Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

igb Intel Gigabit Network Connection I350, lagg, vlan - increase mtu

133 views
Skip to first unread message

Marek Salwerowicz

unread,
Sep 9, 2013, 9:02:22 AM9/9/13
to
Hi,

My server is running with 4 Gigabit NICs:

igb0@pci0:5:0:0: class=0x020000 card=0x152115d9 chip=0x15218086
rev=0x01 hdr=0x00
vendor = 'Intel Corporation'
device = 'I350 Gigabit Network Connection'
class = network
subclass = ethernet
igb1@pci0:5:0:1: class=0x020000 card=0x152115d9 chip=0x15218086
rev=0x01 hdr=0x00
vendor = 'Intel Corporation'
device = 'I350 Gigabit Network Connection'
class = network
subclass = ethernet
igb2@pci0:5:0:2: class=0x020000 card=0x152115d9 chip=0x15218086
rev=0x01 hdr=0x00
vendor = 'Intel Corporation'
device = 'I350 Gigabit Network Connection'
class = network
subclass = ethernet
igb3@pci0:5:0:3: class=0x020000 card=0x152115d9 chip=0x15218086
rev=0x01 hdr=0x00
vendor = 'Intel Corporation'
device = 'I350 Gigabit Network Connection'
class = network
subclass = ethernet


those NICs are connected to switch with LACP aggregation:

lagg0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=401bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,TSO4,VLAN_HWTSO>
ether 00:25:90:c1:1d:18
inet6 fe80::225:90ff:fec1:1d18%lagg0 prefixlen 64 scopeid 0x8
nd6 options=21<PERFORMNUD,AUTO_LINKLOCAL>
media: Ethernet autoselect
status: active
laggproto lacp lagghash l2,l3,l4
laggport: igb3 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>
laggport: igb2 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>
laggport: igb1 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>
laggport: igb0 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>


Using this lagg interface I created 2 vlan interfaces:

vlan14: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=103<RXCSUM,TXCSUM,TSO4>
ether 00:25:90:c1:1d:18
inet 192.168.1.65 netmask 0xffffff00 broadcast 192.168.1.255
inet6 fe80::225:90ff:fec1:1d18%vlan14 prefixlen 64 scopeid 0x9
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
media: Ethernet autoselect
status: active
vlan: 14 parent interface: lagg0


vlan900: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu
1500
options=103<RXCSUM,TXCSUM,TSO4>
ether 00:25:90:c1:1d:18
inet6 fe80::225:90ff:fec1:1d18%vlan900 prefixlen 64 scopeid 0xa
inet 172.25.25.65 netmask 0xffff0000 broadcast 172.25.255.255
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
media: Ethernet autoselect
status: active
vlan: 900 parent interface: lagg0


I would like to increase the MTU on cards, but I can't set it in any way:

% sudo ifconfig igbX mtu 9000 [14:59:37]
ifconfig: ioctl (set mtu): Invalid argument


event setting the MTU to 1500 (default value) results in same error.
same for lagg0, and vlanX


How can I increase the MTU for lagg interface ?
Is it possible without rebooting or only via /etc/rc.conf ?

Regards,
--
Marek Salwerowicz

_______________________________________________
freeb...@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to "freebsd-net...@freebsd.org"

Marek Salwerowicz

unread,
Sep 9, 2013, 9:06:37 AM9/9/13
to
W dniu 2013-09-09 15:02, Marek Salwerowicz pisze:
> My server

just to add - I am running the 9.1-RELEASE:

FreeBSD storage1 9.1-RELEASE-p6 FreeBSD 9.1-RELEASE-p6 #0: Wed Aug 21
20:40:52 UTC 2013
ro...@amd64-builder.daemonology.net:/usr/obj/usr/src/sys/GENERIC amd64

Mark van der Meulen

unread,
Sep 9, 2013, 9:11:55 AM9/9/13
to
As far as I can tell, if you are using the links in a lagg then you are going to need to make the change in /etc/rc.conf and reboot or perhaps restart netif. I recently came across a similar situation where I had to change the MTU across a bunch of different bsdrp boxes running as firewalls and routers and in the end after a lot of research I worked out the only way was either a reboot or possibly a restart of networking. Some machines had Intel(em) and Broadcom(bcd and bge) interfaces.

I never tried restarting networking as there are a few things on my boxes which that will break so I am better off just rebooting.

Mark

Michael Schuh

unread,
Sep 10, 2013, 8:23:48 AM9/10/13
to
Hi Marek,

if i remember well the order of doing so is important.

#1 set the mtu on each physical NIC to 9000
#2 create the lagg
.
.
.
on a running system it might consist of: destroy the vlans, destroy the
lagg,
set the mtu on each nic and rebuild the lagg and add the vlans.


HTH

regards

michael

Marek Salwerowicz

unread,
Sep 13, 2013, 3:42:48 AM9/13/13
to
W dniu 2013-09-09 15:11, Mark van der Meulen pisze:
> As far as I can tell, if you are using the links in a lagg then you are going to need to make the change in /etc/rc.conf and reboot or perhaps restart netif. I recently came across a similar situation where I had to change the MTU across a bunch of different bsdrp boxes running as firewalls and routers and in the end after a lot of research I worked out the only way was either a reboot or possibly a restart of networking. Some machines had Intel(em) and Broadcom(bcd and bge) interfaces.
>
> I never tried restarting networking as there are a few things on my boxes which that will break so I am better off just rebooting.
Hi,

after setting up in /etc/rc.conf:

ifconfig_igb{0,1,2,3}="mtu 9000 up"

and rebooting everything worked and I have mtu 9000 in lagg0 and 2 vlans

Thanks for hints!

--
Marek Salwerowicz
0 new messages