[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Also affects: linux (Ubuntu Precise) Importance: Undecided Status: New ** Description changed: SRU Justification: Impact: - Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. + Users of the 3.2/3.5/3.8 series kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These upstream patches are necessary to fix the issue: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc f45a5c267da35174e22cec955093a7513dc1623d Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) ** Description changed: SRU Justification: Impact: Users of the 3.2/3.5/3.8 series kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These upstream patches are necessary to fix the issue: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc f45a5c267da35174e22cec955093a7513dc1623d Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace + Another testcase without Openstack: + + * create two vms: (vm1, vm2), install iperf on those machines + * connect vms via an isolated bridge + * measure baseline performance + - iperf -s # on machine 1 + - iperf -t 60 -l 4M -c machine 1 IP # on machine 2 + * create veth pairs between vms using attached script: + - ./setup-lp1201869.sh vm1 + - ./setup-lp1201869.sh vm2 + * In the VM's setup static IPs + - sudo ifconfig eth0 10.10.10.1/24 up #vm1 + - sudo ifconfig eth0 10.10.10.2/24 up #vm2 + * measure performance now + * we expect this to be close to the existing performance + -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) ** Changed in: linux (Ubuntu Precise) Assignee: (unassigned) = Chris J Arges (arges) ** Changed in: linux (Ubuntu Precise) Status: New = In Progress ** Changed in: linux (Ubuntu Precise) Importance: Undecided = Medium ** Attachment added: setup-lp1201869.sh https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+attachment/4043420/+files/setup-lp1201869.sh ** Summary changed: - poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels + poor networking throughput through veth interfaces ** Description changed: SRU Justification: Impact: Users of the 3.2/3.5/3.8 series kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These upstream patches are necessary to fix the issue: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc f45a5c267da35174e22cec955093a7513dc1623d + + Minimally 8093315a91340bca52549044975d8c7f673b28a1 provides the proper + features to increase performance dramatically. Testcase: Setup OpenStack Neutron. Test throughput
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
This bug was fixed in the package linux - 3.5.0-41.64 --- linux (3.5.0-41.64) quantal; urgency=low [Brad Figg] * Release Tracking Bug - LP: #1223451 [ Upstream Kernel Changes ] * kernel-doc: bugfix - multi-line macros - LP: #1223920 * Revert zram: use zram-lock to protect zram_free_page() in swap free notify path - LP: #1215513 * x86 thermal: Delete power-limit-notification console messages - LP: #1215748 * x86 thermal: Disable power limit notification interrupt by default - LP: #1215748 * ARM: 7810/1: perf: Fix array out of bounds access in armpmu_map_hw_event() - LP: #1216442 - CVE-2013-4254 * ARM: 7809/1: perf: fix event validation for software group leaders - LP: #1216442 - CVE-2013-4254 * veth: reduce stat overhead - LP: #1201869 * veth: extend device features - LP: #1201869 * veth: avoid a NULL deref in veth_stats_one - LP: #1201869 * veth: fix a NULL deref in netif_carrier_off - LP: #1201869 * veth: fix NULL dereference in veth_dellink() - LP: #1201869 * Bluetooth: Add support for Atheros [0cf3:3121] - LP: #1202477 * efivars: explicitly calculate length of VariableName - LP: #1217745 * xfs: fix _xfs_buf_find oops on blocks beyond the filesystem end - LP: #1151527 - CVE-2013-1819 * drm/i915/lvds: ditch -prepare special case - LP: #1221791 * serial: mxs: fix buffer overflow - LP: #1221791 * fs/proc/task_mmu.c: fix buffer overflow in add_page_map() - LP: #1221791 * af_key: initialize satype in key_notify_policy_flush() - LP: #1221791 * vm: add no-mmu vm_iomap_memory() stub - LP: #1221791 * iwl4965: set power mode early - LP: #1221791 * iwl4965: reset firmware after rfkill off - LP: #1221791 * ASoC: cs42l52: Reorder Min/Max and update to SX_TLV for Beep Volume - LP: #1221791 * can: pcan_usb: fix wrong memcpy() bytes length - LP: #1221791 * ALSA: 6fire: make buffers DMA-able (pcm) - LP: #1221791 * ALSA: 6fire: make buffers DMA-able (midi) - LP: #1221791 * jbd2: Fix use after free after error in jbd2_journal_dirty_metadata() - LP: #1221791 * USB-Serial: Fix error handling of usb_wwan - LP: #1221791 * USB: mos7840: fix big-endian probe - LP: #1221791 * USB: adutux: fix big-endian device-type reporting - LP: #1221791 * USB: ti_usb_3410_5052: fix big-endian firmware handling - LP: #1221791 * m68k/atari: ARAnyM - Fix NatFeat module support - LP: #1221791 * m68k: Truncate base in do_div() - LP: #1221791 * usb: add two quirky touchscreen - LP: #1221791 * USB: mos7720: fix broken control requests - LP: #1221791 * USB: keyspan: fix null-deref at disconnect and release - LP: #1221791 * MIPS: Expose missing pci_io{map,unmap} declarations - LP: #1221791 * microblaze: Update microblaze defconfigs - LP: #1221791 * sound: Fix make allmodconfig on MIPS - LP: #1221791 * sound: Fix make allmodconfig on MIPS correctly - LP: #1221791 * alpha: makefile: don't enforce small data model for kernel builds - LP: #1221791 * MIPS: Rewrite pfn_valid to work in modules, too. - LP: #1221791 * xtensa: fix linker script transformation for .text.unlikely - LP: #1221791 * wusbcore: fix kernel panic when disconnecting a wireless USB-serial device - LP: #1221791 * iwlwifi: dvm: fix calling ieee80211_chswitch_done() with NULL - LP: #1221791 * iwlwifi: pcie: disable L1 Active after pci_enable_device - LP: #1221791 * zd1201: do not use stack as URB transfer_buffer - LP: #1221791 * Hostap: copying wrong data prism2_ioctl_giwaplist() - LP: #1221791 * ARM: at91/DT: fix at91sam9n12ek memory node - LP: #1221791 * drm/i915: Invalidate TLBs for the rings after a reset - LP: #1221791 * libata: apply behavioral quirks to sil3826 PMP - LP: #1221791 * ARM: davinci: nand: specify ecc strength - LP: #1221791 * ARM: 7816/1: CONFIG_KUSER_HELPERS: fix help text - LP: #1221791 * sata_fsl: save irqs while coalescing - LP: #1221791 * xen/events: initialize local per-cpu mask for all possible events - LP: #1221791 * of: fdt: fix memory initialization for expanded DT - LP: #1221791 * zfcp: fix lock imbalance by reworking request queue locking - LP: #1221791 * zfcp: fix schedule-inside-lock in scsi_device list loops - LP: #1221791 * nilfs2: remove double bio_put() in nilfs_end_bio_write() for BIO_EOPNOTSUPP error - LP: #1221791 * nilfs2: fix issue with counting number of bio requests for BIO_EOPNOTSUPP error detection - LP: #1221791 * workqueue: fix possible stall on try_to_grab_pending() of a delayed work item - LP: #1221791 * x86/xen: do not identity map UNUSABLE regions in the machine E820 - LP: #1221791 * jfs: fix readdir cookie incompatibility with NFSv4 - LP: #1221791 * powerpc: Don't Oops when accessing /proc/powerpc/lparcfg
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
This bug was fixed in the package linux - 3.8.0-31.46 --- linux (3.8.0-31.46) raring; urgency=low [ Brad Figg ] * Release Tracking Bug - LP: #1223406 * UBUNTU: [Config] KUSER_HELPERS=y for armhf [ Upstream Kernel Changes ] * Revert cpuidle: Quickly notice prediction failure in general case - LP: #1221794 * Revert cpuidle: Quickly notice prediction failure for repeat mode - LP: #1221794 * Revert zram: use zram-lock to protect zram_free_page() in swap free notify path - LP: #1215513 * x86 thermal: Delete power-limit-notification console messages - LP: #1215748 * x86 thermal: Disable power limit notification interrupt by default - LP: #1215748 * mwifiex: do not create AP and P2P interfaces upon driver loading - LP: #1212720 * ARM: 7810/1: perf: Fix array out of bounds access in armpmu_map_hw_event() - LP: #1216442 - CVE-2013-4254 * ARM: 7809/1: perf: fix event validation for software group leaders - LP: #1216442 - CVE-2013-4254 * veth: reduce stat overhead - LP: #1201869 * veth: extend device features - LP: #1201869 * veth: avoid a NULL deref in veth_stats_one - LP: #1201869 * veth: fix a NULL deref in netif_carrier_off - LP: #1201869 * veth: fix NULL dereference in veth_dellink() - LP: #1201869 * Bluetooth: Add support for Atheros [0cf3:3121] - LP: #1202477 * uvcvideo: quirk PROBE_DEF for Dell SP2008WFP monitor. - LP: #1217957 * usb: dwc3: gadget: don't prevent gadget from being probed if we fail - LP: #1221794 * usb: dwc3: fix wrong bit mask in dwc3_event_type - LP: #1221794 * ASoC: max98088 - fix element type of the register cache. - LP: #1221794 * ata: Fix DVD not dectected at some platform with Wellsburg PCH - LP: #1221794 * Tools: hv: KVP: Fix a bug in IPV6 subnet enumeration - LP: #1221794 * ALSA: usb-audio: 6fire: return correct XRUN indication - LP: #1221794 * usb: serial: cp210x: Add USB ID for Netgear Switches embedded serial adapter - LP: #1221794 * USB: storage: Add MicroVault Flash Drive to unusual_devs - LP: #1221794 * USB: misc: Add Manhattan Hi-Speed USB DVI Converter to sisusbvga - LP: #1221794 * USB: option: append Petatel NP10T device to GSM modems list - LP: #1221794 * usb: cp210x support SEL C662 Vendor/Device - LP: #1221794 * USB: cp210x: add MMB and PI ZigBee USB Device Support - LP: #1221794 * USB: EHCI: Fix resume signalling on remote wakeup - LP: #1221794 * drm/radeon: fix endian issues with DP handling (v3) - LP: #1221794 * drm/radeon: Another card with wrong primary dac adj - LP: #1221794 * drm/radeon: improve dac adjust heuristics for legacy pdac - LP: #1221794 * drm/radeon: fix combios tables on older cards - LP: #1221794 * ARM: footbridge: fix overlapping PCI mappings - LP: #1221794 * [SCSI] isci: Fix a race condition in the SSP task management path - LP: #1221794 * [SCSI] qla2xxx: Properly set the tagging for commands. - LP: #1221794 * [SCSI] sd: fix crash when UA received on DIF enabled device - LP: #1221794 * nfsd: nfsd_open: when dentry_open returns an error do not propagate as struct file - LP: #1221794 * USB: option: add D-Link DWM-152/C1 and DWM-156/C1 - LP: #1221794 * staging: comedi: COMEDI_CANCEL ioctl should wake up read/write - LP: #1221794 * staging: comedi: fix a race between do_cmd_ioctl() and read/write - LP: #1221794 * staging: android: logger: Correct write offset reset on error - LP: #1221794 * usb: option: add TP-LINK MA260 - LP: #1221794 * USB: ti_usb_3410_5052: fix dynamic-id matching - LP: #1221794 * usb: serial: option: Add ONYX 3G device support - LP: #1221794 * md/raid10: remove use-after-free bug. - LP: #1221794 * md/raid5: fix interaction of 'replace' and 'recovery'. - LP: #1221794 * drm/i915: initialize gt_lock early with other spin locks - LP: #1221794 * xhci: Avoid NULL pointer deref when host dies. - LP: #1221794 * usb: host: xhci: Enable XHCI_SPURIOUS_SUCCESS for all controllers with xhci 1.0 - LP: #1221794 * xhci: fix null pointer dereference on ring_doorbell_for_active_rings - LP: #1221794 * usb: serial: option: blacklist ONDA MT689DC QMI interface - LP: #1221794 * usb: serial: option: add Olivetti Olicard 200 - LP: #1221794 * usb: serial: option.c: remove ONDA MT825UP product ID fromdriver - LP: #1221794 * USB: mos7840: fix memory leak in open - LP: #1221794 * usb: Clear both buffers when clearing a control transfer TT buffer. - LP: #1221794 * nl80211: fix mgmt tx status and testmode reporting for netns - LP: #1221794 * mac80211: fix ethtool stats for non-station interfaces - LP: #1221794 * mac80211: fix duplicate retransmission detection - LP: #1221794 * iwlwifi: mvm: refuse connection to APs with BI 16 - LP: #1221794 * iwlwifi: add
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Tags removed: verification-quantal-done ** Tags added: verification-done-quantal -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: In Progress Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These upstream patches are necessary to fix the issue: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc f45a5c267da35174e22cec955093a7513dc1623d Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Changed in: linux (Ubuntu Quantal) Status: In Progress = Fix Committed ** Changed in: linux (Ubuntu Raring) Status: In Progress = Fix Committed -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: Fix Committed Status in “linux” source package in Raring: Fix Committed Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These upstream patches are necessary to fix the issue: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc f45a5c267da35174e22cec955093a7513dc1623d Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
The 3.5 version of this fix has been verified for throughput and stability. Unfortunately the 3.8 verification could not be done at this time. ** Tags removed: verification-needed-quantal ** Tags added: verification-done-quantal ** Tags removed: verification-done-quantal ** Tags added: verification-quantal-done -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: In Progress Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These upstream patches are necessary to fix the issue: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc f45a5c267da35174e22cec955093a7513dc1623d Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
This bug is awaiting verification that the kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-raring- needed' to 'verification-raring-done'. If verification is not done by one week from today, this fix will be dropped from the source code, and this bug will be closed. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you! -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: In Progress Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These upstream patches are necessary to fix the issue: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc f45a5c267da35174e22cec955093a7513dc1623d Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
This bug is awaiting verification that the kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-quantal- needed' to 'verification-quantal-done'. If verification is not done by one week from today, this fix will be dropped from the source code, and this bug will be closed. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you! ** Tags added: verification-needed-quantal ** Tags added: verification-needed-raring -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: In Progress Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These upstream patches are necessary to fix the issue: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc f45a5c267da35174e22cec955093a7513dc1623d Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Tags removed: verification-failed-quantal verification-failed-raring ** Description changed: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: - These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: + These upstream patches are necessary to fix the issue: + 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 + 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc + f45a5c267da35174e22cec955093a7513dc1623d Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. - The following explains an example vlan+namespace configuration: - Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 - netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] + The following explains an example vlan+namespace configuration: + Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 + netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: In Progress Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These upstream patches are necessary to fix the issue: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc f45a5c267da35174e22cec955093a7513dc1623d Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Tags removed: kernel-key -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: Fix Released Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
This bug was fixed in the package linux - 3.5.0-39.60 --- linux (3.5.0-39.60) quantal; urgency=low [Brad Figg] * Release Tracking Bug - LP: #1211872 [ Upstream Kernel Changes ] * Revert veth: avoid a NULL deref in veth_stats_one * Revert veth: extend device features * Revert veth: reduce stat overhead linux (3.5.0-38.59) quantal; urgency=low [Brad Figg] * Release Tracking Bug - LP: #1205440 [ Upstream Kernel Changes ] * hp-wmi: add more definitions for new event_id's - LP: #1152458 * MFD: rtsx_pcr: Fix probe fail path - LP: #1201321 * mfd: rtsx: Add support for RTL8411B - LP: #1201321 * veth: reduce stat overhead - LP: #1201869 * veth: extend device features - LP: #1201869 * veth: avoid a NULL deref in veth_stats_one - LP: #1201869 * Input: elantech - fix for newer hardware versions (v7) - LP: #1166442 * zram: avoid invalid memory access in zram_exit() - LP: #1204600 * zram: use zram-lock to protect zram_free_page() in swap free notify path - LP: #1204600 * zram: destroy all devices on error recovery path in zram_init() - LP: #1204600 * zram: avoid double free in function zram_bvec_write() - LP: #1204600 * zram: avoid access beyond the zram device - LP: #1204600 * zram: protect sysfs handler from invalid memory access - LP: #1204600 * pcmcia: at91_cf: fix gpio_get_value in at91_cf_get_status - LP: #1204600 * usb: gadget: f_mass_storage: add missing memory barrier for thread_wakeup_needed - LP: #1204600 * cgroup: fix umount vs cgroup_event_remove() race - LP: #1204600 * xhci: check for failed dma pool allocation - LP: #1204600 * usb: host: xhci-plat: release mem region while removing module - LP: #1204600 * USB: option,qcserial: move Novatel Gobi1K IDs to qcserial - LP: #1204600 * x86: Fix /proc/mtrr with base/size more than 44bits - LP: #1204600 * genirq: Fix can_request_irq() for IRQs without an action - LP: #1204600 * jbd2: move superblock checksum calculation to jbd2_write_superblock() - LP: #1204600 * ext3,ext4: don't mess with dir_file-f_pos in htree_dirblock_to_tree() - LP: #1204600 * jbd2: fix theoretical race in jbd2__journal_restart - LP: #1204600 * cgroup: fix umount vs cgroup_cfts_commit() race - LP: #1204600 * tracing: Use current_uid() for critical time tracing - LP: #1204600 * ahci: Add AMD CZ SATA device ID - LP: #1204600 * i2c-piix4: Add AMD CZ SMBus device ID - LP: #1204600 * ahci: remove pmp link online check in FBS EH - LP: #1204600 * libata: skip SRST for all SIMG [34]7x port-multipliers - LP: #1204600 * ASoC: wm8962: Remove remaining direct register cache accesses - LP: #1204600 * ACPICA: Do not use extended sleep registers unless HW-reduced bit is set - LP: #1204600 * ALSA: hda - Cache the MUX selection for generic HDMI - LP: #1204600 * ata_piix: IDE-mode SATA patch for Intel Coleto Creek DeviceIDs - LP: #1204600 * ahci: AHCI-mode SATA patch for Intel Coleto Creek DeviceIDs - LP: #1204600 * ARM: 7765/1: perf: Record the user-mode PC in the call chain. - LP: #1204600 * Handle big endianness in NTLM (ntlmv2) authentication - LP: #1204600 * xen/time: remove blocked time accounting from xen clockchip - LP: #1204600 * drivers/dma/pl330.c: fix locking in pl330_free_chan_resources() - LP: #1204600 * ocfs2: xattr: fix inlined xattr reflink - LP: #1204600 * block: do not pass disk names as format strings - LP: #1204600 - CVE-2013-2851 * crypto: sanitize argument for format string - LP: #1204600 * mm/memory-hotplug: fix lowmem count overflow when offline pages - LP: #1204600 * drivers/rtc/rtc-rv3029c2.c: fix disabling AIE irq - LP: #1204600 * nbd: correct disconnect behavior - LP: #1204600 * netfilter: xt_TCPOPTSTRIP: fix possible mangling beyond packet boundary - LP: #1204600 * netfilter: ipt_ULOG: fix non-null terminated string in the nf_log path - LP: #1204600 * netfilter: add nf_ipv6_ops hook to fix xt_addrtype with IPv6 - LP: #1204600 * netfilter: xt_LOG: fix mark logging for IPv6 packets - LP: #1204600 * ipvs: info leak in __ip_vs_get_dest_entries() - LP: #1204600 * netfilter: nfnetlink_cttimeout: fix incomplete dumping of objects - LP: #1204600 * netfilter: nfnetlink_acct: fix incomplete dumping of objects - LP: #1204600 * netfilter: xt_TCPMSS: Fix violation of RFC879 in absence of MSS option - LP: #1204600 * netfilter: xt_TCPOPTSTRIP: don't use tcp_hdr() - LP: #1204600 * netfilter: xt_TCPMSS: Fix missing fragmentation handling - LP: #1204600 * netfilter: xt_TCPMSS: Fix IPv6 default MSS too - LP: #1204600 * ipvs: SCTP ports should be writable in ICMP packets - LP: #1204600 * zfcp: fix adapter (re)open recovery while link to SAN is down - LP: #1204600 * zfcp: block queue limits
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Changed in: linux (Ubuntu Quantal) Status: Fix Released = In Progress -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: Fix Committed Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Tags added: kernel-key -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: Fix Committed Status in “linux” source package in Raring: Fix Committed Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
Looks like with these patch we get the following crash: PID: 6752 TASK: 8817d41e1700 CPU: 13 COMMAND: ip #0 [8817d2be33f0] machine_kexec at 8103bbda #1 [8817d2be3460] crash_kexec at 810bc028 #2 [8817d2be3530] oops_end at 8169f098 #3 [8817d2be3560] no_context at 8168436f #4 [8817d2be35b0] __bad_area_nosemaphore at 81684551 #5 [8817d2be3610] bad_area at 816845ca #6 [8817d2be3640] do_page_fault at 816a1f54 #7 [8817d2be3750] page_fault at 8169e4e5 [exception RIP: netif_carrier_off+9] RIP: 8159f379 RSP: 8817d2be3808 RFLAGS: 00010292 RAX: 0001 RBX: 8817f3e36000 RCX: 81ca9948 RDX: RSI: 0282 RDI: RBP: 8817d2be3808 R8: 880c0fcee400 R9: 8817d2be38e8 R10: 1a60 R11: 0246 R12: 8817f3e36000 R13: 8817d2be3858 R14: 8817fd32be00 R15: ORIG_RAX: CS: 0010 SS: 0018 #8 [8817d2be3810] veth_close at a02221a2 [veth] #9 [8817d2be3830] __dev_close_many at 8157e3be #10 [8817d2be3850] dev_close_many at 8157e500 #11 [8817d2be3890] rollback_registered_many at 8157e658 #12 [8817d2be38c0] unregister_netdevice_many at 8157e7eb #13 [8817d2be38e0] rtnl_dellink at 8159148b #14 [8817d2be3a30] rtnetlink_rcv_msg at 81591be4 #15 [8817d2be3ab0] netlink_rcv_skb at 815aab39 #16 [8817d2be3ae0] rtnetlink_rcv at 8158fa35 #17 [8817d2be3b00] netlink_unicast at 815aa4ad #18 [8817d2be3b50] netlink_sendmsg at 815aa810 #19 [8817d2be3be0] sock_sendmsg at 81568e37 #20 [8817d2be3d60] ___sys_sendmsg at 8156adcc #21 [8817d2be3f00] __sys_sendmsg at 8156c7b9 #22 [8817d2be3f70] sys_sendmsg at 8156c819 #23 [8817d2be3f80] system_call_fastpath at 816a65e9 RIP: 7f7dd23fae30 RSP: 7fff98b25f58 RFLAGS: 00010293 RAX: 002e RBX: 816a65e9 RCX: 0011 RDX: RSI: 7fff98b21ef0 RDI: 0005 RBP: R8: R9: R10: 7fff98b21c60 R11: 0246 R12: 8156c819 R13: 8817d2be3f78 R14: R15: 7fff98b25fe8 ORIG_RAX: 002e CS: 0033 SS: 002b A patch already exists that addresses this crash: Commit 2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc Author: Eric Dumazet eduma...@google.com Date: Thu Jan 10 08:32:45 2013 + veth: fix a NULL deref in netif_carrier_off In commit d0e2c55e7c94 (veth: avoid a NULL deref in veth_stats_one) we now clear the peer pointers in veth_dellink() veth_close() must therefore make sure the peer pointer is set. ** Tags removed: verification-needed-quantal verification-needed-raring ** Tags added: verification-failed-quantal verification-failed-raring -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: Fix Committed Status in “linux” source package in Raring: Fix Committed Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to:
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Tags added: quantal raring ** Changed in: linux (Ubuntu) Importance: Undecided = Medium -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: Fix Committed Status in “linux” source package in Raring: Fix Committed Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
Another related fix for this is here: commit f45a5c267da35174e22cec955093a7513dc1623d Author: Eric Dumazet eduma...@google.com Date: Fri Feb 8 20:10:49 2013 + veth: fix NULL dereference in veth_dellink() commit d0e2c55e7c940 (veth: avoid a NULL deref in veth_stats_one) added another NULL deref in veth_dellink(). # ip link add name veth1 type veth peer name veth0 # rmmod veth We crash because veth_dellink() is called twice, so we must take care of NULL peer. Signed-off-by: Eric Dumazet eduma...@google.com Signed-off-by: David S. Miller da...@davemloft.net -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: Fix Committed Status in “linux” source package in Raring: Fix Committed Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Branch linked: lp:ubuntu/quantal-proposed/linux-ti-omap4 -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: In Progress Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
This bug is awaiting verification that the kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed' to 'verification-done'. If verification is not done by one week from today, this fix will be dropped from the source code, and this bug will be closed. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you! ** Tags added: verification-needed-raring -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: In Progress Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
A bisection reveals that the following patch solves the issue: 8093315a91340bca52549044975d8c7f673b28a1 veth: extend device features However it relies on: 2681128f0ced8aa4e66f221197e183cc16d244fe veth: reduce stat overhead For a clean cherry-pick. And after building and testing the following bug needed to be fixed with: d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 veth: avoid a NULL deref in veth_stats_one A test build is available here that solves the issue backported to 3.5/3.8: http://people.canonical.com/~arges/lp1201869/ -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: In Progress Bug description: OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (3.9) throughput is much higher. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Description changed: + SRU Justification: + + Impact: + Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. + + Fix: + These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: + 2681128f0ced8aa4e66f221197e183cc16d244fe + 8093315a91340bca52549044975d8c7f673b28a1 + d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 + + Testcase: + Setup OpenStack Neutron. Test throughput between internal and external nodes. + + -- + OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of - Linux (3.9) throughput is much higher. + Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much + higher by almost 3 times. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: In Progress Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1201869] Re: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels
** Description changed: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. + The following explains an example vlan+namespace configuration: + Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 + netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] + Where: + 1) tap123+qr-123 and tap234+qg-234 are veth pairs + 2) qr-123 and qg-234 reside inside the qrouter-123 namespace + -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much - higher by almost 3 times. + higher by almost 3 times. For example with some testing throughput is + 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches + applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1201869 Title: poor networking throughput across an OpenStack Neutron router on 3.5/3.8 kernels Status in “linux” package in Ubuntu: Fix Released Status in “linux” source package in Quantal: In Progress Status in “linux” source package in Raring: In Progress Bug description: SRU Justification: Impact: Users of the 3.5/3.8 kernel will have poor network throughput when using OpenStack Neutron depending on their setup. Fix: These patches are all in Ubuntu-3.9.0-7.15 / v3.9-rc1: 2681128f0ced8aa4e66f221197e183cc16d244fe 8093315a91340bca52549044975d8c7f673b28a1 d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 Testcase: Setup OpenStack Neutron. Test throughput between internal and external nodes. The following explains an example vlan+namespace configuration: Internal Node: [10.x.x.2]-eth2.123-br123-tap123-qr-123[10.x.x.1] --- netns: qrouter-123 netns: qrouter-123 --- qg-234[10.x.y.1]-tap234-br234-eth2.234-External Node[10.x.y.2] Where: 1) tap123+qr-123 and tap234+qg-234 are veth pairs 2) qr-123 and qg-234 reside inside the qrouter-123 namespace -- OpenStack Neutron does IP forwarding through a network namespace. A veth pair is used to connect into the namespace. The veth pair appears to be the bottleneck, independent of network namespace. In newer versions of Linux (Ubuntu-3.9.0-7.15 / v3.9-rc1 and greater) throughput is much higher by almost 3 times. For example with some testing throughput is 3.5 Gbps in pre 3.9-rc1 versions and 9.1 Gbps with these patches applied. This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and Raring lts backports) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp