LACP kasutamine - Redigeerimiste ajalugu

Jj: /* Kasulikud lisamaterjalid */

2014-10-12T16:38:37Z

Kasulikud lisamaterjalid

←Vanem redaktsioon		Redaktsioon: 12. oktoober 2014, kell 16:38
182. rida:		182. rida:
	* http://www.wiki.cl.cam.ac.uk/rowiki/SysInfo/LinksysSwitch		* http://www.wiki.cl.cam.ac.uk/rowiki/SysInfo/LinksysSwitch
	* http://www.kernel.org/doc/Documentation/networking/bonding.txt		* http://www.kernel.org/doc/Documentation/networking/bonding.txt

			http://www.linuxfoundation.org/collaborate/workgroups/networking/bonding

Jj: /* Seadistamine */

2014-10-12T16:37:17Z

Seadistamine

←Vanem redaktsioon		Redaktsioon: 12. oktoober 2014, kell 16:37
54. rida:		54. rida:
	* Balance TLB - Adaptive Transmit Load Balancing. Incoming traffic is received on the active slave only, outgoing traffic is distributed according to the current load on each slave. Doesn't require special switch support		* Balance TLB - Adaptive Transmit Load Balancing. Incoming traffic is received on the active slave only, outgoing traffic is distributed according to the current load on each slave. Doesn't require special switch support
	* Balance ALB - Adaptive Load Balancing - provides both transmit load balancing (TLB) and receive load balancing for IPv4 via ARP negotiation. Doesn't require special switch support, but does require the ability to change the MAC address of a device while it is open		* Balance ALB - Adaptive Load Balancing - provides both transmit load balancing (TLB) and receive load balancing for IPv4 via ARP negotiation. Doesn't require special switch support, but does require the ability to change the MAC address of a device while it is open

			Vaikimisi on balance-rr (round robin) ehk 0.

	Vastav seadistamine käsurealt võiks toimuda nt selliselt		Vastav seadistamine käsurealt võiks toimuda nt selliselt

Jj: /* Sissejuhatus */

2014-10-12T16:35:10Z

Sissejuhatus

←Vanem redaktsioon		Redaktsioon: 12. oktoober 2014, kell 16:35
2. rida:		2. rida:

	Ethernet interface bonding tehnika võimaldab teatud tingimustel kasutada mitut arvutisse paigaldatud füüsilist võrguseadet ühe loogilise seadmena.		Ethernet interface bonding tehnika võimaldab teatud tingimustel kasutada mitut arvutisse paigaldatud füüsilist võrguseadet ühe loogilise seadmena.

			'''TODO - Alljärgnev tõlkida ära.'''

	The reason you dont get a throughput greater than 943Mbps is due to the way that link aggregation (LACP) works.		The reason you dont get a throughput greater than 943Mbps is due to the way that link aggregation (LACP) works.
8. rida:		10. rida:
	This means that a connection between two hosts can never use more than 1G, but if you have two connections you might be able to use 2 x 1G; if they choose to use different links in the LACP link (this is based on some calculation using eg. IP address, mac address etc.). In other words, a connection between two hosts, can never use more bandwidth than the original links bandwidth.		This means that a connection between two hosts can never use more than 1G, but if you have two connections you might be able to use 2 x 1G; if they choose to use different links in the LACP link (this is based on some calculation using eg. IP address, mac address etc.). In other words, a connection between two hosts, can never use more bandwidth than the original links bandwidth.

	when ~~i changed to mode0 roundrobin Method~~, ~~Both ethernet~~ interface ~~fully utilized~~ and ~~giving 200MB~~/~~s throughput~~. ~~Please see~~ the ~~attachment~~ for the ~~mode0 throughput in mode0~~.~~png~~		It's just not possible to make 1 x 2 Gbps link using 2 x 1 Gbps links using LACP. The reason is, that when a host sends data using a LACP link, it determines which interface (link) to use by calculating a hash value based on a combination of L2 and L3 information. So if you have two hosts you will get the same hash value each time since the source/destination mac/IP addresses don't change, and the sending host will send data using only one (the same) link/interface. If the sending host is sending data to two different hosts, you might be lucky, that the hash value for the two receiving hosts matches each of the links on the sending hosts, thereby sending data on both links/interfaces.

	===Tööpõhimõte===		===Tööpõhimõte===

Jj: /* Seadistamine */

2014-10-12T16:33:29Z

Seadistamine

←Vanem redaktsioon		Redaktsioon: 12. oktoober 2014, kell 16:33
45. rida:		45. rida:
	Erinevate olekute tähendused		Erinevate olekute tähendused

	~~mode=0 (balance~~-~~rr)~~		* Round Robin - Packets are transmitted in a round robin fashion over the available slave interfaces. Provides both load balancing and fault tolerance
	~~mode=1 (~~active-~~backup~~)		*Active Backup - One slave interface is active at any time. If one interface fails, another interface takes over the MAC address and becomes the active interface. Provides fault tolerance only. Doesn't require special switch support
	~~mode=2~~ (~~balance-xor~~)		* Balance XOR - Tranmissions are balanced across the slave interfaces based on ((source MAC) XOR (dest MAC)) modula slave count. The same slave is selected for each destination MAC. Provides load balancing and fault tolerance.
	~~mode=3 (broadcast)~~		* Broadcast - Transmits everything on all slave interfaces. Provides fault tolerance.
	~~mode=4 (~~802.3ad~~) (Link Aggregation ja LACP)~~		* 802.3ad - This is classic IEEE 802.3ad Dynamic link aggregation. This requires 802.3ad support in the switch and driver support for retrieving the speed and duplex of each slave.
	~~mode=5 (balance~~-~~tlb)~~		* Balance TLB - Adaptive Transmit Load Balancing. Incoming traffic is received on the active slave only, outgoing traffic is distributed according to the current load on each slave. Doesn't require special switch support
	~~mode=6~~ (~~balance-alb~~)		* Balance ALB - Adaptive Load Balancing - provides both transmit load balancing (TLB) and receive load balancing for IPv4 via ARP negotiation. Doesn't require special switch support, but does require the ability to change the MAC address of a device while it is open

	Vastav seadistamine käsurealt võiks toimuda nt selliselt		Vastav seadistamine käsurealt võiks toimuda nt selliselt

Jj: /* Sissejuhatus */

2014-10-12T16:32:37Z

Sissejuhatus

←Vanem redaktsioon		Redaktsioon: 12. oktoober 2014, kell 16:32
2. rida:		2. rida:

	Ethernet interface bonding tehnika võimaldab teatud tingimustel kasutada mitut arvutisse paigaldatud füüsilist võrguseadet ühe loogilise seadmena.		Ethernet interface bonding tehnika võimaldab teatud tingimustel kasutada mitut arvutisse paigaldatud füüsilist võrguseadet ühe loogilise seadmena.

			The reason you dont get a throughput greater than 943Mbps is due to the way that link aggregation (LACP) works.

			When you aggregate two links using LACP eg. 2 x 1G you dont get a 1 x 2G link; you get one link of 2 x 1G.
			This means that a connection between two hosts can never use more than 1G, but if you have two connections you might be able to use 2 x 1G; if they choose to use different links in the LACP link (this is based on some calculation using eg. IP address, mac address etc.). In other words, a connection between two hosts, can never use more bandwidth than the original links bandwidth.

			when i changed to mode0 roundrobin Method, Both ethernet interface fully utilized and giving 200MB/s throughput. Please see the attachment for the mode0 throughput in mode0.png

	===Tööpõhimõte===		===Tööpõhimõte===

Jj – 12. oktoober 2014, kell 16:25

2014-10-12T16:25:57Z

@@ 35. rida: / 35. rida: @@
     bond_miimon 100
     bond-lacp-rate 4
 Vastav seadistamine käsurealt võiks toimuda nt selliselt

Jj: /* Seadistamine */

2014-10-10T13:40:27Z

Seadistamine

←Vanem redaktsioon		Redaktsioon: 10. oktoober 2014, kell 13:40
25. rida:		25. rida:
	====Seadistamine====		====Seadistamine====

	Bonding seadme nimeks on tavaliselt bond0 ning LACP bonding ~~seade~~ seadistatakse /etc/network/interfaces sektsiooniga		Bonding seadme nimeks on tavaliselt bond0 ning LACP bonding (mitte segi ajada failover-bond seadmega) seadistatakse /etc/network/interfaces sektsiooniga

	auto bond0		auto bond0

Jj: /* Seadistamine */

2014-10-10T13:39:46Z

Seadistamine

←Vanem redaktsioon		Redaktsioon: 10. oktoober 2014, kell 13:39
25. rida:		25. rida:
	====Seadistamine====		====Seadistamine====

	Bonding seadme nimeks on tavaliselt bond0 ning ~~nt failover~~ bonding seade seadistatakse /etc/network/interfaces sektsiooniga		Bonding seadme nimeks on tavaliselt bond0 ning LACP bonding seade seadistatakse /etc/network/interfaces sektsiooniga

	auto bond0		auto bond0

Jj – 10. oktoober 2014, kell 13:25

2014-10-10T13:25:44Z

←Vanem redaktsioon		Redaktsioon: 10. oktoober 2014, kell 13:25
32. rida:		32. rida:
	netmask 255.255.255.0		netmask 255.255.255.0
	slaves eth1 eth2		slaves eth1 eth2
	bond_mode ~~active-backup~~		bond_mode 802.3ad
	bond_miimon 100		bond_miimon 100
			bond-lacp-rate 4

	Vastav seadistamine käsurealt võiks toimuda nt selliselt		Vastav seadistamine käsurealt võiks toimuda nt selliselt

Imre: /* Kasulikud lisamaterjalid */

2012-09-26T21:09:31Z

Kasulikud lisamaterjalid

←Vanem redaktsioon		Redaktsioon: 26. september 2012, kell 21:09
149. rida:		149. rida:

	*Ametlik FreeBSD manual http://www.freebsd.org/doc/handbook/network-aggregation.html		*Ametlik FreeBSD manual http://www.freebsd.org/doc/handbook/network-aggregation.html

			===Märkused===

			* Tundub, et bondinguks ettevalmistatud switchi kaudu arvutit ei õnnestu üle PXE alglaadida.

	===Kasulikud lisamaterjalid===		===Kasulikud lisamaterjalid===