From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 0AF40C433FE for ; Wed, 5 Oct 2022 19:55:11 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230175AbiJETzJ (ORCPT ); Wed, 5 Oct 2022 15:55:09 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:55204 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230078AbiJETzI (ORCPT ); Wed, 5 Oct 2022 15:55:08 -0400 Received: from mail-0201.mail-europe.com (mail-0201.mail-europe.com [51.77.79.158]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 564944623F for ; Wed, 5 Oct 2022 12:55:07 -0700 (PDT) Date: Wed, 05 Oct 2022 19:54:53 +0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=protonmail.com; s=protonmail3; t=1664999703; x=1665258903; bh=ygjTo8reVx2XjO/O5ENZntwsGUDohh4oiG5NtkdsuD4=; h=Date:To:From:Subject:Message-ID:Feedback-ID:From:To:Cc:Date: Subject:Reply-To:Feedback-ID:Message-ID; b=zaGRQr9qEBFEBjk4kZeoyQ3abMdkvaPcTzDN7KWtcWMSs7NkIK+yCuyERgHrOZ1CC FqN/ZpS/Fx3OHf7gsRvl1DdpEeZjunmZh9yOQ1Z2ioc/VfIuz7L+SBipcRE27anX29 ityCVxfNAXrp6lPk9yJyhDUm7HHT2MwDB9mVACEOBHUgAc0GZoGYrB/M7/+WeWgJHA CTaz5OJ/DtWGv9PPQJoedKcJ1LDURNO3Tq7ZSYMJB0u/Zs+8DfX3Ou0TYvCtXdb2nT FPQBeFj/SJ0U6GxlSnmSK8XKIVz7z8Nhr+TSiJdlRofNpEiqfoUxSTTLXUlAclyC06 bf8yuPgIm17nA== To: "kvm@vger.kernel.org" From: kvmuser99 Subject: High pings when using SR-IOV w\ macvtap & virtio Message-ID: Feedback-ID: 56853544:user:proton MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: kvm@vger.kernel.org KVM Team, I'm going to the mailing lists because I'm at a loss.=C2=A0 I have thousand= s of Linux (CentOS 7) VMs deployed and a very small percent of them get int= o a state when there's some amount of network IO the guests will constantly= clock high ping times (> 100 ms local network) and performance will suffer= .=C2=A0 Some hosts will always clock the high ping times, other times runni= ng a CLI speedtest will cause the condition to surface. Rebooting the guest does not solve the issue but a reboot of the host does = for awhile.=C2=A0 Running something such as the following helps in most cas= es.=C2=A0 For reference eth0 is the physical interface from which all the V= Fs come from. =3D=3D=3D =C2=A0 126 =C2=A0ethtool -L eth0 combined 16 <-- Arbitrary number=C2=A0=20 128 =C2=A0ethtool -L eth0 combined 4 <--- Number of CPU reserved for the ho= st (it as previously set to this) =3D=3D=3D In most of the scenarios pings even to a 'bridge' interface which just hand= les guest\host communication also shoot high even though there is low netwo= rk IO on that particular interface. Config is as follows - Queues =3D number of CPU assigned to guest - Affinity is enabled. =20 =3D=3D=3D =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0
=C2=A0 =C2=A0 =3D=3D=3D=3D=3D =3D=3D=3D lshw -c network -businfo pci@0000:03:02.5 =C2=A0eth12 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 network = =C2=A0 =C2=A0 =C2=A0 =C2=A0Ethernet Virtual Function 700 Series ip link show | grep eth12 26: eth12: mtu 1500 qdisc mq state UP mod= e DEFAULT group default qlen 1000110: macvtap30@eth12: mtu 1500 qdisc pfifo_fast state UP mode DEFAULT group default= qlen 5000 =3D=3D=3D=3D Info about the interface =3D=3D=3D [root@HOST~]# ethtool -k eth12Features for eth12: rx-checksumming: on tx-checksumming: on =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-checksum-ipv4: on =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-checksum-ip-generic: off [fixed] =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-checksum-ipv6: on =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-checksum-fcoe-crc: off [fixed] =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-checksum-sctp: on scatter-gather: on =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-scatter-gather: on =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-scatter-gather-fraglist: off [fixed] tcp-segmentation-offload: on =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-tcp-segmentation: on =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-tcp-ecn-segmentation: on =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-tcp6-segmentation: on =C2=A0 =C2=A0 =C2=A0 =C2=A0 tx-tcp-mangleid-segmentation: on udp-fragmentation-offload: off [fixed] generic-segmentation-offload: on generic-receive-offload: off large-receive-offload: off [fixed] rx-vlan-offload: on tx-vlan-offload: on ntuple-filters: off [fixed] receive-hashing: on highdma: on rx-vlan-filter: on [fixed] vlan-challenged: off [fixed] tx-lockless: off [fixed] netns-local: off [fixed] tx-gso-robust: off [fixed] tx-fcoe-segmentation: off [fixed] tx-gre-segmentation: on tx-ipip-segmentation: on tx-sit-segmentation: on tx-udp_tnl-segmentation: on fcoe-mtu: off [fixed] tx-nocache-copy: off loopback: off [fixed] rx-fcs: off [fixed] rx-all: off [fixed] tx-vlan-stag-hw-insert: off [fixed] rx-vlan-stag-hw-parse: off [fixed] rx-vlan-stag-filter: off [fixed] busy-poll: off [fixed] tx-gre-csum-segmentation: on tx-udp_tnl-csum-segmentation: on tx-gso-partial: on tx-sctp-segmentation: off [fixed] rx-gro-hw: off [fixed] l2-fwd-offload: off [fixed] hw-tc-offload: on rx-udp_tunnel-port-offload: off [fixed] =3D=3D=3D=3D Versions =3D=3D=3D cat /etc/redhat-release CentOS Linux release 7.6.1810 (Core) modinfo i40efilename: =C2=A0 =C2=A0 =C2=A0 /lib/modules/3.10.0-957.el7.x86_= 64/kernel/drivers/net/ethernet/intel/i40e/i40e.ko.xz version: =C2=A0 =C2=A0 =C2=A0 =C2=A02.3.2-k license: =C2=A0 =C2=A0 =C2=A0 =C2=A0GPL description: =C2=A0 =C2=A0Intel(R) Ethernet Connection XL710 Network Driver author: =C2=A0 =C2=A0 =C2=A0 =C2=A0 Intel Corporation, retpoline: =C2=A0 =C2=A0 =C2=A0Y rhelversion: =C2=A0 =C2=A07.6 uname -r3.10.0-957.el7.x86_64 /usr/libexec/qemu-kvm -version QEMU emulator version 5.1.0 =3D=3D=3D=3D=3D -kvmuser99