Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 22 Feb 2019 11:31:26 +0100
From:      "Patrick M. Hausen" <hausen@punkt.de>
To:        FreeBSD Net <freebsd-net@freebsd.org>
Subject:   Performance issues with VNET/bridge/VLAN
Message-ID:  <9B0EC546-38E6-424E-9CC9-93F4C58B296F@punkt.de>

next in thread | raw e-mail | index | archive | help
Hi all,

please have a look at these two network setups:

------- separate interfaces -------=20
ifconfig_ixl0=3D"up"
ifconfig_ixl1=3D"up"

cloned_interfaces=3D"bridge0 bridge1"

ifconfig_bridge0_name=3D"inet0"
ifconfig_inet0=3D"addm ixl0 up"
ifconfig_inet0_alias0=3D"inet ww.xx.yy.zz/24"
ifconfig_inet0_ipv6=3D"inet6 2a00:b580:8000:11:dead:beef:dead:beef/64 =
auto_linklocal"

ifconfig_bridge1_name=3D"mgmt0"
ifconfig_mgmt0=3D"addm ixl1 up"
ifconfig_mgmt0_alias0=3D"inet 10.x.y.z/16"
ifconfig_mgmt0_ipv6=3D"inet6 auto_linklocal"
-----------------------------------=20

and

------- trunk port w/ VLANs -------=20
ifconfig_ixl0=3D"up"

cloned_interfaces=3D"vlan7 vlan11 bridge0 bridge1"

ifconfig_vlan7=3D"up vlan 7 vlandev ixl0"
ifconfig_vlan11=3D"up vlan 11 vlandev ixl0"

ifconfig_bridge0_name=3D"inet0"
ifconfig_inet0=3D"addm vlan11 up"
ifconfig_inet0_alias0=3D"inet ww.xx.yy.zz/24"
ifconfig_inet0_ipv6=3D"inet6 2a00:b580:8000:11:dead:beef:dead:beef/64 =
auto_linklocal"

ifconfig_bridge1_name=3D"mgmt0"
ifconfig_mgmt0=3D"addm vlan7 up"
ifconfig_mgmt0_alias0=3D"inet 10.x.y.z/16"
ifconfig_mgmt0_ipv6=3D"inet6 auto_linklocal"
-----------------------------------=20

If the switches at the other end are configured correctly, they should =
work
precisely the same, right? In the second case both networks share the =
bandwidth,
but the management network is mostly idle and only used for monitoring =
and
nightly backups.

The machine is an iocage jail host, all jails with VNET.

The problem is: network performance in the jails (not on the host!) is =
abysmal
with the second setup. Not consistently so, everything *seems* to work
but e.g. a customer complained that checking out a project from github
happend at 15k/s =E2=80=A6 that=E2=80=99s when we started to =
investigate.

After not finding anything fundamentally wrong we checked the =
differences
between individual hosts and the only essential one was the network =
setup
as shown above. So early this morning i rewired the host, got rid of the
VLANs - presto, github fast now.

*Any* idea what might be going on here? We use VNET all the same on all =
the
hosts and it is still labelled =E2=80=9Eexperimental", yes. But all the =
parts that
make up the different setups - bridge(4), vlan(4) - have been in FreeBSD
for ages. I=E2=80=99m just combining features orthogonally like every =
good sysadmin ;-)

If someone is willing to do some investigation, I think I can provide a =
test
system and remote access =E2=80=A6

Systems are running 11.2p3 at the moment. To be patched to 11.2p9 next
Tuesday.

Kind regards,
Patric
--=20
punkt.de GmbH			Internet - Dienstleistungen - Beratung
Kaiserallee 13a			Tel.: 0721 9109-0 Fax: -100
76133 Karlsruhe			info@punkt.de	http://punkt.de
AG Mannheim 108285		Gf: Juergen Egeling




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9B0EC546-38E6-424E-9CC9-93F4C58B296F>