- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
Re: XS748T - LAG question
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
XS748T - LAG question
I have 2 Link Aggregations set on my switch. One is 2 10Gpbs ports and another is 4. The first connects to a Freenas box and the second to a Linux box. Both boxes have a LAG set up of 2 or 4 ports. They both work but I can not get data transfers in excess of a single port speed. I test with iperf and get basically the full 10Gbps but never more. I am using 802.3ad mode which according to my server documentation should give me faster speeds. I have included a blurb that describes what I expect to happen. I feel like I am missing something here either on the configuration side or on the expectation side. Does this switch support 802.3ad mode?
Mode 4 (802.3ad): This mode creates aggregation groups that share the same speed and duplex settings, and it requires a switch that supports an IEEE 802.3ad dynamic link. Mode 4 uses all interfaces in the active aggregation group. For example, you can aggregate three 1 GB per second (GBPS) ports into a 3 GBPS trunk port. This is equivalent to having one interface with 3 GBPS speed. It provides fault tolerance and load balancing.
Thanks
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Re: XS748T - LAG question
“This is equivalent to having one interface with 3 GBPS speed.”
The above statement is false. Setting up a LAG may increase the bandwidth of a link, but ‘bandwidth’ is not the same as ‘speed’. And, it just may to some degree. It all depends on the load balancing algorithm being employed and the usage of the environment.
Often, only the load balancing algorithm based on the source and destination MAC addresses is employed when selecting the network interface for a transmission. If those addresses do not change, one and the same interface is always used. That’s what happens when you run just a single iperf session. If you set up two parallel iperf sessions using two different PCs communicating with the same server, each session most likely will run at the speed close to 10 Gb/s. Without a LAG, each session would run with the speed close to 500 Mb/s.
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Re: XS748T - LAG question
Check if the bonding interfaces are active ....
[~] # ifconfig -a | grep bond
...
grab the output using e.g.
cat /proc/net/bonding/bond0
...
cat /proc/net/bonding/bond1
...
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Re: XS748T - LAG question
Thanks for that. I am struggling to be able to logon to the switch remotely. I have enabled remote admin but cannot get the switch to accept a telnet connection. I'll let you know when I get those commands properly executed.
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Re: XS748T - LAG question
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Re: XS748T - LAG question
This is the output for the two interfaces. lagg0 is on a freebsd box and team0 is on a Rocky linux box.
lagg0:
flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500 options=e407bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,TSO4,TSO6,LRO,VLAN_HWTSO,RXCSUM_IPV6,TXCSUM_IPV6>
ether a0:36:9f:3f:b3:8c
inet 10.0.9.31 netmask 0xffffff00 broadcast 10.0.9.255
nd6 options=9<PERFORMNUD,IFDISABLED>
media: Ethernet autoselect
status: active
groups: lagg
laggproto lacp lagghash l2,l3,l4
laggport: ix0 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>
laggport: ix1 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>
team0:
flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500
inet 10.0.9.32 netmask 255.255.255.0 broadcast 10.0.9.255
ether 84:16:0c:c4:b6:10 txqueuelen 1000 (Ethernet)
RX packets 20807488 bytes 379816736216 (353.7 GiB)
RX errors 0 dropped 345 overruns 0 frame 0
TX packets 11810865 bytes 1788077637 (1.6 GiB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Re: XS748T - LAG question
Servers often offer more load balancing algorithms than switches. I am not familiar with the innerworkings of FreeBSD, but I see from the “lagg0” listing that it offers the L3 and L4 algorithms in addition to L2 (“laggproto lacp lagghash l2,l3,l4”). You may be able to improve the performance of that LAG by adjusting the servers configuration. As for your switch, I’m afraid it supports only the L2 algorithm. However, there are switches on the market, even SMB switches not necessarily fully managed, that offer better load balancing support. My SMB switch supports both L2 and L3 algorithms.