|
Blue Forest http://www.lslnet.com at 20:18 on April 6, 2006
FREE BSD can be tied up two fiber!
Locate ZP two bars the use of fiber telecommunications 4M
3 requirements are as follows : a server load balancing so fast NIC
Gateway IP 192.168.1.254 need only one example :
In addition to ask you about heroes tied up in what ways can use Linux, FREE BSD can be used?
If so, please give the information to be grateful. |
FREE BSD can be tied up two fiber!
Iproute2 said Linux combined with a small shell can be achieved, but if one link problems, the whole link failure. I have done useless LINUX. . It is to listen to other people say.
I now intend to do with freeBSD5.3 with two fiber. .
All preparatory work for a good job. . Testing on the clothes, such as read. . . . . My method is as follows :
Kernel :
[quote]machine From beginning for the i386
Cpu I586_CPU
Cpu I686_CPU
Ident GENERIC
Device pf
Device pflog
Device pfsync
Options ALTQ
Options ALTQ_CBQ
# To statically compile in device wiring instead of /boot/device.hints
Kernel options PANIC_REBOOT_WAIT_TIME=0 # abnormal effect immediately
Options DEVICE_POLLING # improve network response time
#man Four polling options HZ=1000
IP transmitting secret support options IPSTEALTH #
#options RANDOM_IP_ID random IP ID # to prevent information leakage
SYN+FIN options TCP_DROP_SYNFIN # discarded packets to prevent detection server
# 4BSD scheduler options SCHED_4BSD
Options INET # InterNETworking
Options FFS # Berkeley Fast Filesystem
SOFTUPDATES # Enable FFS soft updates support options
Options MD_ROOT # MD is a potential root device
Options PROCFS # Process filesystem (requires PSEUDOFS)
Pseudo-filesystem framework options PSEUDOFS #
Options COMPAT_43 # Compatible with BSD 4.3 [KEEP THIS!]
SYSV-style shared memory options SYSVSHM #
Options SYSVMSG # SYSV-style message queues
Options SYSVSEM # SYSV-style semaphores
# POSIX P1003_1B real-time extensions options _KPOSIX_PRIORITY_SCHEDULING
Options KBD_INSTALL_CDEV # install a CDEV entry in /dev
APIC device apic # I/O
# Bus support. Do not remove isa, even if you have no isa slots
Device isa
Pci device
# ATA and ATAPI devices
Device-Quds
Device atadisk # ATA disk drives
ATA_STATIC_ID # Static device numbering options
# SCSI peripherals
Device scbus # SCSI bus (required for SCSI)
Device da # Direct Access (disks)
# Atkbdc0 controls both the PS/2 keyboard and the mouse
# AT keyboard controller device atkbdc
# AT keyboard device atkbd
VGA video card driver device vga #
Splash # device splash screen and screen saver support
# Syscons is the default console driver, resembling an SCO console
Device sc
# Floating point support-do not disable.
Device npx
# Serial (COM) ports
Device sio # 8250, 16[45]50 based serial ports
# PCI Ethernet NICs.
Device de # DEC/Intel DC21x4x (``Tulip '')
Intel PRO/1000 Gigabit Ethernet adapter device em # Card
Intel PRO/10GbE Ethernet Card # device ixgb
Device txp # 3Com 3cR990 (``Typhoon '')
Device vx # 3Com 3c590, 3c595 (``Vortex '')
# PCI Ethernet NICs that use the common MII bus controller code.
# NOTE : Be sure to keep the 'device miibus' line in order to use these NICs!
Miibus # MII bus support device
# Broadcom BCM440x 10/100 Ethernet device bfe
# Broadcom BCM570xx Gigabit Ethernet device bge
DEC/Intel device dc # 21143 and various workalikes
Device fxp # Intel EtherExpress PRO/100B (82557, 82558)
Level 1 LXT1001 gigabit ethernet device lge #
NatSemi DP83820 gigabit ethernet device alleges #
Device pcn # AMD Am79C97x PCI 10/100 (precedence over '-1.0')
# RealTek 8139C+/8169/8169S/8110S device Ampere
Uwe # RealTek 8129/8139 device
# Device sf devaluation AIC-6915 (``Starfire '')
Device sis # Silicon Integrated Systems SiS 900/SiS 7016
Device sk # SysKonnect SK-984x & SK-982x gigabit Ethernet
Sundance local terrorist organizations # device ST201 (D-Link DFE-550TX)
Mixed device ti # Alteon Networks Tigon gigabit Ethernet
Texas Instruments ThunderLAN # device tl
Device tx # SMC EtherPower II (83c170 ``EPIC '')
VIA VT612x gigabit ethernet device vge #
Device vr # VIA Rhine, Rhine II
Winbond W89C840F device dying #
Quantitative # 3Com 3c90x device (``Boomerang '' ``Cyclone '')
# ISA Ethernet NICs. Pccard NICs included.
Crystal Semiconductor device CS89x0 NIC # cs
# 'Device ed' requires' device miibus'
Ed # NE[12]000 device, SMC Ultra, 3c503, DS8390 cards
Device ex # Intel EtherExpress Pro/10 and Pro/10+
HERA ep # Etherlink III device based cards
Fujitsu MB8696x based cards device fe #
Device ie # EtherExpress preoperatively, 3C507, StarLAN 10 etc..
# NE2100 device 10-2, Lance Ethernet cards NE32-VL
SMC device sn # 's 9000 series of Ethernet chips
Pccard # Xircom Ethernet device xe
# ISA devices that use the old ISA shims
#device Le
# Pseudo devices.
Network loopback device loop #
Kernel device mem # Memory and memory devices
Dean # I/O device device
Entropy device device random #
Ethernet support device ether #
PPP device ppp # Kernel
Device tun # Packet tunnel.
# Pseudo-ttys device pty (Telnet etc.)
Device md # Memory "disks"
Gif # IPv6 and IPv4 tunneling device
# The `bpf 'device enables the Berkeley Packet Filter.
# Be aware of the administrative consequences of enabling this!
Device bpf # Berkeley packet filter
# USB support
PCI->;USB interface device uhci # UHCI
Ohci # interface device for OHCI PCI->;USB
Device usb # USB Bus (required)
Generic device ugen #
Keyboard device ukbd #
# Disks/Mass storage device umass - Requires scbus and da
Device ums # Mouse
Options NETGRAPH #netgraph (4) system
Options NETGRAPH_ETHER
Options NETGRAPH_PPPOE
Options NETGRAPH_SOCKET
[/quote] Device vlan # IEEE 802.1Q VLAN Support
/etc/pf.conf Allocation :
(Based on the individual circumstances of the following changes :
Int_if = "dc0"
Ext_if1 = "fxp0"
Ext_if2 = "fxp1"
Ext_gw1 = "68.146.224.1"
Ext_gw2 = "142.59.76.1")
[quote]lan_net = "192.168.0.0/24"
Int_if = "dc0"
Ext_if1 = "fxp0"
Ext_if2 = "fxp1"
Ext_gw1 = "68.146.224.1"
Ext_gw2 = "142.59.76.1"
# Senat outgoing connections on each internet interface
Senat on $ext_if1 from $lan_net to any ->; ($ext_if1)
Senat on $ext_if2 from $lan_net to any ->; ($ext_if2)
# Default deny
Block in from any to any
Block out from any to any
# Pass all outgoing packets on internal interface
To pass out on $int_if from any $lan_net
# Pass in quick any packets destined for the gateway itself
Pass in quick on $int_if from $lan_net to $int_if
# Load balance outgoing tcp traffic from internal network.
Pass in on $int_if route-to \
(($ext_if1 $ext_gw1), ($ext_if2 $ext_gw2)) \ Round-robin
$lan_net Proto tcp from any to modulate state flags S/SA
# Load balance outgoing udp and icmp traffic from internal network
Pass in on $int_if route-to \
(($ext_if1 $ext_gw1), ($ext_if2 $ext_gw2)) \ Round-robin
-- Proto udp, and icmp) from $lan_net to keep any state
# General "pass out" rules for external interfaces
Pass out on $ext_if1 proto tcp from any to any modulate state flags S/SA
-- Pass out on $ext_if1 proto udp, and icmp from any to any keep state)
Pass out on $ext_if2 proto tcp from any to any modulate state flags S/SA
-- Pass out on $ext_if2 proto udp, and icmp from any to any keep state)
# $ext_if1 To route packets from any bands on and the same for $ext_gw1
# $ext_if2 And $ext_gw2
Pass out on $ext_if1 route-to ($ext_if2 $ext_gw2) from $ext_if2 to any
Pass out on $ext_if2 route-to ($ext_if1 $ext_gw1) from $ext_if1 to any [/quote]
Do not know whether it was achievable. . What advantages and disadvantages. . . I have not yet tried it. . :) |
FREE BSD can be tied up two fiber!
Pf really can be achieved with load balancing, but outgoing traffic, do not seem to support failover, that is, if one link cut off, then this route to link the package would be lost. |
FREE BSD can be tied up two fiber!
[quote][i] Note from the original "dennis2"] pf really can be achieved with load balancing, but outgoing traffic, do not seem to support failover, that is, if one link cut off, then this route to link the package would be lost. [/quote [/i] Released :
呢?? how to amend it : ( |
FREE BSD can be tied up two fiber!
How many Linux NIC bonding
--------------------------------------------------------------------------------
NIC NIC become a piece of virtual multi-block with the same IP address. Creating the sun and the fact the technology already exists, and called Trunking etherchannel technologies in Linux, such as bonding technology.
Because bonding had been included in kernel 2.4.x, only in the translation of the network equipment options can be selected by Bonding driver support.
Then, re-editing core, restart the computer, the implementation of the order as follows :
Ismod bonding
Ifconfig eth0 down
Ifconfig eth1 down
Ifconfig bond0 ipaddress
Ifenslave bond0 eth0
Ifenslave bond0 eth1
Now has two network interface card as a piece of the same work. This can increase the data transmission between cluster nodes.
This something that you wrote a script, then /etc/rc.d/rc.local Calling for a switched on the force.
Bonding is a relatively good choice for the server, in the absence of Gigabit network adapter, network card for bonding two or three pieces of 100 megabytes, which can greatly enhance the bandwidth between switches to the servers. However, the need to connect switches installed on the NIC bonding with the two sources for mapping a virtual adapter. |
FREE BSD can be tied up two fiber!
[quote][i] Note from the original "imagecoco" [/i] released :
How many Linux NIC bonding
--------------------------------------------------------------------------------
NIC NIC become a piece of virtual multi-block with the same IP address. In fact, the technology already exists in the sun and cisco. Hours? . . . . . . . . . [/quote]
What you said is right server load balancing. . I need the link load balancing. |
FREE BSD can be tied up two fiber!
[quote][i] Note from the original "jk0wg"] how to amend it 呢?? : ([/quote [/i] released :
I think we can use a running cron Lane in the shell script to monitor for each link, if monitoring link to a certain cut off, then it is immediately replaced by another set pf the rule set (only one link), so that all the packets go from good access link. This may not guarantee that there are no packet loss, but should be regarded as a blessing. |
FREE BSD can be tied up two fiber!
I want to use two WAN lines ah. . Line no problem under normal circumstances. . . Fiber from Rafah at least a year now. . No access routes off too. . There is only slow phenomenon. . . But normal soon. . |
FREE BSD can be tied up two fiber!
The question no one seems to give a satisfactory answer! Master? Given to the proposal. |
FREE BSD can be tied up two fiber!
"; The question no one seems to give a satisfactory answer!
I think the answer would be a blessing jk0wg (I wonder if you understand that? ), But if we can solve the problem better failover. The best way to solve this problem is to use formal BGP peering. |
FREE BSD can be tied up two fiber!
Firewall and failover in OBSD using pfsync http://www.countersiege.com/doc/ CARP
Pfsync-carp/, can not FBSD No, I have not tried.
IP Bonding under the FBSD no problem, the module can be used netgraph. Ngctl because it is quite simple. But Bonding not load balancing. . . . . |
FREE BSD can be tied up two fiber!
Ucarp support in FreeBSD, but there freevrrpd FreeBSD, and similar things. However, we are discussing is the failover link is not the firewall itself :) |
FREE BSD can be tied up two fiber!
[quote][i] Note from the original "dennis2"] ucarp not support FreeBSD, but there freevrrpd FreeBSD, and similar things. However, we are discussing is the failover link is not the firewall itself :) : [/quote published [/i]
I have never understood what he asked. . . : P : P
I tried twice Freevrrpd (http://uni.x-si.org/hamail.html), the heartbeat+drbd Linux does not seem so convenient, because I have forgotten that : mrgreen : Perhaps you can give us means of clear direction? |
FREE BSD can be tied up two fiber!
[quote][i] Note from the original "werix" I tried twice ]Freevrrpd (http://uni.x-si.org/hamail.html), the heartbeat+drbd Linux does not seem so convenient, because I have forgotten that : mrgreen : Perhaps you can give us means of clear direction? [/quote [/i] Released :
Unfortunately, I have not used the heartbeat + drbd. . . : (
I have rarely use Linux, seemed to be a catch up :) |
FREE BSD can be tied up two fiber!
Thank you for your answer. . I would like to ask you a master. Which means I have. If the normal line. 呢?? wealthy handy, I can say with assurance lines have been normal. . . I said which is not in line with normal circumstances like 呢?? |
FREE BSD can be tied up two fiber!
[quote][i] Note from the original "jk0wg"] I want to use two WAN lines ah. . Line no problem under normal circumstances. . . Fiber from Rafah at least a year now. . No access routes off too. . There is only slow phenomenon. . . But normal soon. . [/quote Released : [/i]
Generally speaking it is impossible for the two wan lines tied to the same time, it used the round robin, and not to increase the network bandwidth, but we can solve some of the network traffic load.
If you want to use 102 lines tied together, we can make reference to the relevant article forum. |
FREE BSD can be tied up two fiber!
I only need to be resolved is the network traffic load. . Simply do not know what better way 呢?? |
FREE BSD can be tied up two fiber!
"; I only need to be resolved is the network traffic load. . Simply do not know what better way 呢??
On the way to the second floor, you can hide, if not the link may be down. |
FREE BSD can be tied up two fiber!
But which way do I think of the complicated. . Today, mastering the next. . Not that it comes out. . |
FREE BSD can be tied up two fiber!
Expectations, which made affixed to the trouble, and we now have only one link, it may also have two first months of technical preparations for the first :) |
| |