[ovs-dev] [PATCH 3/3] system-traffic: Add conntrack floating IP test
Eric Garver
e at erig.me
Wed Oct 25 16:38:53 UTC 2017
On Tue, Oct 24, 2017 at 02:10:47PM -0700, William Tu wrote:
> On Tue, Oct 24, 2017 at 1:44 PM, Eric Garver <e at erig.me> wrote:
> > On Tue, Oct 24, 2017 at 12:46:46PM -0700, William Tu wrote:
> >> On Fri, Oct 20, 2017 at 11:23 AM, Eric Garver <e at erig.me> wrote:
> >> > This test cases uses floating IP (FIP) addresses for each endpoint. If
> >> > the destination is a FIP, the packet will undergo a transformation of
> >> > the form (dst=FIP, src=non-FIP) --> (dst=non-FIP, src=FIP) before
> >> > egress. Otherwise the packet is untouched.
> >> >
> >> > This exercises the ct_clear action in the datapath.
> >> >
> >> > Signed-off-by: Eric Garver <e at erig.me>
> >> > ---
> >> > tests/system-traffic.at | 73 +++++++++++++++++++++++++++++++++++++++++++++++++
> >> > 1 file changed, 73 insertions(+)
> >> >
> >> > diff --git a/tests/system-traffic.at b/tests/system-traffic.at
> >> > index 522eaa615834..7cc1e1e21187 100644
> >> > --- a/tests/system-traffic.at
> >> > +++ b/tests/system-traffic.at
> >> > @@ -3996,6 +3996,79 @@ ovs-ofctl -O OpenFlow15 dump-group-stats br0
> >> > OVS_TRAFFIC_VSWITCHD_STOP
> >> > AT_CLEANUP
> >> >
> >> > +AT_SETUP([conntrack - floating IP])
> >> > +AT_SKIP_IF([test $HAVE_NC = no])
> >> > +CHECK_CONNTRACK()
> >> > +OVS_TRAFFIC_VSWITCHD_START()
> >> > +OVS_CHECK_CT_CLEAR()
> >> > +
> >> > +ADD_NAMESPACES(at_ns0, at_ns1)
> >> > +ADD_VETH(p0, at_ns0, br0, "10.1.1.1/24", "f0:00:00:01:01:01") dnl FIP 10.254.254.1
> >> > +ADD_VETH(p1, at_ns1, br0, "10.1.1.2/24", "f0:00:00:01:01:02") dnl FIP 10.254.254.2
> >> > +
> >> > +dnl Static ARPs
> >> > +NS_CHECK_EXEC([at_ns0], [ip neigh add 10.1.1.2 lladdr f0:00:00:01:01:02 dev p0])
> >> > +NS_CHECK_EXEC([at_ns1], [ip neigh add 10.1.1.1 lladdr f0:00:00:01:01:01 dev p1])
> >> > +
> >> > +dnl Static ARP and route entries for the FIP "gateway"
> >> > +NS_CHECK_EXEC([at_ns0], [ip neigh add 10.1.1.254 lladdr f0:00:00:01:01:FE dev p0])
> >> > +NS_CHECK_EXEC([at_ns1], [ip neigh add 10.1.1.254 lladdr f0:00:00:01:01:FE dev p1])
> >> > +NS_CHECK_EXEC([at_ns0], [ip route add default nexthop via 10.1.1.254])
> >> > +NS_CHECK_EXEC([at_ns1], [ip route add default nexthop via 10.1.1.254])
> >> > +
> >> > +NETNS_DAEMONIZE([at_ns0], [nc -l -k > /dev/null], [nc0.pid])
> >> > +
> >> > +AT_DATA([flows.txt], [dnl
> >> > +table=0,priority=10 ip action=ct(table=1)
> >> > +table=0,priority=1 action=drop
> >> > +dnl dst FIP
> >> > +table=1,priority=20 ip,ct_state=+trk+est,nw_dst=10.254.254.0/24 action=goto_table:10
> >> > +table=1,priority=20 ip,ct_state=+trk+new,nw_dst=10.254.254.0/24 action=ct(commit,table=10)
> >> > +dnl dst local
> >> > +table=1,priority=10 ip,ct_state=+trk+est action=goto_table:20
> >> > +table=1,priority=10 ip,ct_state=+trk+new action=ct(commit,table=20)
> >> > +table=1,priority=1 ip,ct_state=+trk+inv action=drop
> >> > +dnl
> >> > +dnl FIP translation (dst FIP, src local) --> (dst local, src FIP)
> >> > +table=10 ip,nw_dst=10.254.254.1 action=set_field:10.1.1.1->nw_dst,goto_table:11
> >> > +table=10 ip,nw_dst=10.254.254.2 action=set_field:10.1.1.2->nw_dst,goto_table:11
> >> > +table=11 ip,nw_src=10.1.1.1 action=set_field:10.254.254.1->nw_src,goto_table:12
> >> > +table=11 ip,nw_src=10.1.1.2 action=set_field:10.254.254.2->nw_src,goto_table:12
> >> > +dnl clear conntrack and do another lookup since we changed the tuple
> >> > +table=12,priority=10 ip action=ct_clear,ct(table=13)
> >> > +table=12,priority=1 action=drop
> >> > +table=13 ip,ct_state=+trk+est action=goto_table:20
> >> > +table=13 ip,ct_state=+trk+new action=ct(commit,table=20)
> >> > +table=13 ip,ct_state=+trk+inv action=drop
> >> > +dnl
> >> > +dnl Output
> >> > +table=20 ip,nw_src=10.1.1.1 action=set_field:f0:00:00:01:01:01->eth_src,goto_table:21
> >> > +table=20 ip,nw_src=10.1.1.2 action=set_field:f0:00:00:01:01:02->eth_src,goto_table:21
> >> > +table=20 ip,nw_src=10.254.254.0/24 action=set_field:f0:00:00:01:01:FE->eth_src,goto_table:21
> >> > +table=21 ip,nw_dst=10.1.1.1 action=set_field:f0:00:00:01:01:01->eth_dst,output:ovs-p0
> >> > +table=21 ip,nw_dst=10.1.1.2 action=set_field:f0:00:00:01:01:02->eth_dst,output:ovs-p1
> >> > +])
> >> > +
> >> > +AT_CHECK([ovs-ofctl --bundle add-flows br0 flows.txt])
> >> > +
> >> > +dnl non-FIP case
> >> > +NS_CHECK_EXEC([at_ns1], [echo "foobar" |nc $NC_EOF_OPT 10.1.1.1])
> >> > +OVS_WAIT_UNTIL([[ovs-appctl dpctl/dump-conntrack | sed -e 's/port=[0-9]*/port=<cleared>/g' -e 's/id=[0-9]*/id=<cleared>/g' |
> >> > +grep "tcp,orig=(src=10.1.1.2,dst=10.1.1.1,sport=<cleared>,dport=<cleared>),reply=(src=10.1.1.1,dst=10.1.1.2,sport=<cleared>,dport=<cleared>),protoinfo=(state=TIME_WAIT)"
> >> > +]])
> >> > +
> >> > +dnl Check that the full session ends as expected (i.e. TIME_WAIT). Otherwise it
> >> > +dnl means the datapath didn't process the ct_clear action. Ending in SYN_RECV
> >> > +dnl (OVS maps to ESTABLISHED) means the initial frame was committed, but not a
> >> > +dnl second time after the FIP translation (because ct_clear didn't occur).
> >> > +NS_CHECK_EXEC([at_ns1], [echo "foobar" |nc $NC_EOF_OPT 10.254.254.1])
> >> > +OVS_WAIT_UNTIL([[ovs-appctl dpctl/dump-conntrack | sed -e 's/port=[0-9]*/port=<cleared>/g' -e 's/id=[0-9]*/id=<cleared>/g' |
> >> > +grep "tcp,orig=(src=10.254.254.2,dst=10.1.1.1,sport=<cleared>,dport=<cleared>),reply=(src=10.1.1.1,dst=10.254.254.2,sport=<cleared>,dport=<cleared>),protoinfo=(state=TIME_WAIT)"
> >> > +]])
> >> > +
> >> > +OVS_TRAFFIC_VSWITCHD_STOP
> >> > +AT_CLEANUP
> >> > +
> >> > AT_BANNER([802.1ad])
> >> >
> >> > AT_SETUP([802.1ad - vlan_limit])
> >> > --
> >>
> >> I'm using
> >> ~/ovs# make check-system-userspace TESTSUITEFLAGS='83'
> >> but fails due to the nc $NC_EOF_OPT.
> >> I suggest we explicitly add network port number, ex: 1234 below
> >>
> >> --- a/tests/system-traffic.at
> >> +++ b/tests/system-traffic.at
> >> @@ -4016,7 +4016,7 @@ NS_CHECK_EXEC([at_ns1], [ip neigh add 10.1.1.254
> >> lladdr f0:00:00:01:01:FE dev p1
> >> NS_CHECK_EXEC([at_ns0], [ip route add default nexthop via 10.1.1.254])
> >> NS_CHECK_EXEC([at_ns1], [ip route add default nexthop via 10.1.1.254])
> >>
> >> -NETNS_DAEMONIZE([at_ns0], [nc -l -k > /dev/null], [nc0.pid])
> >> +NETNS_DAEMONIZE([at_ns0], [nc -l -k 1234 > /dev/null], [nc0.pid])
> >>
> >> AT_DATA([flows.txt], [dnl
> >> table=0,priority=10 ip action=ct(table=1)
> >> @@ -4052,7 +4052,7 @@ table=21 ip,nw_dst=10.1.1.2
> >> action=set_field:f0:00:00:01:01:02->eth_
> >> AT_CHECK([ovs-ofctl --bundle add-flows br0 flows.txt])
> >>
> >> dnl non-FIP case
> >> -NS_CHECK_EXEC([at_ns1], [echo "foobar" |nc $NC_EOF_OPT 10.1.1.1])
> >> +NS_CHECK_EXEC([at_ns1], [echo "foobar" |nc $NC_EOF_OPT 10.1.1.1 1234])
> >> OVS_WAIT_UNTIL([[ovs-appctl dpctl/dump-conntrack | sed -e
> >> 's/port=[0-9]*/port=<cleared>/g' -e 's/id=[0-9]*/id=<cleared>/g' |
> >> grep "tcp,orig=(src=10.1.1.2,dst=10.1.1.1,sport=<cleared>,dport=<cleared>),reply=(src=10.1.1.1,dst=10.1.1.2,sport=<cleared>,dport=<cleared>),protoinfo=(state=TIME_WAIT)"
> >> ]])
> >> @@ -4061,7 +4061,7 @@ dnl Check that the full session ends as expected
> >> (i.e. TIME_WAIT). Otherwise it
> >> dnl means the datapath didn't process the ct_clear action. Ending in SYN_RECV
> >> dnl (OVS maps to ESTABLISHED) means the initial frame was committed, but not a
> >> dnl second time after the FIP translation (because ct_clear didn't occur).
> >> -NS_CHECK_EXEC([at_ns1], [echo "foobar" |nc $NC_EOF_OPT 10.254.254.1])
> >> +NS_CHECK_EXEC([at_ns1], [echo "foobar" |nc $NC_EOF_OPT 10.254.254.1 1234])
> >
> > Are you saying it works for you with the explicit port?
> > If not, can you show the error?
> >
> > FWIW, nmap-ncat defaults to 31337. I guess other nc implementations may
> > be different.
>
> The error message:
>
> 83: conntrack - floating IP FAILED
> (system-traffic.at:4055)
> +This is nc from the netcat-openbsd package. An alternative nc is available
> +in the netcat-traditional package.
> +usage: nc [-46bCDdhjklnrStUuvZz] [-I length] [-i interval] [-O length]
> + [-P proxy_username] [-p source_port] [-q seconds] [-s source]
> + [-T toskeyword] [-V rtable] [-w timeout] [-X proxy_protocol]
> + [-x proxy_address[:port]] [destination] [port]
> ./system-traffic.at:4055: exit code was 1, expected 0
>
> on my system (ubuntu 1604), it is using the BSD
> tests/atlocal.in: NC_EOF_OPT="-q 1 -w 5"
There are other test cases that use NC_EOF_OPT, so I expect those are
failing for you as well. Can you verify?
Indeed, OpenBSD nc(1) man page does not list the -q option. Nor do the
other BSDs. I don't know why that option is being used. The first
evidence of it's usage seems to be 9ac0aadab9f9 ("conntrack: Add support
for NAT.").
I'll look into this and possibly avoid using NC_EOF_OPT entirely.
More information about the dev
mailing list