Re: [ovirt-users] Windows Networking Issues -
I wanted to follow up on this after I found my resolution. I started to see kernel errors when I migrated all but my windows host off a hypervisor and generated traffic. I then took those errors and started looking back at all of the hypervisors only to find this error was on each of them; also it actively reporting on systems with Windows VMs. Tracing back the logs lead me to the bug reports below where I learned that this issue had been re-introduced to the 2.6.32-X kernel. ABRT info - Nov 30 19:28:42 server2.example.com kernel: WARNING: at net/core/dev.c:1915 skb_warn_bad_offload+0x99/0xb0() (Tainted: GW -- ) Nov 30 19:28:42 server2.example.com kernel: Hardware name: PowerEdge M620 Nov 30 19:28:42 server2.example.com kernel: : caps=(0x40c9, 0x0) len=1514 data_len=1460 ip_summed=1 Nov 30 19:28:42 server2.example.com kernel: Modules linked in: sch_prio act_mirred cls_u32 sch_ingress ebt_arp xt_physdev ipt_REJECT nf_conntrack_ipv4 nf_defrag_ipv4 xt_multiport iptable_filter ip_tables fuse nfs lockd fscache auth_rpcgss nfs_acl sunrpc vfat fat bonding ebtable_nat ebtables be2iscsi iscsi_boot_sysfs bnx2i cnic uio cxgb4i iw_cxgb4 cxgb4 cxgb3i libcxgbi iw_cxgb3 cxgb3 ib_iser rdma_cm ib_cm iw_cm ib_sa ib_mad ib_core ib_addr iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi mpt3sas mpt2sas scsi_transport_sas raid_class mptctl mptbase dell_rbu autofs4 bridge 8021q garp stp llc ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 xt_state nf_conntrack ip6table_filter ip6_tables ipv6 dm_round_robin dm_multipath vhost_net macvtap macvlan tun kvm_intel kvm sg ipmi_devintf sr_mod cdrom joydev power_meter acpi_ipmi ipmi_si ipmi_msghandler iTCO_wdt iTCO_vendor_support ixgbe dca ptp pps_core mdio dcdbas sb_edac edac_core lpc_ich mfd_core shpchp usb_storage ext4 jbd2 mbcache sd_mod crc_t10dif megaraid_sas wmi ahci dm_mirror dm_region_hash dm_log dm_mod [last unloaded: ip_tables] Nov 30 19:28:42 server2.example.com kernel: Pid: 0, comm: swapper Tainted: GW -- 2.6.32-573.7.1.el6.x86_64 #1 Nov 30 19:28:42 server2.example.com kernel: Call Trace: Nov 30 19:28:42 server2.example.com kernel: [] ? warn_slowpath_common+0x91/0xe0 Nov 30 19:28:42 server2.example.com kernel: [] ? warn_slowpath_fmt+0x46/0x60 Nov 30 19:28:42 server2.example.com kernel: [] ? __ratelimit+0xd5/0x120 Nov 30 19:28:42 server2.example.com kernel: [] ? skb_warn_bad_offload+0x99/0xb0 Nov 30 19:28:42 server2.example.com kernel: [] ? __skb_gso_segment+0x71/0xc0 Nov 30 19:28:42 server2.example.com kernel: [] ? skb_gso_segment+0x13/0x20 Nov 30 19:28:42 server2.example.com kernel: [] ? dev_hard_start_xmit+0x9b/0x490 Nov 30 19:28:42 server2.example.com kernel: [] ? sch_direct_xmit+0x15a/0x1c0 Nov 30 19:28:42 server2.example.com kernel: [] ? dev_queue_xmit+0x228/0x320 Nov 30 19:28:42 server2.example.com kernel: [] ? __br_forward+0x0/0xd0 [bridge] Nov 30 19:28:42 server2.example.com kernel: [] ? br_dev_queue_push_xmit+0x88/0xc0 [bridge] Nov 30 19:28:42 server2.example.com kernel: [] ? br_forward_finish+0x58/0x60 [bridge] Nov 30 19:28:42 server2.example.com kernel: [] ? __br_forward+0xaa/0xd0 [bridge] Nov 30 19:28:42 server2.example.com kernel: [] ? skb_clone+0x58/0xb0 Nov 30 19:28:42 server2.example.com kernel: [] ? deliver_clone+0x3e/0x60 [bridge] Nov 30 19:28:42 server2.example.com kernel: [] ? br_forward+0x41/0x70 [bridge] Nov 30 19:28:42 server2.example.com kernel: [] ? br_handle_frame_finish+0x17e/0x330 [bridge] Nov 30 19:28:42 server2.example.com kernel: [] ? br_handle_frame+0x1c0/0x270 [bridge] Nov 30 19:28:42 server2.example.com kernel: [] ? br_handle_frame+0x0/0x270 [bridge] Nov 30 19:28:42 server2.example.com kernel: [] ? __netif_receive_skb+0x1c7/0x570 Nov 30 19:28:42 server2.example.com kernel: [] ? netif_receive_skb+0x58/0x60 Nov 30 19:28:42 server2.example.com kernel: [] ? napi_skb_finish+0x50/0x70 Nov 30 19:28:42 server2.example.com kernel: [] ? napi_gro_receive_gr+0x39/0x50 Nov 30 19:28:42 server2.example.com kernel: [] ? vlan_gro_receive+0x1b/0x30 Nov 30 19:28:42 server2.example.com kernel: [] ? ixgbe_clean_rx_irq+0x995/0xc70 [ixgbe] Nov 30 19:28:42 server2.example.com kernel: [] ? ixgbe_poll+0x40a/0x760 [ixgbe] Nov 30 19:28:42 server2.example.com kernel: [] ? net_rx_action+0x103/0x2f0 Nov 30 19:28:42 server2.example.com kernel: [] ? ktime_get+0x6d/0x100 Nov 30 19:28:42 server2.example.com kernel: [] ? __do_softirq+0xc1/0x1e0 Nov 30 19:28:42 server2.example.com kernel: [] ? handle_IRQ_event+0x60/0x170 Nov 30 19:28:42 server2.example.com kernel: [] ? call_softirq+0x1c/0x30 Nov 30 19:28:42 server2.example.com kernel: [] ? do_softirq+0x65/0xa0 Nov 30 19:28:42 server2.example.com kernel: [] ? irq_exit+0x85/0x90 Nov 30 19:28:42 server2.example.com kernel: [] ? do_IRQ+0x75/0xf0 Nov 30 19:28:42 server2.example.com kernel: [] ? ret_from_intr+0x0/0x11 Nov 30 19:28:42 server2.example.com kernel: [] ? intel_idle+0xfe/0x1b0 Nov 30 19:28:42 server2.example.com kernel: [] ? intel_idle+0xe1/0x1b0 Nov 30 19:28:42 serve
Re: [ovirt-users] Windows Networking Issues -
All networking within oVirt seems fine. I can ping devices on other networks, I'm able to use a squid proxy on the same network with no issues. I'm able to use Terminal Services from another VM in oVirt. One oddity is when I first boot, I'm able to hit Google and search for a Dog. On my 2nd search it starts to timeout. All and all ip networking seems to be configured well. The DG lives outside of oVirt and I'm always able to ping it; even increasing the packet size. MTU size is 1500 and matches the VLAN MTU in oVirt; the bond0 is set to set to an 9000 MTU. EM1 and EM2 are in bond0 and show no framing errors. On Wed, Nov 25, 2015 at 1:30 AM, Yaniv Kaul wrote: > On Tue, Nov 24, 2015 at 11:17 PM, Matt Wells > wrote: > >> Hi all, I have a question about Windows VMs and a networking issue I'm >> having. >> >> Here's the setup - >> * oVirt - 3.5.1.1-1 >> * Hypervisors are CentOS 6.7 box with 2 NICs in bond0 'mode=4 miimon=100 >> lacp_rate=1' >> * On bond0 I have a few networks using vlan tagging. >> * Networks are 5,10,15,20 - All on an external switch >> >> Network 15 has a Windows 2012 R2 server and a CentOS 6.7 server on it. >> The rest of the networks have a few linux. >> >> Every linux box on every network is happy. However any and all Windows >> boxes I bring online are incapable of patching or hitting the web. I >> pointed the Windows box to the linux box next to it as a proxy (after >> installing squid on it) When I do that the Windows box has no issues at >> all; it's only when he's attempting to leave on his own. >> >> On my firewall I put in a 'permit any any' on the M$ box IP however all I >> see is tcp resets in PCAPs, >> > > Can you verify basic IP networking is working correctly for those VMs? > For example, we've established that they can get to the Linux VMs - how? > Are they on the same subnet? Or do they go through their default gateway? > Without knowing the IP topology, if the Linux machines were on the same > subnet as the Windows one, and the Windows machine fail to get to their > default gateway for some reason, this may perfectly explain the issue. > Y. > > >> >> I've been playing with for some time but can't seem to find the issue. >> It would be one thing if everything on the 15 was bad but the linux box on >> the network is fine. Here's the rub, I'm 99.999% sure this used to work. >> gggrrr... >> >> Any assistance anyone can offer would be amazingly appreciated. >> >> Thank you for taking the time to read this. >> >> >> ___ >> Users mailing list >> Users@ovirt.org >> http://lists.ovirt.org/mailman/listinfo/users >> >> > ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] Windows Networking Issues -
On Tue, Nov 24, 2015 at 11:17 PM, Matt Wells wrote: > Hi all, I have a question about Windows VMs and a networking issue I'm > having. > > Here's the setup - > * oVirt - 3.5.1.1-1 > * Hypervisors are CentOS 6.7 box with 2 NICs in bond0 'mode=4 miimon=100 > lacp_rate=1' > * On bond0 I have a few networks using vlan tagging. > * Networks are 5,10,15,20 - All on an external switch > > Network 15 has a Windows 2012 R2 server and a CentOS 6.7 server on it. > The rest of the networks have a few linux. > > Every linux box on every network is happy. However any and all Windows > boxes I bring online are incapable of patching or hitting the web. I > pointed the Windows box to the linux box next to it as a proxy (after > installing squid on it) When I do that the Windows box has no issues at > all; it's only when he's attempting to leave on his own. > > On my firewall I put in a 'permit any any' on the M$ box IP however all I > see is tcp resets in PCAPs, > Can you verify basic IP networking is working correctly for those VMs? For example, we've established that they can get to the Linux VMs - how? Are they on the same subnet? Or do they go through their default gateway? Without knowing the IP topology, if the Linux machines were on the same subnet as the Windows one, and the Windows machine fail to get to their default gateway for some reason, this may perfectly explain the issue. Y. > > I've been playing with for some time but can't seem to find the issue. It > would be one thing if everything on the 15 was bad but the linux box on the > network is fine. Here's the rub, I'm 99.999% sure this used to work. > gggrrr... > > Any assistance anyone can offer would be amazingly appreciated. > > Thank you for taking the time to read this. > > > ___ > Users mailing list > Users@ovirt.org > http://lists.ovirt.org/mailman/listinfo/users > > ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] Windows Networking Issues -
Hi all, I have a question about Windows VMs and a networking issue I'm having. Here's the setup - * oVirt - 3.5.1.1-1 * Hypervisors are CentOS 6.7 box with 2 NICs in bond0 'mode=4 miimon=100 lacp_rate=1' * On bond0 I have a few networks using vlan tagging. * Networks are 5,10,15,20 - All on an external switch Network 15 has a Windows 2012 R2 server and a CentOS 6.7 server on it. The rest of the networks have a few linux. Every linux box on every network is happy. However any and all Windows boxes I bring online are incapable of patching or hitting the web. I pointed the Windows box to the linux box next to it as a proxy (after installing squid on it) When I do that the Windows box has no issues at all; it's only when he's attempting to leave on his own. On my firewall I put in a 'permit any any' on the M$ box IP however all I see is tcp resets in PCAPs, I've been playing with for some time but can't seem to find the issue. It would be one thing if everything on the 15 was bad but the linux box on the network is fine. Here's the rub, I'm 99.999% sure this used to work. gggrrr... Any assistance anyone can offer would be amazingly appreciated. Thank you for taking the time to read this. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users