On Thu, Aug 5, 2021 at 6:26 PM Yi Yang (杨燚)-云服务集团 wrote:
>
> Hi, folks
>
>
>
> I’m changing Openstack Neutron to use openflow-based pipeline to implement
> qrouter and floating IP, everything is ok when two VMs (which are in two
> different compute nodes and two subnets but same network, so
Hi, folks
I’m changing Openstack Neutron to use openflow-based pipeline to implement
qrouter and floating IP, everything is ok when two VMs (which are in two
different compute nodes and two subnets but same network, so vlan tag is same),
but NORMAL action will drop packets if they are on
During a netlink transaction, in case of replies of type NLMSG_ERROR,
the current behavior includes the translation of the error number
received into a string that describes the error code.
Netlink replies may carry a more descriptive error message, and
although it is possible to read those
On Thu, Aug 5, 2021 at 9:09 AM Mark Gray wrote:
>
> * The "mac" column may also contain IP addresses.
> * Logical router ports and logical switch ports must not
> share names.
>
> Signed-off-by: Mark Gray
Thanks for fixing the documentation errors. As this patch is
independent of the first
On Thu, Aug 5, 2021 at 9:09 AM Mark Gray wrote:
>
> Currently, ovn-controller generates N x V OpenFlow
> hairpin flows where:
>
> N = number of datapaths
> V = number of LB VIPs
>
> In a scale setup with 120 nodes, 15k VIPs, and 3 Protocols,
> this can generate 5.4M OpenFlows in the
On Mon, Aug 2, 2021 at 6:59 AM Lorenzo Bianconi
wrote:
>
> Check if the nexthop is reachable using router interfaces configuring
> ECMP routes. DDlog northd implementation is already checking the
> condition above.
>
> https://bugzilla.redhat.com/show_bug.cgi?id=1978796
>
> Signed-off-by: Lorenzo
On 8/5/21 4:30 PM, Mark Gray wrote:
> On 05/08/2021 13:57, Ilya Maximets wrote:
>> 'sent_node' is initialized to all zeroes by xzalloc(), but
>> HMAP_NODE_NULL is not all zeroes. hmap_node_is_null() is used
>> to detect if the node is valid, but it will fail and cause
>> segmentation fault on
Eelco Chaudron writes:
> This patch adds a general way of viewing/configuring datapath
> cache sizes. With an implementation for the netlink interface.
>
> The ovs-dpctl/ovs-appctl show commands will display the
> current cache sizes configured:
>
> ovs-dpctl show
> system@ovs-system:
>
Eelco Chaudron writes:
> This patch adds cache usage statistics to the output:
>
> $ ovs-dpctl show
> system@ovs-system:
> lookups: hit:24 missed:71 lost:0
> flows: 0
> masks: hit:334 total:0 hit/pkt:3.52
> cache: hit:4 hit rate:4.21%
> port 0: ovs-system (internal)
> port 1:
Hello,
I was having a look at the calendar and realized that tomorrow (6
August) is our scheduled soft freeze date for OVN 21.09. Since I did a
poor job of notifying people about this, I think it is OK for us to
extend the soft freeze by one week. So soft freeze for OVN 21.09 will be
13
On Thu, Aug 5, 2021 at 11:15 AM Mark Gray wrote:
>
> On 04/08/2021 10:30, Xavier Simonart wrote:
> > When ADD_NAMESPACES or ADD_VETH were executed with variables
> > within the arguments (e.g. ADD_NAMESPACES(sw1-p$i)), the macros were
> > not expanded properly, resulting in bad on_exit calls.
> >
Hi,
I’m trying to implement multiple routing tables support for Logical Routers and
met some difficulties, need help/advice.
How I see it can be used by administrators:
1. In LRP’s options field is added key route_table with the routing table’s
name - any string.
2. When adding new logical
> The commit 1c9e46ab5 removed the outport match from the lflows, which
> leads to a problem for gateway routers that have multiple ports
> configured with different MTUs. For example, R0 has port P1, P2 and P3.
> P2 and P3 both have gateway_mtu configured: P2 mtu = 1400, P3 mtu = 1500.
> Below
On Thu, Aug 5, 2021, at 15:53, anton.iva...@cambridgegreys.com wrote:
> From: Anton Ivanov
>
> time_poll() makes an excessive number of time_msec() calls
> which incur a performance penalty.
>
> 1. Avoid time_msec() call for timeout calculation when time_poll()
> is asked to skip poll()
>
> 2.
On Thu, Aug 5, 2021, at 15:53, anton.iva...@cambridgegreys.com wrote:
> From: Anton Ivanov
>
> If we are not obtaining any useful information out of the poll(),
> such as is a fd busy or not, we do not need to do a poll() if
> an immediate_wake() has been requested.
>
> This cuts out all the
On Wed, Aug 4, 2021 at 1:30 AM Lorenzo Bianconi
wrote:
>
> [...]
> >
> > I think REGBIT_EGRESS_LOOPBACK check is required so that the injected
icmp4
> > packet generated from ovn-controller is skipped in this stage.
> > Otherwise there will
> > be recursion of the packet.
> >
> > @Lorenzo
Good point, OVS is unable to parse them. I don't see anywhere in the OVS
code that relies on these fields being printed.
I could replace them with the default 'tp_src/dst' in this workaround, so
as not to break any scripts.
That being said, I am not sure how these fields would currently even end
On Tue, Aug 3, 2021 at 10:59 AM Numan Siddique wrote:
>
> On Tue, Aug 3, 2021 at 1:55 PM Han Zhou wrote:
> >
> > On Tue, Aug 3, 2021 at 10:14 AM Numan Siddique wrote:
> > >
> > > On Thu, Jul 29, 2021 at 4:03 AM Han Zhou wrote:
> > > >
> > > > From: Ankur Sharma
> > > >
> > > > By default, OVN
On Thu, Aug 5, 2021 at 10:55 AM Mark Gray wrote:
>
> On 04/08/2021 09:59, Xavier Simonart wrote:
> > When running load-balancing test, tcpdump is executed in background.
> > When the backround tcpdump stops, it sends to stderr a message
> > (such as number of packets captured) which might be
On 28/07/2021 18:58, Lorenzo Bianconi wrote:
> Clear ipv6_ra_pd_list option in port_binding option column in order to
> properly update IPv6 prefix info.
>
> https://bugzilla.redhat.com/show_bug.cgi?id=1983862
>
> Fixes: e3a398e914 ("controller: Add ipv6 prefix delegation state machine")
>
On 28/07/2021 18:58, Lorenzo Bianconi wrote:
> Introduce some log messages in IPv6 Prefix delegation processing useful
> for debugging
>
> Signed-off-by: Lorenzo Bianconi
> ---
> controller/pinctrl.c | 25 -
> 1 file changed, 24 insertions(+), 1 deletion(-)
>
> diff
From: Ankur Sharma
By default, OVN support only one DGP (distributed gateway port) per
logical router. While a single DGP port suffices for most of the North
South connectivity, there are requirements where a logical router could
be connected to multiple external networks and based on routing
The commit 1c9e46ab5 removed the outport match from the lflows, which
leads to a problem for gateway routers that have multiple ports
configured with different MTUs. For example, R0 has port P1, P2 and P3.
P2 and P3 both have gateway_mtu configured: P2 mtu = 1400, P3 mtu = 1500.
Below lflows are
V2:
- Converted to a series and added patch-1 that fixes a problem of an upstream
commit, which was found during rebasing.
- Rebased on upstream.
- Addressed comments of V1 from Numan.
Ankur Sharma (1):
ovn-northd: Multiple distributed gateway port support.
Han Zhou (1):
ovn-northd: Match
On 04/08/2021 10:30, Xavier Simonart wrote:
> When ADD_NAMESPACES or ADD_VETH were executed with variables
> within the arguments (e.g. ADD_NAMESPACES(sw1-p$i)), the macros were
> not expanded properly, resulting in bad on_exit calls.
> This caused multiples tests (e.g. 2 LSs IGMP and MLD) to be
On 04/08/2021 10:01, Xavier Simonart wrote:
> This test sends requests to LB and checks that each server receives at
> least one request. However, even if 20 requests are sent to 3 servers,
> there was a possibility that one server did not get any request (0.3%).
Could you add a link to a failing
On 04/08/2021 09:59, Xavier Simonart wrote:
> When running load-balancing test, tcpdump is executed in background.
> When the backround tcpdump stops, it sends to stderr a message
> (such as number of packets captured) which might be catched by the
> following command (wget) causing the test to
When OVN is linked with an appropriate plugging implementation,
CMS can request OVN to plug individual lports into the local
Open vSwitch instance.
The port and instance record will be maintained during the lifetime
of the lport and it will be removed on release of lport.
TODO: The functions
To be able to effectively remove ports previously plugged by us
we need to look up ports by interface records.
Signed-off-by: Frode Nordahl
---
controller/ovn-controller.c | 13 +
1 file changed, 13 insertions(+)
diff --git a/controller/ovn-controller.c
This module contains the infrastructure for registering and
instantiating plugging classes which may be hosted inside or
outside the core OVN repository. The data structures and functions
for interacting with these plugging classes also live here.
Signed-off-by: Frode Nordahl
---
Up until now the controller patch module has been the only
consumer of functions to maintain OVS ports and interfaces.
With the introduction of infrastructure for plugging providers
these functions will also be consumed by the controller binding
module.
As such we introduce a new module called
Make use of the common functions for maintaining OVS ports.
Signed-off-by: Frode Nordahl
---
controller/patch.c | 39 ++-
1 file changed, 6 insertions(+), 33 deletions(-)
diff --git a/controller/patch.c b/controller/patch.c
index a661025da..b26e93e53 100644
ovn-northd will fill this column with UUID of Chassis referenced
in Logical_Switch_Port options:requested-chassis when
options:plug-type is defined.
Signed-off-by: Frode Nordahl
---
northd/ovn-northd.c | 31 +++
ovn-nb.xml | 38
This RFC series is in response to the discussion in a previous RFC patch
[0].
Further work is detailed in TODOs in some of the commit messages as well
as completing unit- and functional- tests and extending the build system
to allow adding external plug library at build time.
Please have a look
On 05/08/2021 13:57, Ilya Maximets wrote:
> 'sent_node' is initialized to all zeroes by xzalloc(), but
> HMAP_NODE_NULL is not all zeroes. hmap_node_is_null() is used
> to detect if the node is valid, but it will fail and cause
> segmentation fault on attempt to remove the non-existent node
>
From: Anton Ivanov
If we are not obtaining any useful information out of the poll(),
such as is a fd busy or not, we do not need to do a poll() if
an immediate_wake() has been requested.
This cuts out all the pollfd hash additions, forming the poll
arguments and the actual poll() after a call
From: Anton Ivanov
time_poll() makes an excessive number of time_msec() calls
which incur a performance penalty.
1. Avoid time_msec() call for timeout calculation when time_poll()
is asked to skip poll()
2. Reuse the time_msec() result from deadline calculation for
last_wakeup and timeout
* The "mac" column may also contain IP addresses.
* Logical router ports and logical switch ports must not
share names.
Signed-off-by: Mark Gray
---
ovn-sb.xml | 21 ++---
1 file changed, 6 insertions(+), 15 deletions(-)
diff --git a/ovn-sb.xml b/ovn-sb.xml
index
Currently, ovn-controller generates N x V OpenFlow
hairpin flows where:
N = number of datapaths
V = number of LB VIPs
In a scale setup with 120 nodes, 15k VIPs, and 3 Protocols,
this can generate 5.4M OpenFlows in the OFTABLE_CT_SNAT_FOR_VIP
table with the following form:
table=70,
'sent_node' is initialized to all zeroes by xzalloc(), but
HMAP_NODE_NULL is not all zeroes. hmap_node_is_null() is used
to detect if the node is valid, but it will fail and cause
segmentation fault on attempt to remove the non-existent node
from the hash map. This can happen if client
On 8/4/21 1:53 PM, Ferriter, Cian wrote:
>
>> -Original Message-
>> From: Ilya Maximets
>> Sent: Tuesday 3 August 2021 18:37
>> To: ovs-dev@openvswitch.org; Stokes, Ian
>> Cc: Flavio Leitner ; Amber, Kumar
>> ; Ferriter, Cian
>> ; Van Haaren, Harry ;
>> Ilya Maximets
>>
>> Subject:
Hi, do you know any infomations about integrating ovs/ovn with evpn/frr.
Thanks.
At 2021-06-25 00:32:46, "Ihar Hrachyshka" wrote:
>It is supported but with a number of limitations. Specifically, the
>number of switches, and ports per switch, is limited to 2^11 when
>VXLAN is
On 7/22/21 3:25 PM, Kevin Traynor wrote:
> The 'Port/Rx Queue Assigment to PMD Threads' section has
> expanded over time and now includes info about stats/commands,
> manual pinning and different options for OVS assigning Rxqs to
> PMDs.
>
> Split them into different sections with sub-headings
43 matches
Mail list logo