Chinaunix首页 | 论坛 | 博客
  • 博客访问: 561031
  • 博文数量: 375
  • 博客积分: 0
  • 博客等级: 民兵
  • 技术积分: 15
  • 用 户 组: 普通用户
  • 注册时间: 2013-09-20 10:21
文章分类

全部博文(375)

文章存档

2015年(1)

2014年(374)

分类:

2014-09-03 13:31:50

   以前有接触过老的Sun Netra系列的服务器,如Netra 20,1125,1405等机器,这类的机器故障最多的要数风扇和电源了,所以会经常看到这类机器前面板FAULT灯亮,这是就要先检查看看是不是风扇或电源的故障了。
   如果你没有带console线,只要检查系统有lom软件已经安装即可,详细如下:
   1.检查系统有没有装LOM软件包:
# pkginfo | grep SUNWlom
system      SUNWlomm                         LOMlite manual pages
system      SUNWlomr                         LOMlite driver (root)
system      SUNWlomu                         LOMlite Utilities (usr)
   2.如果有装的话,执行下面命令即可查看相关信息:
# lominfo event   
# lominfo psu
# lominfo info
# lominfo fan
# lominfo fault
   
   如果没有请用console线连接主机ALARMS的LOM口执行下面命令:
lom>environment     OR     lom>showenvironment
lom>showlogs -v
 
   另在写些lom相关的其他命令吧,呵呵!
lom -a: Display all LOM parameters & variables
lom -p: Display PSU status
lom -f: Display fan status
lom -c: Display LOM configuration settings
lom -l: Display alarm status
lom -w: Display watchdog status
lom -v: Display voltage & circuit breaker status
lom -t: Display internal temperature status
lom -e [n, level]: Display the last 10 [128 max] entries of the event log [that are at or above level (1-4)]
lom -A (on | off) #: Turn alarm # on or off
lom -B: Enable LOM/LOMlite backward compatability mode to Netra t1 100 (See bug 4701814 below)
lom -E (on | off): Turn event reporting on or off
lom -F (on | off): Turn Fault indicator on or off
lom -G default: Reprogram the LOM firmware per 110208-xx instructions
lom -W (on | off): Turn the watchdog on or off
lom -X char: Change the console's LOM escape character (Typically # -> "#.")
lomctl check: Verify status of all monitored components (eg. after repair)
lomctl faulton | faultoff: Turn Fault indicator on or off
lomctl escape=char: Change the console's LOM escape character (Typically # -> "#.")
 
Under the lom prompt, make sure all alarms are turned off.

lom> alarmoff 1
lom> alarmoff 2
lom> alarmoff 3
 
Then set the bootmode to reset_nvram
 
lom> bootmode reset_nvram (bootmode is one word)
 
You must then poweroff and then back on to activate this bootmode change:
 
lom> poweroff
lom> poweron
 
Once the system returns to the ok prompt, then execute a reset-all
 
ok reset-all
 
If your system still refuses to go to the ok prompt, it will tell you this time because an alarm would have been switched on and an event notice will appear on screen. If the alarm does come on, find out the alarm number which came on and consult SunSolve to look for possible causes.
 
How to restart LOM Driver:
 
If watchdog is enabled, then disable it: # lom -W off
 
Determine the LOM driver's module ID via the modinfo command (246 in the case below):
   246 102f786 b5a3 245 1 lom (LOMlite2 driver v1.73)
 
Halt the LOM driver: # lom -U
 
Unload the LOM driver using the module ID: # modunload -i 246
 
Reload the driver: # lom -c
 
Verify that no kern.crit messages are in the messages file &
# modinfo | grep lom Produces (LOMlite2 driver v1.78)
lom and ok Prompt
   /*********************************/
  /* Stop the lom prompt access    */
 /* from the serial port console. */
/*********************************/
/platform/sun4u/kernel/drv/lom.conf
serial_security=1;

   /******************************/
  /* There are three prompts    */
 /* available on the Netra-t1. */
/******************************/
ok>      (normal prompt when the OS is not running)
lom>     (available whether OS is running or not)
#        (the OS prompt)
   /************************************/
  /* To move between the "ok>" prompt */
 /* and the "lom>" prompt            */
/************************************/
ok> #.   There must be less than 1 second
         between the "#" and "."
lom>     This is the prompt you get
  /*******************************************/
 /* Root can change this character sequence */
/*******************************************/
lomctl
// Only the first character ("#") can be changed.
// You don't need to know the
// previous character to do this.
   /***************************/
  /* To find out the current */
 /* sequence do an "init 5" */
/***************************/
init 5
// This shuts down the system and
// gives you the "lom>" prompt
   /*******************************/
  /* Then to show the current    */
 /* two-character sequence type */
/*******************************/
lom> show escape
   /*************************************/
  /* To move between the "lom>" prompt */
 /* and the "ok>" prompt type         */
/*************************************/
lom> console

  /*********************************/
 /* Some useful lom> commands are */
/*********************************/
lom> poweron     (powers on the netra)
lom> poweroff    (powers off the netra)
lom> reset       (the same as the
                  ok> reset command)
lom> reset -x    (resets the CPU only,
                  externally initiated
                  reset XIR)
lom> break       (goes to the ok> prompt
                  from the OS- "Stop-a")
lom> environment (current status of all
                  components in system)
lom> check       (updates component status
                  after repairing it, run
                  "lom> environment" again)
lom> show eventlog (show last 10 events
                    on the system)
lom> set faulton        (turns on the fault LED)
lom> set faultoff       (turns off the fault LED)
lom> set alarmon [1-3]  (sets 1 of 3 alarms on)
lom> set alarmoff [1-3] (sets 1 of 3 alarms off)
lom> show model         (show the server model)
lom> show hostname      (same as uname -n)
lom> show        (help for the "show" command)
lom> help               (list of LOM commands)
   /****************************************/
  /* You can set the boot mode of a       */
 /* netra using the below "lom>" command */
/****************************************/
lom> bootmode [-u] [normal
                    |forth
                    |reset_nvram
                    |diag
                    |skip_diag]
  -u        allows sharing of the console
            and LOM device
  normal    server boots using the
            OBP environment settings
            -must reset server to take effect
  forth      server stops at the "ok>"
             prompt- same as "Stop-f"
            -must reset server to take effect
  reset_nvram -"ok> set-defaults" or "Stop-n"
            -must reset server to take effect
  diag       runs full POST-
             "ok> diag-switch? true" or "Stop-d"
            -must power off server and
             back on within 10 minutes
  skip_diag - ok> setenv diag-switch? false"
            -must power off server and
             back on within 10 minutes
 
阅读(731) | 评论(0) | 转发(0) |
给主人留下些什么吧!~~