Linux 平台下Oracle 9i/10g/11gR1 IO-Fencing 的hangcheck-timer 模块说明
一.官网的说明
参考MOS：
Linux: Hangcheck-Timer Module Requirements for Oracle 9i, 10g, and11gR1 RAC [ID 726833.1]
Hangcheck_timermodule is required to run a supported configuration in Oracle Real ApplicationClusters environments on Linux, with Oracle releases 9i, 10g, or 11gR1RAC. This note identifies and outlines the requirements needed toconfigure hangcheck-timer in an Oracle Enterprise Linux, Red Hat Linux, or SUSELinux environment.
在Linux 环境下Oracle 9i，10g，11gR1 的RAC 需要配置Hangcheck_timer模块。
Note : Hangheck timer is notrequired starting with Oracle Clusterware 11gR2
Starting in release 9.2.0.2and later, Oracle RAC environments required using a new I/O fencing model,named the hangcheck-timer module. This module was implemented to replace theWatchdog module, which provided similar fencing functionality. Hangcheck-timerwas subsequently delivered as part of the standard kernel distribution forLinux kernel releases 2.4 and above.
从9.2.0.2版本开始，ORACLERAC环境需要使用一个新的I/O fencing模块，叫做hangcheck-timer模块。这个模块用来代替Watchdog模块，提供类似的fencing功能。Hangcheck-timer模块是标准的linux2.4以上的内核中的一个子功能被发布。
Hangcheck-timer shouldbe loaded at boot time, and monitors the Linux kernel for long operatingsystem hangs that could affect the reliability of a RAC node. It runs inkernel mode and uses the Time Stamp Counter (TSC) to catch scheduling delays ornode hangs. This is done by setting a timer, then checking when the timerfires as to whether it was delayed by more than the allowed margin oferror. If the duration exceeds the allowed time of (hangcheck_tick +hangcheck_margin seconds), the machine is restarted. Hangcheck-timer willnot cause reboots to occur due to CPU starvation.
--Hangcheck-timer应该在系统启动的时候被加载, 并对于能够影响RAC节点稳定性的长时间的系统操作HANG进行内核监控。它运行在内核级别并使用Time Stamp Counter(TSC)来捕捉调度的延迟和节点HANG。这是通过设置一个timer，然后检查这个timer的fires情况来判断是否延迟是否超过了误差的幅度。如果这个周期超过了允许的时间（也就是hangcheck_tick+hangcheck_margin秒），机器将会被重启，如果是CPU资源不足的时候，Hangcheck-timer将不会导致重启。
Hangcheck-timer requiresthree configuration parameters:
--Hangcheck-timer有三个配置参数:
(1) hangcheck_tick - defines howoften, in seconds, the hangcheck-timer checks the node for hangs. The defaultvalue is 60 seconds.
-- hangcheck_tick：定义了hangcheck-timer检查节点是否hang的频率，单位是秒，缺省是60秒.
(2) hangcheck_margin - defines howmuch margin is allowed, in seconds, between expected scheduling and realscheduling time. The default value is 180 seconds.
--hangcheck_margin：定义期望的和真正的scheduling之间允许的误差，单位是秒，缺省值是180秒.
(3) hangcheck_reboot - determinesif the hangcheck-timer restarts the node if the kernel fails to respond withinthe sum of the hangcheck_tick and hangcheck_margin parameter values. If thevalue of hangcheck_reboot is equal to or greater than 1, then thehangcheck-timer module restarts the system. If the hangcheck_reboot parameteris set to zero, then the hangcheck-timer module will not reboot the node,even if a hang is detected. The default value varies by kernelversion. In the 2.4 kernel, the default is 1. In 2.6 kernels, thedefault is 0.
--hangcheck_reboot：定义了如果内核在hangcheck-tick和hangcheck-margin相加的时间内响应失败的话，hangcheck-timer是否重启节点。如果hangcheck_reboot的值大于等于1,hangcheck-timer模块将会重启系统；如果设置为0，则即使系统hang的时候hangcheck-timer也不会重启系统。在linux 2.4的内核中，这个缺省值是1；在2.6的内核中，缺省值是0。
当hangcheck_reboot=1并且满足下面的公式时，hangcheck-timer将reboot系统: system hang time > (hangcheck_tick + hangcheck_margin)
All hangcheck-timer defaultvalues should be explicitly overridden when loading the kernel module, based onthe Oracle release as follows:
--所有的hangcheck-timer的参数的缺省值必须在加载内核模块的时候被显式的覆盖，不同的oracle版本可以按照下面来设置：
（1）9i: Assuming thedefault setting of "oracm misscount" is set to 220 seconds:
hangcheck_tick=30hangcheck_margin=180 hangcheck_reboot=1
--9i: 假如"oracle misscount"的缺省设置是220秒，则hangcheck_tick=30hangcheck_margin=180 hangcheck_reboot=1
（2）10g/11gR1: Assuming thedefault setting of "CSS misscount" is set to either 30 or 60seconds:
hangcheck_tick=1hangcheck_margin=10 hangcheck_reboot=1
--10g/11gR1: 假如"CSS misscount"的设置是30或者60秒，则hangcheck_tick=1hangcheck_margin=10 hangcheck_reboot=1
You must always ensure thatthe Cluster misscount setting is greater than the sum of the setting forhangcheck_tick + hangcheck_margin.
--注意：你必须设置集群的misscount值大于hangcheck_tick + hangcheck_margin之和。
When running OracleClusterware on Linux, hangcheck-timer should always be configured on each RACcluster node, as the functionality of this module is required to provide I/O Fencingto ensure no stray writes will occur from an evicted node in a RACcluster. To verify if the hangcheck-timer module is running on a nodeexecute as the root or oracle user:
--Linux 平台上的Clusterware，需要在每个节点上配置hangcheck-timer模块，可以用root用户执行如下命令来验证hangcheck-timer是否运行：
# /sbin/lsmod | grep hangcheck
hangcheck-timer 2672 0
If the hangcheck-timer moduleis loaded (running) you will see output similar to above. When hangcheck-timeris not loaded no output is generated, and the command prompt is returned to theuser.
In an Oracle Enterprise Linux,Red Hat 4/5, or SUSE 9/10 environment the hangcheck-timer module is loadedusing the modprobe command:
--使用如下命令来装载hangcheck-timer：
# modprobe hangcheck-timer hangcheck_tick=1 hangcheck_margin=10hangcheck_reboot=1
In order to ensure the moduleis loaded at boot time, you should also place the same command in the appropriatelocal command execution directory (e.g. /etc/rc.d/rc.local, or/etc/init.d/boot.local). In earlier releases, hangcheck-timer was loadedusing insmod in place of modprobe. Consult your release specific documentationto determine which initialization method is required.
--为了确保在系统启动时就装载了hangcheck-timer模块，我们可以将命令添加到/etc/rc.d/rc.local,or /etc/init.d/boot.local中。
Hangcheck-timer will providemessage logging to the system messages log when a failure is detected, and anode restart is initiated by the module:
--当hangcheck-timer检测到系统hang时，会在系统log里记录日志并重启系统。
(1) When Hangcheck-timer reboots itmay leave "Hangcheck: hangcheck is restarting the machine" message in/var/log/messages。
-- hangcheck-timer的启动信息都会记录在系统日志里“ /var/log/messages”，重启时会记录"Hangcheck:hangcheck is restarting the machine"信息到/var/log/messages
(2) If you see the followingmessage in /var/log/messages: "Hangcheck: hangcheck value pastmargin!" this means a reboot was required but was not performed, becausehangcheck_reboot was not set to 1. If this message is seen, you mustreload the hangcheck module as described earlier in this note, with thehangcheck_reboot value set to 1.
--如果你看到/var/log/messages中有"Hangcheck:hangcheck value past margin!"消息，表示系统需要重启但是没有重启，因为hangcheck-reboot参数没有设置为1。
注：
Bug:6125546 which can preventhangcheck-timer from rebooting in RHEL4 (fixed in 2.6.9.56 or RHEL4.6)
二．说明
Hangcheck-timer 是Linux 提供的一个内核级的IO-Fencing 模块， 这个模块会监控Linux 内核运行状态， 如果长时间挂起， 这个模块会自动重启系统。 这个模块在Linux内核空间运行， 不会受系统负载的影响。 这个模块会使用CPU的Time Stamp Counter（TSC） 寄存器,这个寄存器的值会在每个时钟周期自动增加， 因此使用的是硬件时间，所以精度更高。
配置这个模块需要2个参数:hangcheck_tick 和 hangcheck_margin。
hangcheck_tick用于定义多长时间检查一次，缺省值是30秒。 有可能内核本身很忙， 导致这个检查被推迟， 该模块还允许定义一个延迟上限，就是hangcheck_margin， 它的缺省值是180秒。
Hangcheck-timer 模块会根据hangcheck_tick 的设置，定时检查内核。只要2次检查的时间间隔小于 hangcheck_tick +hangchec_margin, 都会认为内核运行正常，否则就意味着运行异常，这个模块会自动重启系统。
CRS本身还有一个MissCount 参数，可以通过crsctl get css miscount 命令查看。
当RAC结点间的心跳信息丢失时， Clusterware 必须确保在进行重构时，故障结点确实是Dead 状态，否则结点仅是临时负载过高导致心跳丢失，然后其他结点开始重构，但是结点没有重启，这样会损坏数据库。 因此MissCount 必须大于 hangcheck_tick+hangcheck_margin的和。
2.1 hangcheck-timer.ko模块安装
hangcheck-timer被默认安装在linux版本 2.4.9-e.12 及之上版本中，可以用如下命令核查hangcheck-timer是否安装。
[root@rac1 ~]# find /lib/modules-name "hangcheck-timer.ko"
/lib/modules/2.6.18-164.el5/kernel/drivers/char/hangcheck-timer.ko
/lib/modules/2.6.18-164.el5xen/kernel/drivers/char/hangcheck-timer.ko
有以上输出，代表已经安装了。
2.2 配置hangcheck-timer 模块
配置hangcheck-timer参数, 在/etc/modprobe.conf 中添加如下内容，这里根据数据库版本不同，内容也不同。
（1）9i: 假如"oracle misscount"的缺省设置是220秒，则hangcheck_tick=30hangcheck_margin=180 hangcheck_reboot=1
（2）10g/11gR1: 假如"CSS misscount"的设置是30或者60秒，则hangcheck_tick=1hangcheck_margin=10 hangcheck_reboot=1
如：
[root@rac1 ~]# vi /etc/modprobe.conf
options hangcheck-timer hangcheck_tick=30hangcheck_margin=180
2.3 配置系统启动时自动加载模块
在/etc/rc.d/rc.local 中添加如下内容
[root@rac1 ~]# modprobe hangcheck-timer
[root@rac1 ~]# vi /etc/rc.d/rc.local
modprobe hangcheck-timer
2.4确认模块加载成功
[root@rac1 ~]# grep Hangcheck/var/log/messages | tail -2
Sep 7 19:53:03 rac1 kernel:Hangcheck: starting hangcheck timer 0.9.0 (tick is 180 seconds, margin is 60seconds).
Sep 7 19:53:03 rac1 kernel:Hangcheck: Using monotonic_clock().
查证hangcheck-timer的参数是否正确的方法
[root@rac2 ~]# /sbin/lsmod |grep hangcheck
hangcheck_timer 7897 0
一般rac这几个参数的设置一般如下：
misscount=220
hangcheck_tick=30
hangcheck_margin=180
-------------------------------------------------------------------------------------------------------
版权所有，文章允许转载，但必须以链接方式注明源地址，否则追究法律责任!
Skype: tianlesoftware
QQ: tianlesoftware@gmail.com
Email: tianlesoftware@gmail.com
Blog: http://www.tianlesoftware.com
Weibo: http://weibo.com/tianlesoftware
Twitter: http://twitter.com/tianlesoftware
Facebook: http://www.facebook.com/tianlesoftware
Linkedin: http://cn.linkedin.com/in/tianlesoftware
-------加群需要在备注说明Oracle表空间和数据文件的关系，否则拒绝申请----
DBA1 群：62697716(满); DBA2 群：62697977(满) DBA3 群：62697850(满)
DBA 超级群：63306533(满); DBA4 群：83829929 DBA5群： 142216823
DBA6 群：158654907 DBA7 群：172855474 DBA总群：104207940
