killer911

客户的一台44P-270，前几天刚给用备份带恢复的系统。最近客户把270换了个地方，机器的系统就起不来了，停在4BA00830，查了service guide，说是引导失败。

处理过程：

光盘引导到维护模式,重建了下引导块，系统就能起来了

bosboot -ad /dev/hdisk0

最近装了一台P55A，一台P52A，都挺顺利

P55A配置是4*1.9G/8G/2*73G/光纤卡1块

P52A配置是2*1.9G/4G/2*146G

1、不同机型系统迁移，在做rootvg镜像的时候宕机，搞得偶很郁闷

2、H85升级内存的时候，有2G（4*512M）内存没认出来，重插了一遍也不行

明天晚上要做做一个系统的迁移，道理说出来很简单，就是P650上的rootvg和datavg全部迁移到一台M85上。

P650的配置为：

4*1.45G CPU

8G内存

2*73.2G内置硬盘

磁盘阵列是DS4300，4块73G硬盘做RAID5

M85配置为：

6*750M CPU

8G内存

2*36.4G内置硬盘

磁盘阵列是7133，5*36.4G硬盘做RAID5

具体操作步骤为：

1、mkszfile,编辑image.data把镜像信息去掉

mksysb /dev/rmt0用内置磁带机做系统备份

2、装3580磁带机驱动

installp -axXd /tmp/Atape.8.2.5.0.bin Atape.driver

3、smitty savevg备份datavg内容

4、恢复p650 mksysb到M85服务器上，记得选择shrink file system选项

5、恢复P650 savevg到M85服务器上，记得选择shrink file system选项

6、配置系统，smitty hostname配置主机名

smitty chinet 配置网卡IP地址

smitty chjfs 改必要的文件系统大小

smitty chaio启动AIO

chmod改变裸设备属性

7、内核从64位降为32位

ln -sf /usr/lib/boot/unix_mp /unix

ln -sf /usr/lib/boot/unix_mp /usr/lib/boot/unix

bosboot -ad /dev/ipldevices

shutdown -Fr

在下列条件下，DBWn写数据文件：

增量或是正常checkpoint

dirty buffer的数量超过阙值

进程扫描free buffers ，扫描了一定数量以后但并没有找到free buffers

超时

RAC环境下的PING请求

表空间设置为只读模式时

drop或是truncate一个表

ALTER TABLESPACE tablespace_name BEGIN BACKUP;

昨天晚上客户处理电源完成以后，重新启动主机的时候发现两台IBM的机器CDE起不来，一台M85显示启动LEGATO后台DAEMON的时候挂起，telnet 进来以后把nsr进程kill以后还是不好使，检查了一下/etc/hosts文件发现这台机器以前的IP还在/etc/hosts文件中，注释这条以后，kill掉所有的dt进程，然后/usr/dt/bin/dtlogin -daemon以后就可以在控制台用CDE界面登陆了。

另外一台650严重一些，CDE，命令行都登陆不了，正准备拿盘进维护模式的时候，有人碰了一下CDE界面居然出来了个命令行窗口，高兴啊，进去一看，靠，/,/var 区都满了，用smitty chjfs扩了这两个分区以后，一切正常。

84分通过HP-UX CSA认证。

主要遇到了以下几个题目：

1、What are the advantages of Online JFS?select three

a. fast resync

b. fast recovery

c. online resizing

d. secure file deletion

e. automatic resizing

f. dynamic inodes creation



2、What are possible uses of the lanadmin command? Select three

a. enable IP multiplexing

b. resets the network interface card

c. shows IP address for network interface card

d. shows basic I/O error for the network interface card

e. shows operational status of the network interface card



3、What does /etc/default set? Select three

a. password policies

b. default swap space

c. default file system type

d. disk space usage limits

e. system time zone

f. some security settings in the directory




4、Which file is used to automatically check and mount file systems?

a. /etc/mnttab

b. /var/adm/fstab

c. /etc/sbtab

d. /etc/fstab


5、Which files can be used to enable a login without a password? Select three

a. .rhost

b. /etc/allow

c. host.allow

d. host.equiv

e. /etc/password

f. /etc/host.aliases



6、During a run-state change from run level 1 to run level 2, which of the following scripts would be run last?

a./etc/rc2.d/k900nfs.server

b./etc/rc2.d/s565snmpTrpDst

c./etc/rc2.d/S565snmpMib2

d./etc/rc2.d/S100nfs.server



7、Which combination of NFS mount options would most likely provide the best chance of avoiding client hangs?

a. retry=2, hard

b. retry=6, hard, intr

c. retry=3, soft, nointr

d. retry=4, soft

1.vi /etc/rc.config.d/nfsconf

修改参数AUTOFS=1

2.重新启动nfs 服务

/sbin/init.d/nfs.server stop

/sbin/init.d/nfs.client stop

/sbin/init.d/nfs.client start

/sbin/init.d/nfs.server start

3.手动启动autofs 服务

/usr/lib/netsvc/fs/autofs/automountd

/usr/sbin/automount

4.修改/etc/auto_master

vi /etc/auto_master

/net -svr1 -nosuid,nobrowse
/- /etc/auto_direct

vi /etc/auto_direct

/suntest -ro e450:/opt

5. ll /net/serv1

ll /suntest

就可以自动挂载相应的资源。

1、划分磁盘分区：
c0t0d0s0 c0t1d0s0 58.42G
c0t0d0s1 c0t1d0s1 4G
c0t0d0s7 c0t1d0s7 5.86G
c0t0d0s3 c0t1d0s3 100M

prtvtoc /dev/rdsk/c0t0d0s2 |fmthard -s - /dev/rdsk/c0t1d0s2
2、根分区镜像：
metadb -f -a -c 3 c0t0d0s3 c0t1d0s3

metainit -f d11 1 1 c0t0d0s0
metainit d12 1 1 c0t1d0s0
metainit d10 -m d11
更新/etc/fstab文件：
cp /etc/vfstab /etc/vfstab.bak

#cp /etc/system /etc/system.bak

#metaroot d10

3、
metainit -f d21 1 1 c0t0d0s1
metainit d22 1 1 c0t1d0s1
metainit d20 -m d21

metainit -f d31 1 1 c0t0d0s1
metainit d32 1 1 c0t1d0s1
metainit d30 -m d31

编辑/etc/vfstab
metainit hsp001

lockfs -fa
init 6

3、挂载第2个sub-mirror
metattach d10 d12
metattach d20 d22
netattach d30 d32
4、生成启动块
installboot /usr/platform/`uname -i`/lib/fs/ufs/bootblk /dev/rdsk/c0t1d0s0

换盘步骤：

1.清除metadb
metadb -d c1t0d0s3 删除坏盘的metadb
luxadm remove_device /dev/rdsk/c1t0d0s2（可跳过，直接拔）
2.换下坏的硬盘
3.luxadm insert_device
4.devfsadm -v后查看一下盘的状况
5.prtvtoc /dev/dsk/c0t1d0s2 | fmthard -s - /dev/rdsk/c0t0d0s2
换上新盘同步分区后后执行
6.metadb -afc3 c0t0d0s3建立新的metadb
7.维护两个子镜像
metarepalce –e d10 c0t0d0s0
metarepalce –e d20 c1t0d0s1
metarepalce –e d30 c0t0d0s7

（可能是OK状态 就不需要维护）
8.通过metastat|grep %

1.
pvcreate -f -B /dev/rdsk/cXtYdZ
vgextend /dev/vg00 /dev/dsk/cXtYdZ
mkboot －l /dev/rdsk/cXtYdZ

lvextend -m 1 /dev/vg00/lvol 1 /dev/dsk/cXtYdZ
lvextend -m 1 /dev/vg00/lvol 2 /dev/dsk/cXtYdZ
lvextend -m 1 /dev/vg00/lvol 3 /dev/dsk/cXtYdZ

mkboot -a "hpux -lq(;0)/stand/vmunix" /dev/rdsk/cXtYdZ

2.
lvlnboot -b /dev/vg00/lvol 1
lvlnboot -s /dev/vg00/lvol 2
lvlnboot -r /dev/vg00/lvol 3
lvlnboot -d /dev/vg00/lvol 2

lvlnboot –Rv( 确保BDRA已经update，这个命令是确认写入的)。直接执行这个命令就不用分个指定boot、swap、root、dump。
Lvlnboot –v (to check it)
b－boot、s－swap、r－root、d－dump
3.修改boot ALT路径为mirror disk
setboot -a <HW-PATCH>
# setboot -p 0/0/0/0.0.0 use your original boot disk’s hw path here
# setboot -a 0/0/0/0.1.0 use your mirrored disk’s hw path here

# passwd user1 reset a user’s password
# passwd -d user1 set a null password
# passwd -f user1 force a password change at next
# passwd -l user1 lock or disable an account
# passwd -n 7 -x 70 –w 7 user1 configure password aging
# passwd -n 7 -x 0 user1 prevent a user from changing passwords
# passwd -s user1 check password aging settings

c0t1d0s0 10g c0t3d0s0
c0t1d0s1 6g c0t3d0s1
c0t1d0s6 0.9g c0t3d0s6
metadb -f -a c 3 c0t1d0s6 c0t3d0s6
metainit -f d11 1 1 c0t1d0s0
metainit d12 1 1 c0t3d0s0
metainit -f d21 1 1 c0t1d0s1
metainit d22 1 1 c0t3d0s1
metainit d10 -m d11
metainit d20 -m d21
metattach d10 d12
metattach d20 d22
newfs /dev/md/rdsk/d10
newfs /dev/md/rdsk/d20
/dev/md/dsk/d10 /dev/md/rdsk/d10 /data ufs 2 yes -
/dev/md/dsk/d20 /dev/md/rdsk/d20 /data1 ufs 2 yes -

根据该文档升级3534-F08光纤交换机

firmware: Firmware Download Instruction Windows 95, NT and Windows 2000 F08/F16 June 24, 2002 Page 1 Before beginning the firmware download, make sure to have IP address, user name of the host machine and IP address of the switch. Make sure that the current switch configuration data has been saved, it must be available for reload if needed. If you are not sure the current switch configuration data is available, see Telnet commands "configUpload" and "configDownload" in the Command Reference for OS 3.0 and 4.0 for procedure to upload and download the switch configuration file.

1. Download New Firmware: 2109 Model F16 or 3534 Model F08

2. Download and save the latest firmware and rsh.zip into a directory. Un-zip all files into the same directory (ie. c:firmware).

3. Have the customer isolate this switch and ensure that all I/O operations through the switch have stopped.

4. Open a DOS windows, go to firmware directory (cd firmware), enter: rshd [to execute the RSH daemon] and keep the windows opened. Example: C:firmware>rshd

5. From a second DOS windows start a Telnet session to the switch over Ethernet LAN-attached that has connectivity to the switch. Format is: c:>telnet [switch ipAddress]

6. Login using the admin userid: login: admin

7. Password: password

8. Type the following command to download firmware: firmwaredownload "ipAddress"], ["userName"], ["fileName"] userName is PC's user name ipAddress is PC's ipAddress. Example: switch:admin> firmwareDownload “9.111.23.1", "John", "c:/firmware/v3.0.2f" 89640+6024+130948, csum 6132 DO NOT power cycle the switch! writing flash 0 ............................... writing flash 1 ............................... download complete switch:admin>

9. The switch will need to be rebooted following the firmware download. Firmware Download Instruction Windows 95, NT and Windows 2000 F08/F16 June 24, 2002 Page 2 switch:admin> fastboot NOTE: When reboot the switch the Telnet session will be disconnected. 10. Do step 5, 6 and 7 to Telnet back to the switch. 11. Verify the new level of firmware has been installed and everything is functional. switch:admin> version

1、到客户现场后，诊断两台270服务器故障。其中一台三块内置盘均报错（P270-1），hdisk0属于rootvg，hdisk1,hdisk2属于datavg,另外一台（P270-2）两块硬盘报错，其中hdisk0 属于rootvg，hdisk1,hdisk2属于datavg，报错的为hdisk1,hdisk2；

注：两台服务器均没有内置磁带机

2、为了保证应用的正常运行，另外搭建一个应用环境。安装B80服务器操作系统，建相应的文件系统、扩空间、建用户，远程拷贝运行环境内容到B80中；

rcp -r * b80:/data

3、加外置磁带机，备份P270-1关键文件系统

tar -cvf /dev/rmt0.1 /data

3、停p270-1，拔掉hdisk1,hdisk2，安装一块新盘，重新启动系统

，把新盘加到rootvg 中，mirrorvg rootvg 然后lsvg -l rootvg 发现/data文件系统状态为stale，删除该文件系统，重新同步vg，均正常，改变启动顺序从hdisk1启动系统；

4、P270-1从hdisk1启动成功后，unmirrorvg rootvg hdisk0

reducevg rootvg hdisk0

shutdown -F

5、更换P270-1内置硬盘hdisk0,重新启动系统

mirrrorvg rootvg hdisk0

6、为P270-1做mksysb

P270-2恢复方式大致与P270-1相同

一、升级目标

H85主机，备机每台原有1个450MHZ CPU，1G内存，内置硬盘为2*9G要把每台机器的配置升级为2个450MHZ CPU，4G内存7133增加4块硬盘，做RAID10

二、升级过程

1，系统备份smitty mksysb，没有什么好讲的，大约要1个小时

2，关闭备机

3,升级备机CPU(1颗450M升级到2颗),内存(由1G升级到4G),内置硬盘(由9G*2升级到36G*2)

4,备机开机,由磁带启动恢复操作系统.需要1个半小时完成

5,将数据卷组挂到备机上,smitty savevg备份数据卷组内容

6,升级主机CPU(1颗450M升级到2颗),内存(由1G升级到4G),内置硬盘(由9G*2升级到36G*2)

7,恢复主机系统

8,主机开机,做HA切换

9,双机升级完成