Hi Andreas,
Are you using ifconfig to set the MTU? Those settings can be removed
at any time at XenServer's discretion. You may want to consider using
interface reconfigure / xapi / xe / the UI if you want persistent settings.
-Reid
On Mon, Nov 11, 2013 at 1:39 AM, Andreas Pflug wrote:
> I
You have configured the loop yourself, as you described above.
Broadcast packets are forwarded out all ports except the ingress
port by the NORMAL action (you didn't mention configuring any
flows), and different tunnels are different ports. Since each Host
has 3 ports, the broadcast packets are se
t starting with XEN and all machines in the network are
> configured with 2 NIC’s and STP switches.
>
> ** **
>
> ** **
>
> Best Regards,
>
> ** **
>
> *From:* Reid Price [mailto:rpr...@nicira.com]
> *Sent:* Thursday, September 19, 2013 11:51 PM
>
A few quick comments:
1) You say there are no custom flows, but your ofctl dump flows
show custom flows. In particular, I see 'drop' flows on both that
appear to have been used in the last second before you dumped.
2) A common misconception is that tcpdump is
dumping traffic on 'the bridge' in
Hi Kaja,
Take a look at the man page of ovs-vsctl.
In particular, the controller and manager tables
are probably what you're thinking of.
-Reid
On Wed, Oct 9, 2013 at 10:30 PM, RASOOL, Kaja Mohideen (Kaja Mohideen) <
kaja_mohideen.ras...@alcatel-lucent.com> wrote:
> Hi all,
>
> ** **
>
Hi Mark,
The 'drop' is implicit. If the switch executions no actions,
either from 'actions=', 'actions=drop', or any other sequence
of events that has no output, then the switch will appear
to have dropped the packet, since it was not forwarded
anywhere.
-Reid
On Wed, Oct 9, 2013 at 8:54 AM,
vs-ofctl add-flow xenbr0 in_port=3,actions=output:5
> but it is not right, i test the flow, i find the windows xp can ping the
> firewall success,but can't ping the xenserver or other ip success
> i guess i'm wrong ,so ,i hope someone can help me!
> thanks!
>
>
>
>
&g
Hi there,
I think you will have to describe what you want more carefully, perhaps by
including a diagram of the components and desired vs current behavior.
-Reid
On Tue, Sep 24, 2013 at 7:57 PM, huanghj wrote:
>
>
>
> 原始邮件
> 主题：[ovs-discuss] about ovs
>
> 发件人：huanghj
> 收件
ri, Sep 20, 2013 at 4:45 AM, Alexandru Coseru wrote:
> Hello,
>
> ** **
>
> Can you give some more details please ?
>
> ** **
>
> ** **
>
> Best Regards,
>
> ** **
>
> *From:* Reid Price [mailto:rpr...@nicira.com]
> *Sent:* Thursday, Septem
gt; ** **
>
> Did somebody found a solution for the problem below?
>
> ** **
>
> We want to use openvswitch to a configure a host with 2 nic with STP ON (for
> switch redundacy).
>
> ** **
>
> Best Regards,
>
> Alex
>
> ** **
&
t; the config? The reason I ask is I read it in a blog that it does save it.
>
> On Aug 26, 2013, at 3:40 AM, Reid Price wrote:
>
>> Hi Faisal,
>>
>> Xen only persists its own OVS information, and recreates it from scratch
>> (xapi information) at boot.
>>
Hi Faisal,
Xen only persists its own OVS information, and recreates it from scratch (xapi
information) at boot.
-Reid
On Aug 25, 2013, at 5:41 PM, Faisal Ali wrote:
> Hi,
>
> I am using xenserver 6.2 which has openvswitch version 1.4.6 running.
>
> When I configure ovs using ovs commands
n surely it wouldn't need to be
> shown, since it would be implementation detail. There must be a reason why
> it's exposed. Is it able to be used in some way?
>
> Cheers,
> Tudor.
>
> On 2013-08-21 11:10, Reid Price wrote:
>
> Because it is the "local"
Because it is the "local" port for a bridge. A bridge always has a port
with the same name.
There is a patch [1] by Ben that changes this message from
ovs-vsctl: no port named br-eth1
to
ovs-vsctl: cannot delete port br-eth1 because it is the local port for
bridge br-eth1 (deleting this po
ers some traffic for unkown reasons:
>
> - screen1-wire: shows capturing traffic at ifInt2 (internal VM network)
> receiving ping packets from VLAN 100 which should be unreachable.
>
> - screen2-tcpdump: shows tcpdump (-i eth2) receiving traffic at VM's iface
> eth2 which should
Hi Edu Serra,
The environment you describe seems straightforward and is known to work.
Does the attached image really reflect your topology? There seems to be
an issue somewhere in these statements:
- The very first moment a machine had 2 interfaces up, some traffic
would happen to be rep
For what it's worth, you are not seeing a drop "rule", you are seeing no
action being taken for this packet.
There is one way ([#1], below), and 3 or 4 typical situations in which
packets have no action:
1) Matching a flow that results in no actions (or equivalently,
'actions=drop', which is an a
Hi Prashant,
One note, in your email the dash (-) between replace and flows is a hyphen (−).
I don't know if that is a red herring though.
-Reid
On May 24, 2556 BE, at 22:49, Prashant Chougule wrote:
>
> hi,
>
> I am trying to replace the flows from the table but found an error
Hi sunayana,
The output of the following commands might be helpful to determine if you
followed Justin's suggestion correctly:
*python -c 'import ovs.json'*
*
*
*find / -path '*/ovs/json.py'*
*
*
*python -c 'import sys; print "\n".join(sys.path)'*
-Reid
On Fri, May 3, 2013 at 2:07 AM, Sunay
Having no other way to reach the machine will not make things easier.
You can do silly tricks like
ssh user@remote "; sleep 10; ; sleep 10;
"
where teardown would remove eth0 from br0 and put the IP back on it,
and debugging commands could be things like starting a ping, dumping
the datapath fl
Moved conversation to discuss@ instead of bugs@ - haven't seen anything to
indicate a bug so far.
Just to check, you aren't actually leaving "" as a literal
string, correct? It should be 10.20.30.40 or some real IP. I also notice that
you are missing the create bridge command, I assume that was
Is the script local or remote? You need to run the command atomically from the
remote perspective. This means you need the ssh command to be a single one,
likely along the lines of:
ssh user@remote remote_script
or
ssh user@remote "command && command ; command"
. A local script would not
So how did you come up with this transaction?
One thing you might want to note
echo [What-You-Have-Below]
hangs for me in bash (can't ctrl-C, etc). Try escaping things
properly (add single-quotes perhaps), otherwise you are killing your
machine with an absurd shell expansion.
For example:
For discussions, try google (ovs flow governor) or looking at the
commit history for the files Ben just pointed out.
On Thu, Dec 13, 2012 at 11:18 AM, Ben Pfaff wrote:
>
> ofproto/ofproto-dpif-governor.c
> ofproto/ofproto-dpif-governor.h
>
> On Thu, Dec 13, 2012 at 10:15:12AM -0800, Masoud Moshre
Hi priyanka,
It may be helpful if you replace "I am facing some weird behavior' with a more
traditional:
* Actions
* Expected behavior
* Observed behavior
Also, do you see the same behavior in a simpler setup?
-Reid
On Nov 22, 2555 BE, at 20:59, priyanka singhal <06priyanka.sing...@gmail.co
Hi Marco,
You will want to provide much more information in order to get a useful
response. OVS typically uses near-zero CPU when it is in an idle state.
Perhaps look at the instructions for reporting bugs to get an idea of what
information to collect. I would start with at least
# ovs-dpctl
Hi Arindam,
About 1/3 of the way through the FAQ there is a section entitled:
Configuration Problems
The first Q/A answers several of the questions you have asked and gives
examples.
You might also find luck with the mailing list archive, which is almost
certainly
the top block of results if y
Hello Ryan Spammable :)
On Tue, Jul 24, 2012 at 11:16 AM, Ryan White
wrote:
> Hi all,
>
> I connect two ovs machines to a physical openflow switch which is
> connected to another physical switch.
>
> The controller (beacon, connected to one of the physical switches)
> indicates that all physical
Hi Yimin,
Take a look at the information in the REPORTING_BUGS section for things to help
you get a response. For this particular issue, I would look at the following:
- assign ips to br0 on each bridge
- make sure you can ping between bridges
(if not, your connection is probably not what you thi
"action=NORMAL,goto
> table2", where table2 contains egress flow entries? I am not sure
> whether after NORMAL action is done, whether packet is already
> forwarded and goto table2 will not have no packet to deal with.
>
>
> Thanks!
> Yimin
>
>
> On Fri, J
Hi Yimin,
The behavior you describe would emerge from the flows you choose to
utilize, and may not necessarily be a single modification. One way to
think about this is as a side effect of the flow table being processed a
single time. Since the matching only occurs on the packet entering
the swit
Hi Yimin,
You are using nw_dst as a flow match criteria, but nw_dst only makes sense
in the context of IP. Try doing
ovs-ofctl add-flow br0 ip,nw_dst=11.1.1.5,actions=output:3
Note that "ip" is simply shorthand for "dl_type=0x0800".
I've seen this question go through (and be answered) in the
Missed this response, sorry. Thanks Robin for the assistance and bringing
this back near the top of my emails =)
On Sun, May 20, 2012 at 7:39 AM, Salman Malik wrote:
> Thanks Reid, please find the inline comments.
>
>
> --
> From: r...@nicira.com
> Date: Sun, 20 May
Some options:
* Check to see if eth1 exists =)
* Try collecting some of the information listed in REPORTING_BUGS, and
posting again
* I believe it's recommended to use ovs-vsctl for this; 1.1.1 is almost a
year old though, didn't check. man pages would tell you
-Reid
On Thu, May 24, 2012 at 6
On Sat, May 19, 2012 at 3:42 PM, Salman Malik wrote:
> Thanks a lot Reid. That was one awesome explanation. Few more questions:
>
> 1) In the first figure VMs can not communicate with the outside world
> (except using NAT) ?
>
Yup. I'm not sure what NAT would mean in this context, all tap1/2 t
I drew a little ascii art, does this clarify things?
Hypervisor
++
| |
+-- | -- iface <--+
One thing I have seen in the past: if ethX has an ip as well as brX, it usually
means that routing will try to send traffic out ethX. That traffic will be
dropped, since brX is now the one hooked up to the ip stack (not ethX).
On May 17, 2555 BE, at 20:49, Robin Wang wrote:
> eth1, sorry for t
Make sure you have removed the old IP addresses from ethX as well
On Thu, May 3, 2012 at 1:37 AM, 星言 墨虚 wrote:
> I have assigned IP address to br0, and let OpenvSwitch to connect
> controller. I can find the ARP Request packet and ARP Reply packet between
> OpenvSwitch and NOX by wireshark. But
Hi Mike,
Yep, the switch does initiate the connection.
If you check the man page for ovs-vsctl (search for set-controller), you can
see that
the controller target should include a protocol, and be of one of these
forms:
ssl:ip[:port]tcp:ip[:port]unix:file
Which one is your controller?
Hi Jen-Wei,
My understanding was that if you want to simply pass all VLANs through
the switch, you can make all the relevant ports (in your case, the two
connected to the servers) trunk ports. Try putting something like
switch trunk port configuration
(or the equivalent) into your favorite
Attempted to answer your question inline.
As a side note: I'm quite not sure what you mean by "br0.10 and br0.20 are
two interfaces
with vlan 10 and 20", but is it possible that you are just double-tagging
your packets (once
in the flow rule action, once on the port), or not even hitting your rul
Quick add-on to Justin's answer, in case it wasn't implied:
ovs-vsctl *set-controller* [ssl|tcp]:
-Reid
2011/1/18 Justin Pettit
> If you have all the pieces installed, you can just configure it as you
> would any OpenFlow controller in OVS:
>
>ovs-vsctl ssl:
>
> or, if it's unenc
Hi Derek, comment inline
On Wed, Oct 13, 2010 at 10:51 PM, Derek Cormier wrote:
> One thing I've noticed is that when I use vswitchd, it uses some kind of
> default flow table in unix:/usr/local/var/run/openvswitch/vbr0.mgmt. I tried
> deleting the flows so that pinging from one host to another
Trimmed, see inline response
On May 14, 2010, at 15:59, Jesse Gross wrote:
On Fri, May 14, 2010 at 5:58 AM, Axle Judd wrote:
On Thu, May 13, 2010 at 4:08 AM, Axle Judd wrote:
What i want to accomplish use this single bridge with 6 aggregated
NICS
What i want to do is create a Network that i
44 matches
Mail list logo
