分类: 系统运维
2012-11-08 01:31:21
VRTS-raid5卷不能启动
# vxvol -g userdg2 stop v21
# vxprint -g userdg2 -ht
# vxvol -g userdg2 -f start v21
如果不行,就
# vxplex -g userdg2 -o rm dis v21-02
# vxvol -g userdg2 start v21
# vxassist -g userdg2 addlog v21 userdg201
DETACHED REPLAY修复方法:
enters this state during a log replay.
if the log replay dies with signal 10 or 11
run these commands.
vxplex -f -o rm dis
vxassist -g
vxvol -g
(1M)是什么意思呢?
number 1~9 stands for different section of the manuals, 1 and 1M usually means commands,often you will use such command to get what your need: "man 1 drvconfig" .
1: User Commands
1M: System Administration Commands
2: System Calls
3: Library Interfaces and Headers
3: Basic Library Functions
3: Networking Library Functions
3: Threads and Realtime Library Functions
3: Extended Library Functions
3: Curses Library Functions
4: File Formats
5: Standards, Environments, and Macros
6: Demos [Download this Book]
7: Device and Network Interfaces
9: DDI and DKI Overview
9F: DDI and DKI Kernel Functions
9S: DDI and DKI Data Structures
关于提示符请教
在.profile加入
ccd()
{
cd $1
PS1="`id |cut -d '(' -f2|cut -d ')' -f1' '`@`uname -n``pwd`>"
}
alias cd=ccd
SOLARIS下有抓图的键或者命令吗?
/usr/dt/bin/sdtimage可以取窗口/屏幕/矩形框截取图片,可存.gif/.jpg等
ASC - The additional sense code
ASCQ - The additional sense code qualifier
Please get more information about SCSI standards.
如何在主机中看HBA卡的WWN号
luxadm -e dump_map {enclosure[,dev]... | pathname...}
Upon startup, the Platform Information and Control Library daemon, loads and
initializes plug-in modules. To get information from the PICL tree, you can
use the prtpicl command interface.
In the output following, I have noted the port & node WWN with this
symbol: <======
To determine the WWN of a Sun[TM] Qlogic HBA, run the following
command(sample output included):
#prtpicl -v -c scsi-fcp
SUNW,qlc (scsi-fcp, 3e00000272)
eviceID 0x1
:UnitAddress 1
:vendor-id 0x1077
:device-id 0x2312
:revision-id 0x2
:subsystem-id 0x10a
:subsystem-vendor-id 0x1077
:min-grant 0x40
:max-latency 0
:interrupts 00 00 00 01
:class-code 0xc0400
:devsel-speed 0x1
:fast-back-to-back
:66mhz-capable
:fcode-rom-offset 0
:version ISP2312 FC-AL Host Adapter Driver: 1.13.08 04/10/02
:device_type scsi-fcp
:#address-cells 0x2
:#size-cells 0
:manufacturer QLGC
:compatible (3e0000027dTBL)
| pci1077,2312.1077.10a.2 |
| pci1077,2312.1077.10a |
| pci1077,10a |
| pci1077,2312.2 |
| pci1077,2312 |
| pciclass,0c0400 |
| pciclass,0c04 |
:reg
00 02 08 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 01 02
08 10 00 00 00 00 00 00 00 00 00 00 00 00 00 00 01 00 02
02 08 14
00 00 00 00 00 00 00 00 00 00 00 00 00 00 10 00
:node-wwn 20 00 00 e0 8b 0a 57 77 <====================
:port-wwn 21 00 00 e0 8b 0a 57 77 <====================
:assigned-addresses
81 02 08 10 00 00 00 00 00 00 20 00 00 00 00 00 00 00 01
00 82 02
08 14 00 00 00 00 00 20 00 00 00 00 00 00 00 00 10 00
:devfs-path /pci@1f,0/pci@1/SUNW,qlc
:binding-name SUNW,qlc
:instance -1
or try this script, it report only the wwn and the device path :
echo
printf "See below the results :\n"
echo
prtpicl -v -c scsi-fcp | grep port-wwn > /tmp/port-wwn
prtpicl -v -c scsi-fcp | grep devfs-path > /tmp/devfs-path
paste -d "\t" /tmp/devfs-path /tmp/port-wwn
To determine the WWN of a Sun JNI HBA, run the following command:
# prtpicl -v -c scsi
SUNW,jfca (scsi, 8e0000023c)
eviceID 0x2
:UnitAddress 2
:vendor-id 0x1242
:device-id 0x1560
:revision-id 0xb2
:subsystem-id 0x656b
:subsystem-vendor-id 0x1242
:min-grant 0x8
:max-latency 0x8
:interrupts 00 00 00 01
:class-code 0xc0400
:devsel-speed 0x1
:66mhz-capable
:fcode-rom-offset 0
:device_type scsi
:#size-cells 0
:#address-cells 0x2
:manufacturer JNIC
:copyright Copyright (c) 2000-2003 by JNIC
:fcode_revision Version 1.0
:version Version 1.0
:reg
00 00 10 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
01 00 10 10 00 00 00 00 00 00 00 00 00 00 00 00 00 00 01 00
03 00 10 14 00 00 00 00 00 00 00 00 00 00 00 00 00 00 10 00
01 00 10 1c 00 00 00 00 00 00 00 00 00 00 00 00 00 00 01 00
02 00 10 30 00 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
:pci_slot 0x1000
:power-consumption
00 00 00 00 00 00 00 00 03 30 00 00 03 61 50 00 00 00 00 00
00 00 00 00 00 60 00 00 01 00 00 00 00 00 00 00 00 00 00 00
:model FCX2-6562-L
:my_wwn_lo 0x7300e0ca
:my_wwn_hi 0x10000001
:board_id 0xe0ca
:board_revision 0x8001
:lport_cfg 0xffffffff
:seq_vers MIL_32_v4eng9
:fcode_cksum 0xf16
:sequencer_cksum 0x598a
:boots_wwn_lo 0xffffffff
:boots_wwn_hi 0xffffffff
:bootp_wwn_lo 0xffffffff
:bootp_wwn_hi 0xffffffff
:assigned-addresses
81 00 10 10 00 00 00 00 00 00 18 00 00 00 00 00 00 00 01 00
83 00 10 14 00 00 00 00 00 11 00 00 00 00 00 00 00 00 10 00
81 00 10 1c 00 00 00 00 00 00 1c 00 00 00 00 00 00 00 01 00
82 00 10 30 00 00 00 00 00 18 00 00 00 00 00 00 00 08 00 00
:cache-line-size 16
:latency-timer 64
:node-wwn 10 00 00 01 73 00 e0 ca <===============
:port-wwn 20 00 00 01 73 00 e0 ca <===============
:devfs-path /pci@1f,4000/SUNW,jfca@2
:driver-name jfca
:binding-name SUNW,jfca
:bus-addr 2
:instance 0
:_class scsi
:name SUNW,jfca
or try this script, it report only the wwn and the device path :
echo
printf "See below the results :\n"
echo
prtpicl -v -c scsi | grep port-wwn > /tmp/port-wwn
prtpicl -v -c scsi | grep devfs-path > /tmp/devfs-path
paste -d "\t" /tmp/devfs-path /tmp/port-wwn
NOTE: You will not be able to locate a SUN Branded Emulex cards WWN while the system is up and running.
you will need to use .properties command at the ok prompt.
The prtpicl command does not work with SBus Sun Qlogic cards (x6757, Ivory). To
find the WWN of these cards, 'prtconf -vp' can be used instead (reference infodoc 76953:
Sun StorEdge[TM] SAN Software: How to Get the WWN of Sun FiberChannel HBAs for LUN Masking.)
如何查看系统中安装了哪些shell,并且还能支持哪些shell?
# pkginfo | grep -i shell
and,
ls -al /bin/*sh /usr/bin/*sh /usr/local/bin/*sh | egrep -v "hash|remsh|rsh|p.*sh"
solaris怎么样查看一个文件是哪个包安装出来的?
pkgchk -lp yourfilename
请教一个双控制器连接服务器的问题
要做冗余,当然要对2个端口做map
SOLARIS中自己就带有multi path功能。修改/kernel/drv/scsi_vh*.conf,将mpxio_disable打开
如何实现文件的即时监控?
可能徐对文件做校验,并比较,
如用Solaris 10里有cksum,可比较文件变化前后的校验,如校验Solaris 10 DVD光盘media文件:
# cksum solarisdvd.iso
3842565141 3133603840 solarisdvd.iso
Solaris 10里还有BART,Basic Auditing and Reporting Tool,校验目录,帮助快速可靠的检查已部署的系统,创建已配置好的系统上的文件列表manifest,在系统上发现文件一级的改变,所以在配置好系统后应尽快建立系统或关键目录的文件列表manifest,并按系统的变化经常更新,如校验/etc目录:
# bart create -R /etc > etc2005-05-20
# bart create -R /etc > etc2006-01-16
Solaris 10里算法很多,自己琢磨下,选个速度满意的:
# cryptoadm list
用户级的提供者:
提供者: /usr/lib/security/$ISA/pkcs11_kernel.so
提供者: /usr/lib/security/$ISA/pkcs11_softtoken.so
内核软件提供者:
des
aes
arcfour
blowfish
sha1
md5
rsa
swrand
。。。
再如用md5签名校验:
# digest -l
sha1
md5
# digest -a md5 /etc/hosts
6cda1541bb58d86f3d3c9b13cf87aea7
校验越频繁,CPU消耗越大,
solaris下如何把登陆日志全部删除呢?
# cd /var/adm
# /usr/lib/acct/fwtmp < wtmpx > wtmpx.ascii
# vi + wtmpx.ascii
remove some lines including your login information
# /usr/lib/acct/fwtmp -ic < wtmpx.ascii > wtmpx
# rm wtmpx.ascii
# last | head
scripting crontab 动态修改
Remove #
CODE:
#!/bin/sh
EDITOR=/bin/ed; export EDITOR
/bin/crontab -e <
w
EOF
Add #
CODE:
#!/bin/sh
EDITOR=/bin/ed; export EDITOR
/bin/crontab -e <
w
EOF
System information:
Fire v880 operates on Solaris 8 with 6 disks.disk1 was organized as mirror of root disk disk0 and disk2 disk3 disk4 disk5 were oragnized as a raid 5 volume by disksuite4.2 . 2 replicas were placed at disk0-5 slice 4 so that system can boot correctly in case there is one disk failure .
Error information:
metastat command display that logic devices at disk0 (c1t0d0s2) need maintenace .
solution:
replace the failed disk c1t0d0s2 (disk0)
procedure:
1. gether system information
check the meta device state
#metastat
#metadb –i
check the slice location
#format
#more /vfstab
2. reboot from mirror disk
you can create the alias previously
ok nvalias disk1 /sbus@1f,0/espdma@e,8400000/esp@e,8800000/sd@1,0:a
pathname listed here is not in a 880 server but from a test machine Ultra1
ok setenv boot-device disk disk1
#reboot - - disk1
if there is no sufficient replicas , not in this case , when rebooting system will enter system maintenance due to the replica quorum .You can delete the replicas on the bad disk by:
#metadb –d –f c1t0d0s4
then you can boot system
3. replace hard disk (hot plug)
#luxadm remove_device /dev/rdsk/c1t0d0s2
while OK-to-Remove LED is lighting physically pull up the error disk from disk bay and insert the new disk. Then use command
#luxadm insert_device FCloop,s0
or
#devfsadm –C
In the command luxadm enclosure_name can get by command
#luxadm probe
Identify the new disk by compare the WWN number in the output of belowing command
#format
#ls –al /dev/rdsk/c1t0d0s2
4. repair replica database
#prtvtoc /dev/rdsk/c1t1d0s2 | fmthard - /dev/rdsk/c1t0d0s2
#metadb –a –c 2 /dev/rdsk/c1t0d0s4
5. resyncing the sub-mirrors
#metadetach d0 d10
#metadetach d1 d11
#metadetach d3 d13
#metadetach d5 d15
#metadetach d6 d16
#metadetach d7 d17
#metaclear d10 d11 d13 d15 d16 d17
#metainit d10 1 1 c1t0d0s0
#metainit d11 1 1 c1t0d0s1
#metainit d13 1 1 c1t0d0s3
#metainit d15 1 1 c1t0d0s5
#metainit d16 1 1 c1t0d0s6
#metainit d17 1 1 c1t0d0s7
#metattach d0 d10
#metattach d1 d11
#metattach d3 d13
#metattach d5 d15
#metattach d6 d16
#metattach d7 d17
#metastat |grep %
#reboot
#metastat
tricks:
1. sds user guide had mentioned the command metareplace to auto replace error submirror .But we have failed in using this command .You can find the same problem at InfoPartner Document BugReports Doc ID: 4777440.
2. when you mirror the root disk you must use command installboot bootblk ,but when you replace the root disk you will not need the same procedure.
开机时报warning:pcisch0:in0 0x16 has been blocked
warning:usba10_ehcio:interrupt #0 has been blocked
Try a workaround.
1. Add a line to /etc/system
set pcisch:pci_unclaimed_intr_max=0x7fffffff
2. reboot
我在solarse下面uncompress 一个4。5g大小的文件,刚开始都是执行的,
但是大概到了半个小时后,uncompress的cpu值就降低下来了,但是进程又不结束
1. Check what happened for uncompress process by truss ( sleeping? )
2. Install patch 116245-01 for Solaris 9
在SUN下用snoop抓包,看到如下内容,只有*** Header ,内部的DD在哪啊?
# snoop -v port 7777
# snoop -v -x 0 port 7777
将T3阵列的datadg导入之后
vxprint -ht
显示
v trunk001 gen DISABLED ACTIVE 4194304 SELECT -
pl trunk107-01 trunk107 DISABLED ACTIVE 4200448 CONCAT - RW
sd disk01-108 trunk107-01 disk01 487194624 4200448 0 c2t1d0 ENA
... ...
... ...
v trunk107 gen DISABLED ACTIVE 4194304 SELECT -
pl trunk107-01 trunk107 DISABLED ACTIVE 4200448 CONCAT - RW
sd disk01-108 trunk107-01 disk01 487194624 4200448 0 c2t1d0 ENA
~~~~~~~~~~~~~~~~~~~~~~~~~~~请问以上是代表什么意思
v vol01 fsgen DISABLED ACTIVE 41943040 SELECT -
pl vol01-01 vol01 ENABLED ACTIVE 41947136 CONCAT - RW
sd disk01-01 vol01-01 disk01 0 41947136 0 c2t1d0 ENA
# vxrecover -g datadg -sb
使用priocntl命令可以:
显示或设置指定进程的调度参
显示系统调度程序的当前配置信息
执行带有指定调度参数的命令
每个进程都有不同的级,每级分配有各自不同的调度策略。
系统可能配置的级包括:
系统(SYS)
交互(IA)
实时(RT)
分式共享(TS)
对于分式共享级,用户提供的优先权范围为:-20~+20。一个
分时共享进程的优先权(用户模式优先权)是从父进程继承
而来。系统在分时调度表中查找用户模式优先权,加入任何
nice值或这者priocntl优先权值(用户提供),保证创建的全
局优先权在0~59之间。
在默认配置中,一个可运行的实时进程在任何其它进程之前运行。实时进程使用不当会导致系统性能急剧下降。
1. 显示进程级的基本信息(进程级或系统调度参数):
输入priocntl -l并回车
2. 显示进程的全局优先级
使用ps -ecl命令显示进程的全局优先权。全局优先权在PRI栏列出
3、指定进程的优先级:
输入priocntl -e -c class -m user -limit -p priority command - name回车
-e 执行命令
-c 指定级(默认选项是TS(分时共享)或RI(实时)
-m 用户限制选项,指定-p选项提升或降低优先权的最大量。
-p 允许指定用户提供的优先权,范围为-20~+20。
4.更改分时共享进程的调度参数
依照以下步骤更改分时共享进程的调度参数:
1)成为超级用户。
2)输入priocntl -s -m user -limit [-p priority] -i id type id -list回车。
-s允许设置用户优先权范围的上限
-m 用户限制选项,指定-p选项提升或降低优先权的最大量。
-p优先权命令名选项,允许指定优先权。
-I id -type和id -list选项使用id -type和id -list的组合标识进程, id -
type指定id类型,如PID或UID。
3)输入ps -ecl|grep id -list回车。
4)检查PRI栏的输出,验证已经成功地更改了进程状态。
% priocntl -s -c RT -t 500 -p 20 myprog
% ps -ecl|grep myprog
5. 更改进程
依照以下步骤更改进程级:
1)成为超级用户。
2)输入priocntl -s -c class -i id type id -list回车。
-s允许设置用户优先权范围的上限并更改优先权
-c 指定想更改的进程级为TS(分时共享)级或RI(实时)级
-I id -type和id -list选项使用id -type和id -list的组合标识进程, id -type指定id
类型,如PID或UID。
3)输入ps -ecl|grep id -list回车。
4)检查PRI栏的输出,验证已经成功地更改了进程状态。
注意:必须是超级用户或者在实时shell中操作,更改进程级为非实时级或者实时级。
6、更改进程优先级
使用nice命令可以提高或降低一个命令或进程的优先权。
不带参数使用该命令则默认nice值增4个单位,使进程优先
权降低。
注意:必须是超级用户才能。
/usr/bin/nice command -name(默认,进程的优先权降低4个单位)
/usr/bin/nice +10 command -name(进程的优先权降低10个单位)
/usr/bin/nice -10 command -name(进程的优先权增加10个单位)
/usr/bin/nice - -10 command -name(nice 值减少10个单位,提高一个命令的优先权)
所有root拥有的文件全部改为ROot,如何快速恢复?
find dir-path -user ROot -exec chown root {} \;
try
# crle -u
# ldd /usr/sbin/vxdg
# crle
If failed
# mv /var/ld/ld.config /var/ld/ld.config.old
# crle -u
# ldd /usr/sbin/vxdg
# crle
[原创]
ufsdump and ufsrestore ,刚开始用的时候都以为他就是用来备份的(到TAPE),最近由于我们公司有几十台服务器要装,由于HDD又不一样大,所以觉得用DD太浪费了(72G变成了36G),于是就想到用UFSDUMP来试试。
具体方法为:
1。首选将你需要的HDD放在SERVER中,启动后,看一下分区表,大体上相同就可以,不能比第一个HDD的分区大小小。
2。用NEWFS命令将分区好的HDD都给初始化一下,newfs /dev/dsk/c0t1d0s0 .....
3。这时准备用了,
mount /dev/dsk/c0t0d0s0 /mnt
ufsdump 0f - /dev/dsk/c0t0d0s0|(cd /mnt; ufsrestore rf -)
等到执行完了后,表示这个分区就同步了。
umount /mnt
再用ufsdump 将其它的分区同步就可以了。
4。重要的一步是,在做完这些后,新的HDD是不能启动的,因为启动分区中是没有内容的,所以就需要用到
installboot
用法为: installboot /usr/platform/'uname -i'/lib/fs/ufs/bootblk /dev/rdsk/c0t0d0s0
注意此处的/dev/rdsk/c0t0d0s0 中RDSK 不能用DSK。
5。重起就可以了。
这样就会很方便,希望对大家有帮助。
下面是我测试过的,可用
1.在sunos中设置两块网卡pcn0(168.9.203.239)、pcn1(192.168.9.11)
# ifconfig -a
lo0: flags=1000849
inet 127.0.0.1 netmask ff000000
pcn0: flags=1000843
inet 168.9.203.239 netmask fffffc00 broadcast 168.9.203.255
ether 0:c:29:7d:3a:2a
pcn1: flags=1000843
inet 192.168.9.11 netmask ffffff00 broadcast 192.168.9.255
ether 0:c:29:7d:3a:34
2.
127.0.0.1 localhost
168.9.203.239 sophia loghost sophia.partner.com
192.168.9.11 partner
3.增加/etc/hostname.pcn1
partner
4.设置ip包转发
# ndd -set /dev/ip ip_forwarding 1
# ndd -set /dev/ip ip_forward_directed_broadcasts 1
# ndd -set /dev/ip ip_forward_src_routed 1
5.在客户端(16 设置网关为168.9.203.239
在客户端(192)设置网关为192.168.9.11
这样就可以通过sun机器访问不同网段了。
kill -9 $(ps -ef | grep username | awk '{print $3}')
对你的问题的比较赶兴趣.
可以把PING之间的路由表,PING之后的路由表给大家共享一下?
还有最好对照一下PING前后的Cache内容:
ndd -get /dev/ip ipv4_ire_status
如果没发现什么问题,就继续查交换机吧.
在两台主机加上如下文件 格式如下 加完后 两边都重起 在做METASET 前不久刚有人问过我 呵呵
vi /var/run/nodelist
1 host1 172.16.193.1
2 host2 172.16.193.2
powermt display dev=all可以显示native设备与pseudo设备之间的对应关系!
$ sqlplus /nolog
SQL*Plus: Release 8.1.7.0.0 - Production on Thu Dec 8 16:05:28 2005
(c) Copyright 2000 Oracle Corporation. All rights reserved.
SQL> connect /as sysdba
Connected to an idle instance.
SQL> startup
ORACLE instance started.
Total System Global Area 283291324 bytes
Fixed Size 102076 bytes
Variable Size 215015424 bytes
Database Buffers 67108864 bytes
Redo Buffers 1064960 bytes
Database mounted.
Database opened.
SQL> exit
这个问题我遇到过了,你肯定是在跑POST或者系统运行过程的时候强行使用setkeyswitch off或者poweroff all中断了.
SUN 800那边给过我一个解决办法,但是那个passwd是不是所有机器都通用我就不知道了!!
可以试一下,不行的话打800吧!!
Please kindly help to check about this:
SC> showcomponent -d a
Component Status Pending POST Description
--------- ------ ------- ---- -----------
/N0/SB0/P0 enabled - pass UltraSPARC-III+, 1050MHz,
8M ECache
/N0/SB0/P1 enabled - pass UltraSPARC-III+, 1050MHz,
8M ECache
/N0/SB0/P2 enabled - pass UltraSPARC-III+, 1050MHz,
8M ECache
/N0/SB0/P3 enabled - pass UltraSPARC-III+, 1050MHz,
8M ECache
/N0/SB0/P0/B0/L0 enabled - pass 512M DRAM
/N0/SB0/P0/B0/L2 enabled - pass 512M DRAM
/N0/SB0/P0/B1/L1 enabled - pass 512M DRAM
/N0/SB0/P0/B1/L3 enabled - pass 512M DRAM
/N0/SB0/P1/B0/L0 enabled - pass 512M DRAM
/N0/SB0/P1/B0/L2 enabled - pass 512M DRAM
/N0/SB0/P1/B1/L1 enabled - pass 512M DRAM
/N0/SB0/P1/B1/L3 enabled - pass 512M DRAM
/N0/SB0/P2/B0/L0 enabled - pass 512M DRAM
/N0/SB0/P2/B0/L2 enabled - pass 512M DRAM
/N0/SB0/P2/B1/L1 enabled - pass 512M DRAM
/N0/SB0/P2/B1/L3 enabled - pass 512M DRAM
/N0/SB0/P3/B0/L0 enabled - pass 512M DRAM
/N0/SB0/P3/B0/L2 enabled - pass 512M DRAM
/N0/SB0/P3/B1/L1 enabled - pass 512M DRAM
/N0/SB0/P3/B1/L3 enabled - pass 512M DRAM
/N0/SB2/P0 disabled - chs UltraSPARC-III+, 1050MHz,
8M ECache
/N0/SB2/P1 disabled - chs UltraSPARC-III+, 1050MHz,
8M ECache
/N0/SB2/P2 disabled - chs UltraSPARC-III+, 1050MHz,
8M ECache
/N0/SB2/P3 disabled - chs UltraSPARC-III+, 1050MHz,
8M ECache
/N0/SB2/P0/B0/L0 disabled - chs 512M DRAM
/N0/SB2/P0/B0/L2 disabled - chs
If the SB4 is disabled, please check the col named POST of SB4, if it is
chs, please help to check:
SC> showcomponent -v SB4
SC> showerrorbuffer
and then enter service mode:
SC> service
The use of service commands can be destructive and destabilize the system.
Use of service mode is authorized only by personnel trained in its use.
Enter Password: gGmmlHslPK
SC[service]> showchs -b
....
SC[service]> setchs -r "enable brd" -s ok -c sb4
and then please poweroff domain A by setkeyswitch off, after the domain A power off, please power it on and do max POST.
TRAP 3E
Cause
The Ultra system fails to boot with TRAP 3E. The system sometimes also displays bad magic number errors.
This error is caused by a bad super block on the boot disk. Which, in turn, could have been caused by a SCSI configuration problem.
Action
To fix:
1.
Check the SCSI bus for illegal configuration, bad cables, and duplicate SCSI addresses.
2.
Boot from CD-ROM as single user.
OK boot cdrom -sw
3.
Attempt to fsck(1M) boot disk. This could fail with a super block error.
# fsck /dev/rdsk/device
4.
Find the locations of alternate super blocks. BE SURE TO USE AN UPPERCASE -N. For example:
# newfs -N /dev/rdsk/c0t0d0s0
/dev/rdsk/c0t0d0s0: 2048960 sectors in 1348 cylinders of 19 tracks,
80 sectors 1000.5MB in 85 cyl groups (16 c/g, 11.88MB/g, 5696 i/g)
super-block backups (for fsck -F ufs -o b=#) at:
32, 24432, 48832, 73232, 97632, 122032, 146432, 170832, 195232, 219632,
244032, 268432, 292832, 317232, 341632, 366032, 390432, 414832, 439232,
463632, 488032, 512432, 536832, 561232, 585632, 610032, 634432, 658832,
683232, 707632, 732032, 756432, 778272, 802672, 827072, 851472, 875872,
900272, 924672, 949072, 973472, 997872, 1022272, 1290672, ...
5.
Using an alternate super block, run fsck(1M) on the disk. You might have to try more than one alternate super block to make this to work. Pick a couple from the beginning, the middle, and the end.
# fsck -o b=
6.
The boot block is probably bad too. Restore it while you are booted from the CD-ROM.
# /usr/sbin/installboot /usr/platform/architecture/lib/fs/ufs/bootblk
/dev/rdsk/c0t0d0s0
7.
Reboot the operating environment.
# reboot
usb键盘上用STOP+N是没用的,应该按照以下步骤来恢复NVRAM到缺省值:
1. After turning on the power to your system, wait until the front panel power indicator
LED begins to blink and you hear an audible beep.
2. Quickly press the front panel power switch twice (similar to the way you would
double-click a mouse).
# luxadm probe
# luxadm display Logical Path或# luxadm -e dump_map node_wwn
这么复杂啊,IPMP管理的公共网络,怎么说视心跳问题呢?心跳是由cluster去管理的,再安装的时候就会选择配置的啊!
给你一个我配的IPMP,可以参考一下:
1) ok setenv local-mac-address? true
2) 使用vi编辑器修改hosts文件,在/etc/hosts文件中,增加双机中另一节点的机器名和IP地址,以方便双机之间的访问。
在scha1的hosts文件中增加行:
#vi /etc/hosts
127.0.0.1 localhost
156.36.16.37 scha1 loghost
156.36.16.38 ce0test
156.36.16.39 ce2data
156.36.16.40 ce2test
156.36.16.41 scha2
156.36.16.45 ha-ora-lh
在scha2的hosts文件中增加行:
#vi /etc/hosts
127.0.0.1 localhost
156.36.16.41 scha2 loghost
156.36.16.42 ce0test
156.36.16.43 ce2data
156.36.16.44 ce2test
156.36.16.37 scha1
156.36.16.45 ha-ora-lh
3). 配置主备双机的IP多路径
对外公布的IP地址为:156.36.16.45
双机中每台主机有4块网卡ce0、ce1、ce2、ce3,规定ce0、ce2作数据网卡,互为主备用。两个网卡的逻辑组名要求一致,这里组名是abc。
当ce0链路出现问题时,ce1自动生成一个ce1:1的逻辑网卡,其IP地址为ce0的IP地址,当ce0恢复正常后,ce1:1 自动撤消,ce0仍然使用原来的IP地址。同理,当ce1链路出现问题时,ce0:1接替ce1工作,保证了网络使用的可靠性。(注意,该处的所有的IP地址要求处于同一网段,)
在主机scha1上使用vi编辑器编辑/etc/hostname.ce0和/etc/hostname.ce1两个文件,具体操作如下:
#vi /etc/hostname.ce0
scha1 group abc up \
addif ce0test -failover deprecated up
#vi /etc/hostname.ce2
ce2data group abc up \
addif ce2test -failover deprecated up
在主机scha2上使用vi编辑器编辑/etc/hostname.ce0和/etc/hostname.ce1两个文件,具体操作如下:
#vi /etc/hostname.ce0
scha2 group abc up \
addif ce0test -failover deprecated up
#vi /etc/hostname.ce2
ce2data group abc up \
addif ce2test -failover deprecated up
4) 在双机上使用vi编辑器修改/kernel/drv/scsi_vhci.conf文件,具体操作步骤如下:
#vi /kernel/drv/scsi_vhci.conf
将以下行:
mpxio-disable="yes"
修改为:
mpxio-disable="no"
5.修改/etc/system,添加ce网卡bug修正参数:
set ce:ce_reclaim_pending=1
set ce:ce_taskq_disable=1
IPMP从Solaris 8
(01/01)开始就已经成为Solaris的标准配置了,所以大家都可以用,对于这些新的feature,大家不要客气啊,不用白不用啊!
下面就用配置文件的方式来配置IPMP,假设如下:
multipathing test IP 10.170.1.81 测试地址1
multipathing test IP 10.170.1.82 测试地址1
Logical IP 10.170.1.80 逻辑地址,这是我们访问的地址
Group Name mswitch 这个名字,随便取的
Network Card ce0,ce1
1、首先更改hosts文件
#vi /etc/hosts
10.170.1.80 host1 loghost
10.170.1.81 host1-1
10.170.1.82 host1-2
2、然后创建文件hostname.ce0 , hostname.ce1
#vi /etc/hostname.ce0
host1 group mswitch up
addif host1-1 -failover deprecated up
#vi /etc/hostname.ce1
host1-2 group mswitch -failover deprecated up
3、重新启动计算机,就可以了
#reboot
4、after reboot , you can use ifconfig -a to view the configure
information.
#ifconfig -a
做做测试,看看会不会failover!
创建NAFO 组
# pnmset -c nafo-group -o create adapter [adapter ...]
-c nafo-group 对指定的NAFO 组执行配置子命令。必须将NAFO 组命
名为nafoN,其中N 是一个非负的整数,用来标识该组。对每个节点而言,组名都是本地名称。因而,同一NAFO 组名可在多个节点上使用。
-o create 创建新的NAFO 组。
adapter [ adapter ... ] 指定用作备份适配器的公共网络适配器。
2.向NAFO 组添加适配器
# pnmset -c nafo-group -o add adapter
-c nafo-group 指定要添加新适配器的NAFO 组。
-o add adapter 指定要添加到指定的NAFO 组的公共网络适配器。
3.如何删除NAFO 组
要想能够删除NAFO 组,逻辑主机资源组或共享地址资源组不得使用该组。
1. 成为包含要删除的NAFO 组的节点的超级用户。
2. 确定是否有任何逻辑主机资源或共享地址资源正在使用NAFO 组。
# scrgadm -pv
您还可以使用scrgadm-pvv(带有两个v 标志)来确定那些正在使用将要删除的NAFO 组的资源。
3. 切换使用该NAFO 组的逻辑主机资源组和共享地址资源。
# scswitch -z -g resource-group -h nodelist
-z -g resource-group 切换指定的资源组。
-h nodelist 指定要将资源组切换到的节点的名称。
4. 删除NAFO 组。
# pnmset -c nafo-group -o delete
-c nafo-group 指定要删除的NAFO 组。
-o delete 删除NAFO 组。
5. 检验该NAFO 组的状态。
列表中不应出现已删除的NAFO 组。
# pnmstat -l
示例— 切换NAFO 组的活动适配器
以下示例说明了将活动适配器从qfe0 切换到qfe1 的过程。
# pnmstat -l
group adapters status fo_time act_adp
nafo0 qfe0:qfe1 OK NEVER qfe0
# pnmset -c nafo0 -o switch qfe1
# mv /etc/hostname.qfe0 /etc/hostname.qfe1
# pnmstat -l
group adapters status fo_time act_adp
nafo0 qfe0:qfe1 OK 11 qfe
查找NAFO 组的活动适配器
# pnmptor nafo0
查找适配器的NAFO 组
# pnmrtop qfe5
4.如何更改公共网络管理可调参数
/etc/cluster/pnmparams
inactive_time 当前活动适配器的包计数器在连续进行两次探测之间的秒数。缺省值为5。
ping_timeout ALL_HOST_MULTICAST 和子网广播ping 的超时值(以秒为单位)。缺省值为4。
repeat_test 在声明活动适配器有故障并触发故障转移前执行ping 序列的次数。缺省值为3。
slow_network 每次从执行ping 序列之后到检查包计数器以查看是否有任何更改之间所等待的秒数。缺省值为2。
warmup_time 从将故障转移到备份适配器后到恢复故障监视前所等待的秒数。此设置为慢速驱动程序或缓慢的端口初始化过程保留了额外的时间。缺省值为0。
一 文件方式:
IP配置方式
hosts 文件
#
# Internet host table
#
127.0.0.1 localhost
10.0.2.1 WEB1-test1
10.0.2.2 WEB1-test2
10.0.2.3 WEB1-data1 WEB1-data1.mbox.monternet.com WEB1 loghost
10.0.2.4 WEB1-data2 WEB1-data2.mbox.monternet.com
root@WEB1 # ls host*
hostname.eri0 hostname.hme0 hosts
hostname.eri0 文件
WEB1-data1 netmask + broadcast + group web1 failover up \
addif WEB1-test1 netmask + broadcast + deprecated -failover up
hostname.hme0文件
WEB1-data2 netmask + broadcast + group web1 failover up \
addif WEB1-test2 netmask + broadcast + deprecated -failover up
IP配置方式
hostname.qfe1 文件
db1data netmask + broadcast + group dbt1 failover up \
addif db1test1 netmask + broadcast + deprecated -failover up
hostname.qfe2 文件
db1test2 netmask + broadcast + group dbt1 deprecated -failover standby up
二 命令行方式:
IP配置方式
# ifconfig qfe1 plumb 192.2.0.6 netmask + broadcast + group dbt2 failover up \
addif 192.2.0.4 netmask + broadcast + deprecated -failover up
# ifconfig qfe2 plumb 192.2.0.7 netmask + broadcast + group dbt2 failover up \
addif 192.2.0.5 netmask + broadcast + deprecated -failover up
其中 192.2.0.4 192.2.0.5 为测试IP 192.2.0.6 192.2.0.7 为浮动IP
IP配置方式
# ifconfig qfe1 plumb 192.2.0.6 netmask + broadcast + group dbt2 failover up \
addif 192.2.0.4 netmask + broadcast + deprecated -failover up
# ifconfig qfe2 plumb 192.2.0.5 netmask + broadcast + group dbt2 deprecated -failover standby up
其中 192.2.0.4 192.2.0.5 为测试IP 192.2.0.6 为浮动IP
DETACHED REPLAY修复方法:
enters this state during a log replay.
if the log replay dies with signal 10 or 11
run these commands.
vxplex -f -o rm dis
vxassist -g
vxvol -g
1、boot -s
2、fcsk root file system
3、rm /dev/null
rm /dev/null
4、re-create /dev/null link
ln -s /devices/pseudo/mm@0:null /dev/null
5、reboot system
用metaset建立的磁盘怎么删除
#metaset -s dg-schost -d -h sol1 sol2
metaset: sysb: setname "dg-schost": no such set
#metaset -s
metaset: sol1: setname "dg-schost-3": no such set
我注意到sysb和sol1两个主机名
如果是你更改过主机名了,按顺序执行下面几条命令看看能不能行:
#metaimport -r
#metaimport -r -v
#metaimport -n -s
# metaset -s
# mount
Maintaining File Systems