Juniper-EX交换机系统恢复.doc

1 交换机无法正常启动的解决办法

EX交换机在突然掉电或非正常关机的情况下，设备重新启动后，可能会出现一直卡着启动进程或OS引导失败的场景，本文列出几种常见的模式及解决方法（注：本文档中提到的操作可能出现设备原有配置丢失情况出现，另外，对于通过命令关机时，强烈建议等到所有进程都halt时再对进行掉电操作）。
通常，交换机无法正常启动时，通过console输出为：
1.系统直接进入Loader模式，Loader Prompt  ( loader >) ；
2.系统进入Debug模式，Debug Prompt  (db>) ；
3.系统进入UBoot模式，UBoot Prompt    (=>)；
4.系统能正常启动，但各系统进程无法正常加载；
5.系统提示Can't load kernel error  ；

2 系统直接进入Loader模式解决办法
方法一 TFTP方式恢复交换机
1.准备好TFTP服务器，然后把Junos安装文件上传至TFTP的root目录层次下；
2.用console连接交换机，并对设备进行加电；

# When you see the "loading /boot/defaults/loader.conf" display hit ENTER.
Then press [Enter] to boot immediately, or space bar for command prompt.
Hit the space bar to enter the manual loader. The loader > prompt displays.
(NOTE: There is a 1 second delay for hitting the space bar)
(TIPS: you can hit space bar after you see "Loading /boot/defaults/loader.conf" message)
3.配置交换机的管理ip地址；
loader> set ipaddr=192.168.100.2
loader> set netmask=255.255.255.0
loader> set serverip=192.168.100.1
4.从TFTP安装Junos;
loader> install tftp://192.168.100.1/jinstall-ex-9.1R2.7-domestic-signed.tgz
==============================================
This part of the boot log is related to the steps listed above
==============================================

Consoles: U-Boot console 
Found compatible API, ver. 7

FreeBSD/PowerPC U-Boot bootstrap loader, Revision 2.1
(marcelm@apg-, Wed Feb  6 11:23:55 PST 2008)
Memory: 1024MB
Loading /boot/defaults/loader.conf
/kernel data=0x9dc348+0x6df44 syms=[0x4+0x87bb0+0x4+0x8e152]

Hit [Enter] to boot immediately, or space bar for command prompt.             <-------HIT SPACE BAR-------

Type '?' for a list of commands, 'help' for more detailed help.
loader>       
loader> set ipaddr=192.168.100.2
loader> set netmask=255.255.255.0
loader> set serverip=192.168.100.1
loader> install tftp://192.168.100.1/jinstall-ex-9.1R2.7-domestic-signed.tgz
Speed: 100, full duplex
Package /jinstall-ex-9.1R2.7-domestic-signed.tgz is signed...
/kernel data=0x504104+0x32e60 syms=[0x4+0x50a00+0x4+0x6366f]
Kernel entry at 0xa0000100 ...
GDB: no debug ports present
KDB: debugger backends: ddb
KDB: current backend: ddb
Copyright (c) 1996-2008, Juniper Networks, Inc.
All rights reserved.
Copyright (c) 1992-2006 The FreeBSD Project.
... snip ...

==============================================
或参考KB11752 - Installing EX-Series software from a TFTP server from Loader prompt  OR Recovering from a Failed Software Upgrade on an EX Series Switch. 
方法二、USB方式恢复交换机
步骤一、将需要安装的Junos软件copy至U盘根目录下；
步骤二、用console连接交换机，并对设备进行加电；

# When you see the "loading /boot/defaults/loader.conf" display hit ENTER.
Then press [Enter] to boot immediately, or space bar for command prompt.
Hit the space bar to enter the manual loader. The loader > prompt displays.
(NOTE: There is a 1 second delay for hitting the space bar)
(TIPS: you can hit space bar after you see "Loading /boot/defaults/loader.conf" message)
步骤三、输入install    file:///jinstall-ex-3200-10.0S1.1-domestic-signed.tgz
注：交换机将从U盘进行安装；
或参考KB10386 - Boot from USB and Recovering from a Failed Software Upgrade on an EX-series Switch OR Booting an EX Series Switch Using a Software Package Stored on a USB Flash Drive

3 系统进入Debug模式解决办法
步骤一，在db>模式下输入以下命令收集相关信息：
db> x/s version     
db> alltrace
db> ps
db> show allvms
db> show msgbuf
db> show registers
db> show lock
db> show uma
步骤二，完成以上步骤后，输入cont
db> cont
Do you get the CLI prompt?
No - Continue with 步骤三.
Yes - Jump to 步骤四.
步骤三,重启交换机，从loader模式进行恢复；如果交换机无法进入loader模式，请开case进行处理或采用install format方式解决；
步骤四，在cli下用show system core-dumps看是否存在core-dump文件，如果有的话，可先进入shell模式层下，然后进入/var/core-dumps directory目录下，将文件上传至FTP服务器；
或参考KB20635 - While booting up, switch stuck in db> mode

4 系统进入UBoot模式解决办法
按照以下三个步骤进行操作：
1. 在UBoot模式输入reset重启交换机；
2. Break启动顺序进入loader>模式
# When you see the "loading /boot/defaults/loader.conf" display hit ENTER.
Then press [Enter] to boot immediately, or space bar for command prompt.
Hit the space bar to enter the manual loader. The loader > prompt displays.
(NOTE: There is a 1 second delay for hitting the space bar)
(TIPS: you can hit space bar after you see "Loading /boot/defaults/loader.conf" message)
3. 然后按照方法一中的Loader模式进行处理；

5 系统能正常启动，但其它各系统进程无法加载的解决方法
设备启动后存在以下报错
root@GZ-EX4200-1> show chassis alarms no-forwarding
error: the alarm-control subsystem is not running
root@GZ-EX4200-1> show chassis hardware detail no-forwarding
error: the chassis-control subsystem is not running
1. 对于10.4R3以前的版本，可按照Loader模式的方法进行处理；
2. 对于10.4R3之后的版本，存在双分区双系统，而Junos loader默认是用disk99,如果这个分区的系统坏了之后，会一直卡着Loader>模式；
步骤一，设备加电后，输入以下命令；
=> setenv loaddev disk99: <enter>
=> saveenv <enter>
=> reset <enter>
步骤二，在JunOS loader模式下选择从internal NAND flash 启动Junos
loader> set currdev=disk0: <enter>
loader> boot <enter>
步骤三，在shell模式删除以下文件；
<login> (Login as root)
# ls -al /boot/loader.conf.local <enter>
# rm /boot/loader.conf.local
步骤四，Revert the change made in step one:
# nvram unsetenv loaddev
步骤五，Reboot
6 Can't load kernel error解决办法
如果设备重启后，一直显示以下错误:
can't load '/kernel'
can't load '/kernel.old'
可参照DB模式进行恢复；

7 Junos在备用分区启动（KB23180）

开机提示：
Warning:****************************************************************************************
** **
** WARNING: THIS DEVICE HAS BOOTED FROM THE BACKUP JUNOS IMAGE **
** **
** It is possible that the primary copy of JUNOS failed to boot up **
** properly, and so this device has booted from the backup copy. **
** **
** Please re-install JUNOS to recover the primary copy in case **
** it has been corrupted. **
** **
***************************************************************************************

设备告警信息：
user@switch> show chassis alarms
1 alarms currently active
Alarm time Class Description
2011-02-17 05:48:49 PST Minor Host 0 Boot from backup root
查看加载路径：
root> show system storage partitions
fpc0:
--------------------------------------------------------------------------
Boot Media: internal (da0)
Active Partition: da0s1a
Backup Partition: da0s2a <-- this is the backup slice
Currently booted from: backup (da0s2a) <-- shows booted from that slice

Partitions information:
Partition Size Mountpoint
s1a 184M altroot
s2a 184M /
s3d 369M /var/tmp
s3e 123M /var
s4d 62M /config
s4e unused (backup config)
root> show system snapshot media internal
Information for snapshot on internal (/dev/da0s1a) (primary)
Creation date: Feb 24 11:32:07 2012
JUNOS version on snapshot:
jbase : 10.4I20120224_1123_bshekar
jcrypto-ex: 10.4I20120224_1123_bshekar
jdocs-ex: 10.4I20120224_1123_bshekar
jkernel-ex: 10.4I20120224_1123_bshekar
jroute-ex: 10.4I20120224_1123_bshekar
jswitch-ex: 10.4I20120224_1123_bshekar
jweb-ex: 10.4I20120224_1123_bshekar
jpfe-ex42x: 10.4I20120224_1123_bshekar
Information for snapshot on internal (/dev/da0s2a) (backup) <-- provides info for this slice/partition the switch booted off of and the date the file system was created
Creation date: Feb 14 05:42:42 2012 <-- if less than alarm date then customer should snapshot (it is a good way to confirm
JUNOS version on snapshot:
jbase : 11.2-20120214.0
jcrypto-ex: 11.2-20120214.0
jdocs-ex: 11.2-20120214.0
jkernel-ex: 11.2-20120214.0
jroute-ex: 11.2-20120214.0
jswitch-ex: 11.2-20120214.0
jweb-ex: 11.2-20120214.0
jpfe-ex42x: 11.2-20120214.0
解决办法：
把备用junos镜像copy到主用junos镜像
request system snapshot media internal slice alternate
查看junos启动镜像
show system storage partitions
确保开机启用主用junos镜像命令：
request system reboot slice alternate media internal
user@switch>show system snapshot media internal slice 1
user@switch>show system snapshot media internal slice 2