*Hi Justin,* Yes they are built into the kernel, and I inserted them now (manually by insmod)
Now, they are added in module list. cadlab:~/Documents # lsmod | grep cls_basic cls_basic 5326 0 cadlab:~/Documents # lsmod | grep sch_ingress sch_ingress 2386 2 cadlab:~/Documents # lsmod | grep act_police act_police 5101 0 Then, I again configured it. cadlab:~/Documents # tc qdisc add dev eth0 ingress cadlab:~/Documents # ovs-vsctl set Interface vif1.0 ingress_policing_rate=1000 cadlab:~/Documents # ovs-vsctl set Interface vif1.0 ingress_policing_burst=100 But still, it's not showing any change in Netperf output. root@mohit-HP-Compaq-nx6325-EQ422AV:~# netperf -H 10.112.10.51 TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.112.10.51 (10.112.10.51) port 0 AF_INET : demo Recv Send Send Socket Socket Message Elapsed Size Size Size Time Throughput bytes bytes bytes secs. 10^6bits/sec 87380 30000000 30000000 12.32 93.37 I still see 93 Mbps.. :( * ---------------------------- Thanks & Regards Mohit Dhingra +919611190435* On 17 February 2012 23:55, Justin Pettit <jpet...@nicira.com> wrote: > Did you follow the instructions for enabling support for ingress policing > in "INSTALL.Linux"? > > For optional support of ingress policing, you must enable kernel > configuration options NET_CLS_BASIC, NET_SCH_INGRESS, and > NET_ACT_POLICE, either built-in or as modules. (NET_CLS_POLICE is > obsolete and not needed.) > > If you're running an older version of OVS, the instructions are as follows: > > For optional support of ingress policing, you must enable kernel > configuration options NET_CLS_ACT, NET_CLS_U32, NET_SCH_INGRESS, > and NET_ACT_POLICE, either built-in or as modules. > (NET_CLS_POLICE is obsolete and not needed.) > > I don't see some of the modules in your list below. Are they built into > your kernel? > > --Justin > > > On Feb 17, 2012, at 10:11 AM, Mohit Dhingra wrote: > > > Hello, > > > > I looked at the debugging tip, it says, check whether your kernel is > built with "ingress qdisc" or not. I am not really sure of that. I searched > for it, and found "tc qdisc add dev eth0 ingress" to enable it. I ran it, > again configured the same QoS parameters for vif2.0, but no change in > netperf output. > > > > Here is the mod list, if it helps. > > cadlab:~/Documents # lsmod > > Module Size Used by > > sch_ingress 2386 0 > > usb_storage 53657 0 > > uas 10350 0 > > iptable_filter 1946 0 > > ip_tables 22302 1 iptable_filter > > x_tables 27449 2 iptable_filter,ip_tables > > openvswitch_mod 87483 1 > > loop 20091 4 > > bridge 92152 0 > > stp 2331 1 bridge > > llc 6103 2 bridge,stp > > fuse 82307 3 > > usbbk 25163 0 > > gntdev 8554 6 > > netbk 42268 0 [permanent] > > blkbk 29670 0 [permanent] > > blkback_pagemap 2806 1 blkbk > > blktap 126872 2 [permanent] > > domctl 6437 2 blkbk,blktap > > xenbus_be 4261 4 usbbk,netbk,blkbk,blktap > > evtchn 6002 7 > > edd 9664 0 > > snd_pcm_oss 53007 0 > > snd_mixer_oss 19233 1 snd_pcm_oss > > snd_seq 66355 0 > > snd_seq_device 7770 1 snd_seq > > tcp_bic 3363 27 > > dm_mod 85824 0 > > joydev 12518 0 > > snd_hda_codec_realtek 356107 1 > > snd_hda_intel 29349 4 > > snd_hda_codec 107570 2 snd_hda_codec_realtek,snd_hda_intel > > snd_hwdep 7612 1 snd_hda_codec > > snd_pcm 106732 3 snd_pcm_oss,snd_hda_intel,snd_hda_codec > > pcspkr 2190 0 > > snd_timer 27510 2 snd_seq,snd_pcm > > serio_raw 5222 0 > > sr_mod 16493 0 > > snd 83254 18 > snd_pcm_oss,snd_mixer_oss,snd_seq,snd_seq_device,snd_hda_codec_realtek,snd_hda_intel,snd_hda_codec,snd_hwdep,snd_pcm,snd_timer > > soundcore 8494 1 snd > > sg 33121 0 > > firewire_ohci 31303 0 > > i7core_edac 20315 0 > > firewire_core 66559 1 firewire_ohci > > iTCO_wdt 12566 0 > > iTCO_vendor_support 3118 1 iTCO_wdt > > edac_core 50431 1 i7core_edac > > i2c_i801 10856 0 > > snd_page_alloc 9473 2 snd_hda_intel,snd_pcm > > e1000e 203595 0 > > crc_itu_t 1779 1 firewire_core > > ext4 398734 3 > > jbd2 88998 1 ext4 > > crc16 1747 1 ext4 > > usbhid 50613 0 > > hid 87441 1 usbhid > > uhci_hcd 29265 0 > > nouveau 610369 2 > > ttm 76417 1 nouveau > > drm_kms_helper 36694 1 nouveau > > ehci_hcd 61639 0 > > drm 230792 4 nouveau,ttm,drm_kms_helper > > i2c_algo_bit 6342 1 nouveau > > usbcore 213666 7 > usb_storage,uas,usbbk,usbhid,uhci_hcd,ehci_hcd > > i2c_core 32282 5 > i2c_i801,nouveau,drm_kms_helper,drm,i2c_algo_bit > > video 15865 1 nouveau > > output 2727 1 video > > button 6797 1 nouveau > > xenblk 26089 0 > > cdrom 42800 2 sr_mod,xenblk > > xennet 36109 0 > > fan 3215 0 > > processor 36333 0 > > ata_generic 3995 0 > > thermal 14850 0 > > thermal_sys 17334 4 video,fan,processor,thermal > > hwmon 2648 1 thermal_sys > > > > cadlab:~/Documents # brctl show > > bridge name bridge id STP enabled interfaces > > eth0 0000.7071bc62737a no peth0 > > vif2.0 > > vif5.0 > > > > cadlab:~/Documents # ovs-vsctl show > > d3f38d68-2623-4a60-a169-fa7bf14e6ca4 > > Bridge "eth0" > > Port "vif2.0" > > Interface "vif2.0" > > Port "peth0" > > Interface "peth0" > > Port "vif5.0" > > Interface "vif5.0" > > Port "eth0" > > Interface "eth0" > > type: internal > > > > I configured vif2.0 for 1Mbps, but of no use. Can you please help me out. > > > > ---------------------------- > > Thanks & Regards > > Mohit Dhingra > > +919611190435 > > > > > > On 14 February 2012 09:24, Jesse Gross <je...@nicira.com> wrote: > > On Mon, Feb 13, 2012 at 7:07 PM, Mohit Dhingra <mohitdhing...@gmail.com> > wrote: > > > Hi All, > > > > > > I am following the same instructions as given in > > > http://openvswitch.org/support/config-cookbooks/qos-rate-limiting/ > > > > > > I have exactly the same setup, except for the measurement host, but I > don't > > > see any difference in the netperf results, I have given more details > in the > > > previous post. Any suggestions/input please? > > > > Did you look at the debugging tips at the bottom of the page? > > > > > Hi Jesse, > > > Ok, So that means hardware should be configurable enough, that it > > > understands VMs, and QoS for each VM? Is there any way to set egress > QoS on > > > the NIC? > > > > It's not a hardware feature. It's software configured on the NIC > > device. As I mentioned before, there's more information in the > > ovs-vsctl man page. It tends to work better than ingress policing but > > you should check that that's at least loaded and running correctly > > first since it's easier to configure. > > > > _______________________________________________ > > discuss mailing list > > discuss@openvswitch.org > > http://openvswitch.org/mailman/listinfo/discuss > >
_______________________________________________ discuss mailing list discuss@openvswitch.org http://openvswitch.org/mailman/listinfo/discuss
