau]

Jul  :: 192.168.10.205 [20239.423355]  [<ffffffffa01fd6fb>] nv04_timer_intr+0x6b/0xb0 [nouveau]

Jul  :: 192.168.10.205 [20239.423384]  [<ffffffffa01fd174>] nvkm_timer_intr+0x14/0x20 [nouveau]

Jul  :: 192.168.10.205 [20239.423419]  [<ffffffffa01ada87>] nvkm_subdev_intr+0x17/0x20 [nouveau]

Jul  :: 192.168.10.205 [20239.423458]  [<ffffffffa01ef7f9>] nvkm_mc_intr+0x79/0x110 [nouveau]

Jul  :: 192.168.10.205 [20239.423486]  [<ffffffffa01f4155>] nvkm_pci_intr+0x55/0xa0 [nouveau]

Jul  :: 192.168.10.205 [20239.423503]  [<ffffffff8113015e>] handle_irq_event_percpu+0x3e/0x1e0

Jul  :: 192.168.10.205 [20239.423521]  [<ffffffff8113033d>] handle_irq_event+0x3d/0x60

Jul  :: 192.168.10.205 [20239.423536]  [<ffffffff81133007>] handle_edge_irq+0x77/0x130

Jul  :: 192.168.10.205 [20239.424012]  [<ffffffff8102d26f>] handle_irq+0xbf/0x150

Jul  :: 192.168.10.205 [20239.424491]  [<ffffffff810f3c8a>] ? tick_check_idle+0x8a/0xd0

Jul  :: 192.168.10.205 [20239.424967]  [<ffffffff8169201a>] ? atomic_notifier_call_chain+0x1a/0x20

Jul  :: 192.168.10.205 [20239.425445]  [<ffffffff81698bef>] do_IRQ+0x4f/0xf0

Jul  :: 192.168.10.205 [20239.425921]  [<ffffffff8168dd6d>] common_interrupt+0x6d/0x6d

Jul  :: 192.168.10.205 [20239.426389]  <EOI>

Jul  :: 192.168.10.205  [<ffffffff81514052>] ? cpuidle_enter_state+0x52/0xc0

Jul  :: 192.168.10.205 [20239.426863]  [<ffffffff81514199>] cpuidle_idle_call+0xd9/0x210

Jul  :: 192.168.10.205 [20239.427314]  [<ffffffff8103516e>] arch_cpu_idle+0xe/0x30

Jul  :: 192.168.10.205 [20239.427793]  [<ffffffff810e7c95>] cpu_startup_entry+0x245/0x290

Jul  :: 192.168.10.205 [20239.428222]  [<ffffffff8104f12a>] start_secondary+0x1ba/0x230

Jul  :: 192.168.10.205 [    2.633081] nouveau ::00.0: priv: HUB0:  ffffffff (1b70820b)

Jul  :: S214 systemd: Started Session  of user root.

Jul  :: S214 systemd: Starting Session  of user root.

Jul  :: S214 systemd: Started Session  of user root.

这是正确的处理方式，不是去深入调查原因，也不是去hacking。

1. 升至最新版稳定内核。

2. 回退至前一版稳定内涵。

[root@S205 ~]# yum upgrade
Installing: 
 kernel                                                x86_64                              3.10.0-514.26.2.el7                                    updates                               37 M

已升最新，待观察：

[root@S205 ~]# cat /etc/redhat-release

CentOS Linux release 7.3. (Core)

[root@S205 ~]# uname -a

Linux S205 3.10.-514.26..el7.x86_64 # SMP Tue Jul  :: UTC  x86_64 x86_64 x86_64 GNU/Linux

三：仍然重启，现象与错误信息一致。

　　参考：https://stackoverflow.com/questions/44039958/kernel-panic-not-syncing-watchdog-detected-hard-lockup

　　好像是 nvidia 显卡的问题。

[root@S205 ~]# cat /etc/default/grub |grep CMDLINE

GRUB_CMDLINE_LINUX="crashkernel=auto rd.lvm.lv=cl/root rd.lvm.lv=cl/swap rd.driver.blacklist=nouveau nomodeset rhgb quiet"

[root@S205 ~]#

　　增加内核参数：rd.driver.blacklist=nouveau nomodeset

　　再观察。

四：Fixed

连续24小时未重启。

完。