Re: [ovs-discuss] Simple test-setup: lots of TCP Retransmission, TCP Dup ACK

2012-01-08 Thread Dave Whitla
d/70-ethtool.rules This is cleaner than rc.local and the ethtool support in the network up/down scripts is completely broken on RHEL (it only supports setting duplex and speed). Dave On 09/01/2012, at 7:01 AM, Hans van Kranenburg wrote: > On 1/8/12 10:33 AM, Dave Whitla wrote: >> &g

Re: [ovs-discuss] Simple test-setup: lots of TCP Retransmission, TCP Dup ACK

2012-01-08 Thread Dave Whitla
Hi Jesse, The issue persists in all kernels on the 2.6 branch. I've not experienced it with 3.x kernels but my scripts have disabled TX checksumming since I first came across the issue. The problem is that many people run VMs to support "legacy" systems on new high-capacity hardware or in prod

[ovs-discuss] Fwd: Newbie question

2012-01-05 Thread Dave Whitla
Forgot to CC the list. Begin forwarded message: > From: Dave Whitla > Date: 6 January 2012 12:14:40 PM GMT+10:00 > To: Tyrion Lannister > Subject: Re: [ovs-discuss] Newbie question > > You must assign an IP to your bridge. It isn't migrated automagically. If y

Re: [ovs-discuss] H/w offloading list for OVS 1.2.2

2011-12-15 Thread Dave Whitla
ecked) the latest available from our repos. It was less effort to patch that than build my own 1.3 distribution packages only to have them conflict when 1.3 reaches the Debian repos. Dave On 16/12/2011, at 11:52 AM, Jesse Gross wrote: > On Thu, Dec 15, 2011 at 5:31 PM, Dave Whitla wrote: &g

Re: [ovs-discuss] H/w offloading list for OVS 1.2.2

2011-12-15 Thread Dave Whitla
Sorry Sheili, I don't understand exactly what you are asking? Also please reply to the list rather than directly to me. Thanks, Dave On 15/12/2011, at 7:44 PM, Sheili Mittal wrote: > Hi Dave, > > My question was can we enable /disable this on virtual machine also on > OVS? > ___

Re: [ovs-discuss] H/w offloading list for OVS 1.2.2

2011-12-15 Thread Dave Whitla
disable hardware tx checksum offload on F5's LTM VE. I can only comment from experience - if that's not good enough then I don't know what else I can offer. Dave On 16/12/2011, at 11:27 AM, Jesse Gross wrote: > On Wed, Dec 14, 2011 at 11:21 PM, Dave Whitla > wrote

Re: [ovs-discuss] H/w offloading list for OVS 1.2.2

2011-12-14 Thread Dave Whitla
Forgot to say - see the effects using ethtool -k eth0 before and after. Dave On 15/12/2011, at 5:36 PM, Dave Whitla wrote: > If you enter the following in a shell you will see that disabling tx csum > offload implies that other features are disabled also: > > ethtool -K eth0 tx o

Re: [ovs-discuss] H/w offloading list for OVS 1.2.2

2011-12-14 Thread Dave Whitla
? > > Regards, > Sheili > > -Original Message- > From: Dave Whitla [mailto:dave.whi...@ocean.net.au] > Sent: 15 December 2011 12:51 > To: Jesse Gross > Cc: Sheili Mittal; discuss@openvswitch.org > Subject: Re: [ovs-discuss] H/w offloading list for OVS 1.2.2 &g

Re: [ovs-discuss] H/w offloading list for OVS 1.2.2

2011-12-14 Thread Dave Whitla
Sheili, A particular case to be aware of is when combining Linux bridging with 802.1q vlans. Hardware tx checksum offload does not work properly in this scenario, so it is advisable to disable tx checksum offloading in virtual machines which connect to vifs mapped to a vlan or a trunk port to

Re: [ovs-discuss] discuss Digest, Vol 30, Issue 5

2011-12-14 Thread Dave Whitla
When you "enable" tx checksum on the vm you are actually enabling "tx checksum offloading". That is the vm delegates this to the switch. OVS doesn't so your syn is dropped by the receiver. On 14/12/2011, at 3:52 PM, Jesse Gross wrote: > 2011/12/7 likunyun : >> >> >>> From: je...@nicira.com

Re: [ovs-discuss] Plans to support 3.1 kernel?

2011-12-11 Thread Dave Whitla
I have a working patch for Wheezy which I am currently using in our test environment. It selectively applies the Git commits in support of the 3.1 kernel made since 1.2.2. To use the patch: apt-get install linux-source-3.1 linux-headers-3.1.0-1-all openvswitch-datapath-source cd /usr/src tar j