刘佳的技术日志、博客

/usr/bin/java ---> /etc/alternatives/java ---->　某个版本的JDK的/bin/java  

首先从http://java.sun.com/products/archive/选择相应的版本，我们这里1.4.2的最高版本1.4.2_19为例，包名如下：
j2re-1_4_2_19-linux-i586-rpm.bin
j2sdk-1_4_2_19-linux-i586-rpm.bin

安装方法：
A：以root身份执行以下操作:
首先将文件赋予可以执行权限
chmod 700 j2re-1_4_2_19-linux-i586-rpm.bin  j2sdk-1_4_2_19-linux-i586-rpm.bin
B：执行bin软件包，读完sun的License声明，敲yes，就可以将其解压为rpm文件，对了，这个有自行校验功能。可以检查出解压缩的rpm的完整性。
./j2re-1_4_2_19-linux-i586-rpm.bin
./j2sdk-1_4_2_19-linux-i586-rpm.bin
C：在执行目录下面下，产生rpm文件如下：
j2re-1_4_2_19-linux-i586-rpm j2sdk-1_4_2_19-linux-i586-rpm
D：安装rpm包,这个简单吧。
rpm -ivh j2re-1_4_2_19-linux-i586-rpm j2sdk-1_4_2_19-linux-i586-rpm

auto
/usr/bin/java
jre
/usr/lib/jvm/jre
jre_exports
/usr/lib/jvm-exports/jre
keytool
/usr/bin/keytool
rmiregistry
/usr/bin/rmiregistry

/usr/lib/jvm/jre-1.4.2-gcj/bin/java
1420
/usr/lib/jvm/jre-1.4.2-gcj
/usr/lib/jvm-exports/jre-1.4.2-gcj
/usr/lib/jvm/jre-1.4.2-gcj/bin/keytool
/usr/lib/jvm/jre-1.4.2-gcj/bin/rmiregistry
+++++++++++++++使用时，请去掉分割线+++++++++++++++++
/usr/java/j2re1.4.2_19/bin/java
1500
/usr/java/j2re1.4.2_19
/usr/java/j2re1.4.2_19/lib
/usr/java/j2re1.4.2_19/bin/keytool
/usr/java/j2re1.4.2_19/bin/rmiregistry
+++++++++++++++使用时，请去掉分割线+++++++++++++++++
/usr/java/j2sdk1.4.2_19/bin/java
1600
/usr/java/j2sdk1.4.2_19/jre
/usr/java/j2sdk1.4.2_19/lib
/usr/java/j2sdk1.4.2_19/bin/keytool
/usr/java/j2sdk1.4.2_19/bin/rmiregistry
+++++++++++++++使用时，请去掉分割线+++++++++++++++++

[root@kook alternatives]# alternatives --config java

共有 3 个程序提供“java”。

  选择    命令
-----------------------------------------------
+ 1           /usr/lib/jvm/jre-1.4.2-gcj/bin/java
   2           /usr/java/j2re1.4.2_19/bin/java
*  3           /usr/java/j2sdk1.4.2_19/bin/java

按 Enter 来保存当前选择[+]，或键入选择号码：2

最后查看一下版本：
[root@kook ~]# java -version
java version "1.4.2_19"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2_19-b04)
Java HotSpot(TM) Client VM (build 1.4.2_19-b04, mixed mode)

1:修改/etc/modprobe.conf，添加如下：
[root@ha01 network-scripts]# tail -2 /etc/modprobe.conf
alias bond0 bonding
options bond0 miimon=100 mode=1

关于mode见本文最下面。

2:在/etc/sysconfig/network-scripts/下创建ifcfg-bond0，ifcfg-eth1,ifcfg-eth2
[root@ha01 network-scripts]# cat ifcfg-bond0
DEVICE=bond0
BOOTPROTO=none
ONBOOT=yes
NETWORK=192.168.137.0
NETMASK=255.255.255.0
IPADDR=192.168.137.9
USERCTL=no
GATEWAY=
TYPE=Ethernet
[root@ha01 network-scripts]# cat ifcfg-eth1
DEVICE=eth1
BOOTPROTO=none
ONBOOT=yes
MASTER=bond0
SLAVE=yes
USERCTL=yes
[root@ha01 network-scripts]# cat ifcfg-eth2
DEVICE=eth2
BOOTPROTO=none
ONBOOT=yes
MASTER=bond0
SLAVE=yes
USERCTL=yes

mode

        Specifies one of the bonding policies. The default is
        balance-rr (round robin).  Possible values are:

        balance-rr or 0

                Round-robin policy: Transmit packets in sequential
                order from the first available slave through the
                last.  This mode provides load balancing and fault
                tolerance.
        active-backup or 1

                Active-backup policy: Only one slave in the bond is
                active.  A different slave becomes active if, and only
                if, the active slave fails.  The bond's MAC address is
                externally visible on only one port (network adapter)
                to avoid confusing the switch.

                In bonding version 2.6.2 or later, when a failover
                occurs in active-backup mode, bonding will issue one
                or more gratuitous ARPs on the newly active slave.
                One gratuitous ARP is issued for the bonding master
                interface and each VLAN interfaces configured above
                it, provided that the interface has at least one IP
                address configured.  Gratuitous ARPs issued for VLAN
                interfaces are tagged with the appropriate VLAN id.

                This mode provides fault tolerance.  The primary
                option, documented below, affects the behavior of this
                mode.

        balance-xor or 2

                XOR policy: Transmit based on the selected transmit
                hash policy.  The default policy is a simple [(source
                MAC address XOR'd with destination MAC address) modulo
                slave count].  Alternate transmit policies may be
                selected via the xmit_hash_policy option, described
                below.

                This mode provides load balancing and fault tolerance.

        broadcast or 3

                Broadcast policy: transmits everything on all slave
                interfaces.  This mode provides fault tolerance.

        802.3ad or 4

                IEEE 802.3ad Dynamic link aggregation.  Creates
                aggregation groups that share the same speed and
                duplex settings.  Utilizes all slaves in the active
                aggregator according to the 802.3ad specification.

                Slave selection for outgoing traffic is done according
                to the transmit hash policy, which may be changed from
                the default simple XOR policy via the xmit_hash_policy
                option, documented below.  Note that not all transmit
                policies may be 802.3ad compliant, particularly in
                regards to the packet mis-ordering requirements of
                section 43.2.4 of the 802.3ad standard.  Differing
                peer implementations will have varying tolerances for
                noncompliance.

                Prerequisites:

                1. Ethtool support in the base drivers for retrieving
                the speed and duplex of each slave.

                2. A switch that supports IEEE 802.3ad Dynamic link
                aggregation.

                Most switches will require some type of configuration
                to enable 802.3ad mode.

        balance-tlb or 5

                Adaptive transmit load balancing: channel bonding that
                does not require any special switch support.  The
                outgoing traffic is distributed according to the
                current load (computed relative to the speed) on each
                slave.  Incoming traffic is received by the current
                slave.  If the receiving slave fails, another slave
                takes over the MAC address of the failed receiving
                slave.

                Prerequisite:

                Ethtool support in the base drivers for retrieving the
                speed of each slave.

        balance-alb or 6

                Adaptive load balancing: includes balance-tlb plus
                receive load balancing (rlb) for IPV4 traffic, and
                does not require any special switch support.  The
                receive load balancing is achieved by ARP negotiation.
                The bonding driver intercepts the ARP Replies sent by
                the local system on their way out and overwrites the
                source hardware address with the unique hardware
                address of one of the slaves in the bond such that
                different peers use different hardware addresses for
                the server.

                Receive traffic from connections created by the server
                is also balanced.  When the local system sends an ARP
                Request the bonding driver copies and saves the peer's
                IP information from the ARP packet.  When the ARP
                Reply arrives from the peer, its hardware address is
                retrieved and the bonding driver initiates an ARP
                reply to this peer assigning it to one of the slaves
                in the bond.  A problematic outcome of using ARP
                negotiation for balancing is that each time that an
                ARP request is broadcast it uses the hardware address
                of the bond.  Hence, peers learn the hardware address
                of the bond and the balancing of receive traffic
                collapses to the current slave.  This is handled by
                sending updates (ARP Replies) to all the peers with
                their individually assigned hardware address such that
                the traffic is redistributed.  Receive traffic is also
                redistributed when a new slave is added to the bond
                and when an inactive slave is re-activated.  The
                receive load is distributed sequentially (round robin)
                among the group of highest speed slaves in the bond.

                When a link is reconnected or a new slave joins the
                bond the receive traffic is redistributed among all
                active slaves in the bond by initiating ARP Replies
                with the selected MAC address to each of the
                clients. The updelay parameter (detailed below) must
                be set to a value equal or greater than the switch's
                forwarding delay so that the ARP Replies sent to the
                peers will not be blocked by the switch.

                Prerequisites:

                1. Ethtool support in the base drivers for retrieving
                the speed of each slave.

                2. Base driver support for setting the hardware
                address of a device while it is open.  This is
                required so that there will always be one slave in the
                team using the bond hardware address (the
                curr_active_slave) while having a unique hardware
                address for each slave in the bond.  If the
                curr_active_slave fails its hardware address is
                swapped with the new curr_active_slave that was
                chosen.

#!/bin/bash
# chkconfig: 35 99 05
# description: Test chkconfig
# file: /etc/init.d/foo
env > /tmp/ENVLOG

[root@kook init.d]# chkconfig --add foo
foo 服务不支持 chkconfig
[root@kook init.d]# chkconfig --add foo
service foo does not support chkconfig

35:就是在哪种runlevel下启动。
99:在rc3.d和rc5.d下产生S99foo
05:在rc0.d、rc1.d、rc2.d、rc4.d、rc6.d下产生K05foo

S代表Start
K代表Kill

<Directory "/var/www/html">
   Order allow,deny
   Allow from all
</Directory>

<Directory "/var/www/html">
   Order allow,deny
</Directory>
未定义Allow，结果全部deny

<Directory "/var/www/html">
   Order deny,allow
</Directory>
未定义Deny,结果全部allow

<Directory "/var/www/html">
   Order allow,deny
   Deny from .kook.com
</Directory>
全部Deny，定义的那个kook.com无意义。

<Directory "/var/www/html">
   Order deny,allow
   Allow from .kook.com
</Directory>
全部Allow，定义的那个kook.com无意义。



<Directory "/var/www/html">
   Order deny,allow
   Deny from .kook.com
</Directory>
定义生效，除了kook.com其他的域都可以访问。

<Directory "/var/www/html">
   Order allow,deny
   Allow from .kook.com
</Directory>
定义生效，除了kook.com其他的域都不可以访问。