Configuring Linux for the Oracle 10g VIP or private interconnect using bonding driver [ID 298891.1]
2011年07月19日
[b]Configuring Linux for the Oracle 10g VIP or private interconnect using bonding driver [ID 298891.1][/b] 修改时间 06-AUG-2010 类型 BULLETIN 状态 ARCHIVED
PURPOSE-------In order to avoid the public LAN from being a single point of failure, Oracle highly recommends configuring a redundant set of public network interface cards (NIC's) on each cluster node. Network redundancy can be achieved On Linux platforms using NIC Teaming (configuring multiple interfaces to a team or using the Linux kernel bonding module).This note will go over the two possible choices for achieving redundancy on Linux.As inter-node IP address failover is achieved by using the Oracle managed VIP, 3rd party clusterware based inter-node IP address failover technologies should not be configured on the same set of NIC's that are used by the Oracle VIP. Only intra-node IP address failover functionalities should be used in conjunction with the Oracle VIP.SCOPE & APPLICATION-------------------This article is intended for experienced DBAs and Support Engineers.1. NIC TEAMING BY CONFIGURING MULTIPLE INTERFACES TO A TEAM-----------------------------------------------------------Various hardware vendors provide network interface drivers and utilities to achieve NIC teaming. Please consult your hardware vendor for details on how to configure your system for NIC teaming2. NIC TEAMING USING THE LINUX KERNEL BONDING MODULE----------------------------------------------------The Linux kernel includes a bonding module that can be used to achieve software level NIC teaming. The kernel bonding module can be used to team multiple physical interfaces to a single logical interface, which is used to achieve fault tolerance and load balancing. The bonding driver is available as part of the Linux kernel version 2.4.12 or newer versions. Since the bonding module is delivered as part of the Linux kernel, it can be configured independently from the interface driver vendor (different interfaces can constitute a single logical interface).The configuration steps are different among Linux distributions. This note will go over the steps required to configure the bonding module in RedHat Enterprise Linux 3.0.In the following example, two physical interfaces (eth0 and eth1) will be bonded together to a single logical interface (bond0), and the VIP will run on top of the single logical interface.A sample network configuration is as follows: Default Gateway: 192.168.1.254 Netmask: 255.255.255.0 Interface configuration before bonding: eth0: IP Address 192.168.1.1 eth1: IP Address 192.168.1.2After configuring the bonding driver, a logical interface named bondX (where X is a number higher than zero) representing the team of interfaces. Interface configuration after bonding: bond0: IP Address 192.168.1.10Note: Before continuing, be sure the bonding module is mounted. The command "modprobe bonding" can be used to mount the module.2-1 CONFIGURING THE BONDING DRIVER----------------------------------Since the bonding driver is delivered as a kernel module, the following lines need to be added to /etc/modules.conf as root.alias bond0 bondingoptions bond0 miimon=100For details on the "options" parameter, please refer to the documents referred to in section 2.8. In the above configuration, the MII link monitoring interval is set to 100ms. MII is used to monitor the interface link status, and this is a typical configuration for mission critical systems that require fast failure detection.自己备注：实际上还有一个mode选项。mode=0 表示负载均衡方式，两块网卡都工作。mode=1 表示冗余方式，网卡只有一个工作，一个出问题启用另外的。如options bond0 miimon=100 mode=0 Note: MII is an abbreviation of "Media Independent Interface". Many popular fast ethernet adapters use MII to autonegotiate the link speed and duplex mode.By default, the bonding driver will transmit outgoing packets in a round-robin fashion using each "slave" interface. The above example uses this default behavior. For details on changing this behavior, please also refer to the documents referred to in section 2.8.If you want to use muitlple bonding interfaces you should modify /etc/modules.conf like below example.alias bond0 bondingalias bond1 bondingoptions bond0 miimon=100 max_bonds=2options bond1 miimon=100 max_bonds=2(In this example we have 2 bonding interfaces.)The "max_bonds" parameter defines how many bonding interfaces we aregoing to have.For details on the "max_bonds" parameter, please refer to the documentsreferred to in section 2.8. 2-2. CONFIGURING THE bond0 INTERFACE------------------------------------On RHEL 3.0, network interface parameters are configured in configuration files named "ifcfg-", found in the /etc/sysconfig/network-scripts directory. In order to enable the bonding driver, a configuration file "ifcfg-bond0" needs to be created with appropriate parameters. As root, create the file "/etc/sysconfig/network-scripts/ifcfg-bond0" as shown below.DEVICE=bond0IPADDR=192.168.1.10NETMASK=255.255.255.0NETWORK=192.168.1.0BROADCAST=192.168.1.255ONBOOT=yesBOOTPROTO=noneUSERCTL=no (Please change the IP address, Netmask, Broadcast to match your network configuration)2-3. CHANGING THE CONFIGURATION FOR THE EXISTING INTERFACES-----------------------------------------------------------As root, please change the configuration file "/etc/sysconfig/network-scripts/ifcfg-eth0" as shown below:DEVICE=eth0USERCTL=noONBOOT=yesMASTER=bond0SLAVE=yesBOOTPROTO=nonePlease also change the configuration file "/etc/sysconfig/network-scripts/ifcfg-eth1" as shown below:DEVICE=eth1USERCTL=noONBOOT=yesMASTER=bond0SLAVE=yesBOOTPROTO=noneThese steps are necessary to associate the bond0 interface to its slave interfaces (eth0 and eth1).2-4. RESTART THE NETWORK------------------------Execute the following commands as root to reflect the changes.# service network stop# service network startIf the configuration is correct, the above commands should both return [ OK ].2-5. CONFIRMING THE NEW CONFIGURATION-------------------------------------The following messages should appear in your syslog (/var/log/messages).Jan 28 16:00:09 rac01 kernel: bonding: MII link monitoring set to 100 msJan 28 16:00:09 rac01 kernel: ip_tables: (C) 2000-2002 Netfilter core teamJan 28 16:00:11 rac01 ifup: Enslaving eth0 to bond0Jan 28 16:00:11 rac01 kernel: bonding: bond0: enslaving eth0 as a backup interface with a down link.Jan 28 16:00:11 rac01 kernel: e1000: eth0: e1000_watchdog: NIC Link is Up 1000 Mbps Full DuplexJan 28 16:00:11 rac01 kernel: bonding: bond0: link status definitely up for interface eth0.Jan 28 16:00:11 rac01 kernel: bonding: bond0: making interface eth0 the new active one.Jan 28 16:00:11 rac01 ifup: ENslaving eth1 to bond0Jan 28 16:00:11 rac01 kernel: bonding: bond0: enslaving eth1 as a backup interface with a down link.Jan 28 16:00:11 rac01 kernel: e1000: eth1: e1000_watchdog: NIC Link is Up 1000 Mbps Full DuplexJan 28 16:00:11 rac01 kernel: bonding: bond0: link status definitely up for interface eth1.Jan 28 16:00:11 rac01 network: Bringing up interface bond0: succeededThe "ifconfig -a" command should return the following output.bond0 Link encap:Ethernet HWaddr 00:0C:29:DC:83:E8 inet addr:192.168.1.10 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1 RX packets:27 errors:0 dropped:0 overruns:0 frame:0 TX packets:1 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:3462 (3.3 Kb) TX bytes:42 (42.0 b)eth0 Link encap:Ethernet HWaddr 00:0C:29:DC:83:E8 inet addr:192.168.1.10 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1 RX packets:13 errors:0 dropped:0 overruns:0 frame:0 TX packets:1 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:1701 (1.6 Kb) TX bytes:42 (42.0 b) Interrupt:10 Base address:0x1424 eth1 Link encap:Ethernet HWaddr 00:0C:29:DC:83:E8 inet addr:192.168.1.10 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1 RX packets:14 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:1761 (1.7 Kb) TX bytes:0 (0.0 b) Interrupt:11 Base address:0x14a4 (Note that other interfaces will also appear in a typical RAC installation)2-6. CONSIDERATIONS FOR CRS INSTALLATION -----------------------------------------During CRS installation, choose "Public" for the "bond0" interface in the "Specify Network Interface Usage" OUI page. If the "eth0" and "eth1" interfaces appear in OUI, then make sure to choose "Do not use" for their types.2-6a. IF YOU WANT TO CONFIGURE BONDING DEVICES AFTER INSTALLATION OF CRS.-----------------------------------------You can change your interconnect/public interface configuration using oifcfg command. Please refer Note 283684.12-7. VIPCA CONFIGURATION------------------------The single interface name (i.e. "bond0") representing the redundant set of NIC's is the interfaces that should be specified in the second screen in VIPCA (VIP Configuration Assistant, 1 of 2). Make sure not to select any of the underlying non-redundant NIC names in VIPCA, as they should not be used by Oracle in a NIC teaming configuration.2-8. OPTIONS FOR THE BONDING DRIVER-----------------------------------Various advanced interface, driver and switch configurations are available for achieving a highly available network configuration. Please refer to the "Linux Ethernet Bonding Driver mini-howto" for more details.http://www.kernel.org/pub/linux/kernel/people/marcelo/linux-2.4/Documentation/networking/bonding.txtRELATED DOCUMENTS-----------------Linux Ethernet Bonding Driver mini-howto:http://www.kernel.org/pub/linux/kernel/people/marcelo/linux-2.4/Documentation/networking/bonding.txtRed Hat Enterprise Linux 3: Reference Guide -> Appendix A. General Parameters and Modules -> A.3 Ethernet Parameters:http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/ref-guide/s1-modules-ethernet.htmlNote.283684.1 - How to Change Interconnect/Public Interface IP Subnet in a 10g ClusterNote.291962.1 - Setting Up Bonding on SLES 9Note.291958.1 - Setting Up Bonding in Suse SLES8
2011年07月19日
[b]Configuring Linux for the Oracle 10g VIP or private interconnect using bonding driver [ID 298891.1][/b] 修改时间 06-AUG-2010 类型 BULLETIN 状态 ARCHIVED
PURPOSE-------In order to avoid the public LAN from being a single point of failure, Oracle highly recommends configuring a redundant set of public network interface cards (NIC's) on each cluster node. Network redundancy can be achieved On Linux platforms using NIC Teaming (configuring multiple interfaces to a team or using the Linux kernel bonding module).This note will go over the two possible choices for achieving redundancy on Linux.As inter-node IP address failover is achieved by using the Oracle managed VIP, 3rd party clusterware based inter-node IP address failover technologies should not be configured on the same set of NIC's that are used by the Oracle VIP. Only intra-node IP address failover functionalities should be used in conjunction with the Oracle VIP.SCOPE & APPLICATION-------------------This article is intended for experienced DBAs and Support Engineers.1. NIC TEAMING BY CONFIGURING MULTIPLE INTERFACES TO A TEAM-----------------------------------------------------------Various hardware vendors provide network interface drivers and utilities to achieve NIC teaming. Please consult your hardware vendor for details on how to configure your system for NIC teaming2. NIC TEAMING USING THE LINUX KERNEL BONDING MODULE----------------------------------------------------The Linux kernel includes a bonding module that can be used to achieve software level NIC teaming. The kernel bonding module can be used to team multiple physical interfaces to a single logical interface, which is used to achieve fault tolerance and load balancing. The bonding driver is available as part of the Linux kernel version 2.4.12 or newer versions. Since the bonding module is delivered as part of the Linux kernel, it can be configured independently from the interface driver vendor (different interfaces can constitute a single logical interface).The configuration steps are different among Linux distributions. This note will go over the steps required to configure the bonding module in RedHat Enterprise Linux 3.0.In the following example, two physical interfaces (eth0 and eth1) will be bonded together to a single logical interface (bond0), and the VIP will run on top of the single logical interface.A sample network configuration is as follows: Default Gateway: 192.168.1.254 Netmask: 255.255.255.0 Interface configuration before bonding: eth0: IP Address 192.168.1.1 eth1: IP Address 192.168.1.2After configuring the bonding driver, a logical interface named bondX (where X is a number higher than zero) representing the team of interfaces. Interface configuration after bonding: bond0: IP Address 192.168.1.10Note: Before continuing, be sure the bonding module is mounted. The command "modprobe bonding" can be used to mount the module.2-1 CONFIGURING THE BONDING DRIVER----------------------------------Since the bonding driver is delivered as a kernel module, the following lines need to be added to /etc/modules.conf as root.alias bond0 bondingoptions bond0 miimon=100For details on the "options" parameter, please refer to the documents referred to in section 2.8. In the above configuration, the MII link monitoring interval is set to 100ms. MII is used to monitor the interface link status, and this is a typical configuration for mission critical systems that require fast failure detection.自己备注：实际上还有一个mode选项。mode=0 表示负载均衡方式，两块网卡都工作。mode=1 表示冗余方式，网卡只有一个工作，一个出问题启用另外的。如options bond0 miimon=100 mode=0 Note: MII is an abbreviation of "Media Independent Interface". Many popular fast ethernet adapters use MII to autonegotiate the link speed and duplex mode.By default, the bonding driver will transmit outgoing packets in a round-robin fashion using each "slave" interface. The above example uses this default behavior. For details on changing this behavior, please also refer to the documents referred to in section 2.8.If you want to use muitlple bonding interfaces you should modify /etc/modules.conf like below example.alias bond0 bondingalias bond1 bondingoptions bond0 miimon=100 max_bonds=2options bond1 miimon=100 max_bonds=2(In this example we have 2 bonding interfaces.)The "max_bonds" parameter defines how many bonding interfaces we aregoing to have.For details on the "max_bonds" parameter, please refer to the documentsreferred to in section 2.8. 2-2. CONFIGURING THE bond0 INTERFACE------------------------------------On RHEL 3.0, network interface parameters are configured in configuration files named "ifcfg-", found in the /etc/sysconfig/network-scripts directory. In order to enable the bonding driver, a configuration file "ifcfg-bond0" needs to be created with appropriate parameters. As root, create the file "/etc/sysconfig/network-scripts/ifcfg-bond0" as shown below.DEVICE=bond0IPADDR=192.168.1.10NETMASK=255.255.255.0NETWORK=192.168.1.0BROADCAST=192.168.1.255ONBOOT=yesBOOTPROTO=noneUSERCTL=no (Please change the IP address, Netmask, Broadcast to match your network configuration)2-3. CHANGING THE CONFIGURATION FOR THE EXISTING INTERFACES-----------------------------------------------------------As root, please change the configuration file "/etc/sysconfig/network-scripts/ifcfg-eth0" as shown below:DEVICE=eth0USERCTL=noONBOOT=yesMASTER=bond0SLAVE=yesBOOTPROTO=nonePlease also change the configuration file "/etc/sysconfig/network-scripts/ifcfg-eth1" as shown below:DEVICE=eth1USERCTL=noONBOOT=yesMASTER=bond0SLAVE=yesBOOTPROTO=noneThese steps are necessary to associate the bond0 interface to its slave interfaces (eth0 and eth1).2-4. RESTART THE NETWORK------------------------Execute the following commands as root to reflect the changes.# service network stop# service network startIf the configuration is correct, the above commands should both return [ OK ].2-5. CONFIRMING THE NEW CONFIGURATION-------------------------------------The following messages should appear in your syslog (/var/log/messages).Jan 28 16:00:09 rac01 kernel: bonding: MII link monitoring set to 100 msJan 28 16:00:09 rac01 kernel: ip_tables: (C) 2000-2002 Netfilter core teamJan 28 16:00:11 rac01 ifup: Enslaving eth0 to bond0Jan 28 16:00:11 rac01 kernel: bonding: bond0: enslaving eth0 as a backup interface with a down link.Jan 28 16:00:11 rac01 kernel: e1000: eth0: e1000_watchdog: NIC Link is Up 1000 Mbps Full DuplexJan 28 16:00:11 rac01 kernel: bonding: bond0: link status definitely up for interface eth0.Jan 28 16:00:11 rac01 kernel: bonding: bond0: making interface eth0 the new active one.Jan 28 16:00:11 rac01 ifup: ENslaving eth1 to bond0Jan 28 16:00:11 rac01 kernel: bonding: bond0: enslaving eth1 as a backup interface with a down link.Jan 28 16:00:11 rac01 kernel: e1000: eth1: e1000_watchdog: NIC Link is Up 1000 Mbps Full DuplexJan 28 16:00:11 rac01 kernel: bonding: bond0: link status definitely up for interface eth1.Jan 28 16:00:11 rac01 network: Bringing up interface bond0: succeededThe "ifconfig -a" command should return the following output.bond0 Link encap:Ethernet HWaddr 00:0C:29:DC:83:E8 inet addr:192.168.1.10 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1 RX packets:27 errors:0 dropped:0 overruns:0 frame:0 TX packets:1 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:3462 (3.3 Kb) TX bytes:42 (42.0 b)eth0 Link encap:Ethernet HWaddr 00:0C:29:DC:83:E8 inet addr:192.168.1.10 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1 RX packets:13 errors:0 dropped:0 overruns:0 frame:0 TX packets:1 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:1701 (1.6 Kb) TX bytes:42 (42.0 b) Interrupt:10 Base address:0x1424 eth1 Link encap:Ethernet HWaddr 00:0C:29:DC:83:E8 inet addr:192.168.1.10 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1 RX packets:14 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:1761 (1.7 Kb) TX bytes:0 (0.0 b) Interrupt:11 Base address:0x14a4 (Note that other interfaces will also appear in a typical RAC installation)2-6. CONSIDERATIONS FOR CRS INSTALLATION -----------------------------------------During CRS installation, choose "Public" for the "bond0" interface in the "Specify Network Interface Usage" OUI page. If the "eth0" and "eth1" interfaces appear in OUI, then make sure to choose "Do not use" for their types.2-6a. IF YOU WANT TO CONFIGURE BONDING DEVICES AFTER INSTALLATION OF CRS.-----------------------------------------You can change your interconnect/public interface configuration using oifcfg command. Please refer Note 283684.12-7. VIPCA CONFIGURATION------------------------The single interface name (i.e. "bond0") representing the redundant set of NIC's is the interfaces that should be specified in the second screen in VIPCA (VIP Configuration Assistant, 1 of 2). Make sure not to select any of the underlying non-redundant NIC names in VIPCA, as they should not be used by Oracle in a NIC teaming configuration.2-8. OPTIONS FOR THE BONDING DRIVER-----------------------------------Various advanced interface, driver and switch configurations are available for achieving a highly available network configuration. Please refer to the "Linux Ethernet Bonding Driver mini-howto" for more details.http://www.kernel.org/pub/linux/kernel/people/marcelo/linux-2.4/Documentation/networking/bonding.txtRELATED DOCUMENTS-----------------Linux Ethernet Bonding Driver mini-howto:http://www.kernel.org/pub/linux/kernel/people/marcelo/linux-2.4/Documentation/networking/bonding.txtRed Hat Enterprise Linux 3: Reference Guide -> Appendix A. General Parameters and Modules -> A.3 Ethernet Parameters:http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/ref-guide/s1-modules-ethernet.htmlNote.283684.1 - How to Change Interconnect/Public Interface IP Subnet in a 10g ClusterNote.291962.1 - Setting Up Bonding on SLES 9Note.291958.1 - Setting Up Bonding in Suse SLES8
扫一扫
235
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
