带有Ubuntu 14.04的DPDK OVS

Question

我们试图在Ubuntu14.04内核3.19的openstack之上安装DPDK。我们指的是执行相同操作的以下步骤。

https://software.intel.com/en-us/blogs/2015/06/09/building-vhost-user-for-ovs-today-using-dpdk-200

在执行过程中，我们遇到了一些与ovs服务有关的问题，因为它在启动时会挂起。

命令./ovs-vswitchd --dpdk -c 0x0FF8 -n 4 --socket-mem 1024,0 -- --log-file=/var/log/openvswitch/ovs-vswitchd.log --pidfile=/var/run/op penvswitch/ovs-vswitchd.pid的输出如下

AL: PCI device 0000:04:00.0 on NUMA socket 0
EAL:   probe driver: 8086:10fb rte_ixgbe_pmd
EAL:   Not managed by a supported kernel driver, skipped
EAL: PCI device 0000:04:00.1 on NUMA socket 0
EAL:   probe driver: 8086:10fb rte_ixgbe_pmd
EAL:   Not managed by a supported kernel driver, skipped
Zone 0: name:<MALLOC_S0_HEAP_0>, phys:0x30400000, len:0xb00000, virt:0x7f2d1ba00000, socket_id:0, flags:0
Zone 1: name:<RG_MP_log_history>, phys:0x35400000, len:0x2080, virt:0x7f2d1b600000, socket_id:0, flags:0
Zone 2: name:<MP_log_history>, phys:0xe3ce00000, len:0x28a0c0, virt:0x7f2d09400000, socket_id:0, flags:0
2015-11-24T13:14:12Z|00002|vlog|INFO|opened log file /var/log/openvswitch/ovs-vswitchd.log
2015-11-24T13:14:12Z|00003|ovs_numa|INFO|Discovered 24 CPU cores on NUMA node 0
2015-11-24T13:14:12Z|00004|ovs_numa|INFO|Discovered 24 CPU cores on NUMA node 1
2015-11-24T13:14:12Z|00005|ovs_numa|INFO|Discovered 2 NUMA nodes and 48 CPU cores
2015-11-24T13:14:12Z|00006|reconnect|INFO|unix:/var/run/openvswitch/db.sock: connecting...
2015-11-24T13:14:12Z|00007|reconnect|INFO|unix:/var/run/openvswitch/db.sock: connected
2015-11-24T13:14:12Z|00008|ofproto_dpif|INFO|netdev@ovs-netdev: Datapath supports recirculation
2015-11-24T13:14:12Z|00009|ofproto_dpif|INFO|netdev@ovs-netdev: MPLS label stack length probed as 3
2015-11-24T13:14:12Z|00010|ofproto_dpif|INFO|netdev@ovs-netdev: Datapath supports unique flow ids
2015-11-24T13:14:12Z|00011|dpif_netlink|ERR|Generic Netlink family 'ovs_datapath' does not exist. The Open vSwitch kernel module is probably not loaded.
2015-11-24T13:14:12Z|00012|dpif|WARN|failed to enumerate system datapaths: No such file or directory
2015-11-24T13:14:12Z|00013|dpif|WARN|failed to create datapath ovs-system: No such file or directory
2015-11-24T13:14:12Z|00014|ofproto_dpif|ERR|failed to open datapath of type system: No such file or directory
2015-11-24T13:14:12Z|00015|ofproto|ERR|failed to open datapath br-eth1: No such file or directory
2015-11-24T13:14:12Z|00016|bridge|ERR|failed to create bridge br-eth1: No such file or directory
2015-11-24T13:14:12Z|00017|bridge|INFO|bridge br-ex: added interface br-ex on port 65534
2015-11-24T13:14:12Z|00018|bridge|INFO|bridge br-int: added interface tapb6327a6f-e3 on port 1
2015-11-24T13:14:12Z|00019|bridge|INFO|bridge br-int: added interface br-int on port 65534
2015-11-24T13:14:12Z|00020|bridge|INFO|bridge br-ex: using datapath ID 0000be3137b62448
2015-11-24T13:14:12Z|00021|connmgr|INFO|br-ex: added service controller "punix:/var/run/openvswitch/br-ex.mgmt"
2015-11-24T13:14:12Z|00022|bridge|INFO|bridge br-int: using datapath ID 000016d2ab534949
2015-11-24T13:14:12Z|00023|connmgr|INFO|br-int: added service controller "punix:/var/run/openvswitch/br-int.mgmt"
2015-11-24T13:14:12Z|00024|bridge|INFO|ovs-vswitchd (Open vSwitch) 2.4.90
2015-11-24T13:14:19Z|00025|memory|INFO|25200 kB peak resident set size after 10.3 seconds
2015-11-24T13:14:19Z|00026|memory|INFO|handlers:17 ports:3 revalidators:7 rules:9

• 我们试着建立了echo 0 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages，但没有成功。

如果我们丢失了任何东西，并且在开始时导致ovs卡住，你能帮我们弄到这些东西吗？

此外，当我们使用DPDK在openstack中创建vm时，dpdkvhost-用户类型接口将自动创建。如果这个接口是用常规的br-int桥(而不是DPDK桥br0 )映射的，那么这是否意味着我们已经成功地使用netdev数据路径启用了DPDK？

Answer 1

您的问题是“不受支持的内核驱动程序管理，跳过”。

您必须将设备从内核驱动程序中取消到例如uio驱动程序，以便DPDK可以使用它们。

对于Ubuntu16.04，以下内容将帮助您：https://help.ubuntu.com/16.04/serverguide/DPDK.html#dpdk-config-dev。

这可以用dpdk_nic_bind手动完成。

从--help输出：

显示当前设备状态: dpdk_nic_bind --从当前驱动程序绑定eth1并移动到使用igb_uio dpdk_nic_bind -- bind =igb_uio eth1解除绑定0000:01:00.0从使用任何驱动程序dpdk_nic_bind -u 0000:01:00.0绑定0000:02:00.0和0000:02:00.1绑定到ixgbe内核驱动程序dpdk_nic_bind -b ixgbe 02:00.0 :00.1