Tracy R Reed
2006-Aug-17 22:39 UTC
[Xen-users] Error: Device 0 (vif) could not be connected. Backend device not found.
There is probably a simple solution to this but I can''t seem to find it despite googling and asking around and playing with things for a couple of days. I am running xen-3.0-testing.hg which I just pulled a new copy of last night. uname -a produces: Linux cpu01 2.6.16.13-xen #1 SMP Thu Aug 17 12:51:50 PDT 2006 x86_64 x86_64 x86_64 GNU/Linux The OS is CentOS 4.3. I am running it on an AMD64 motherboard with disk provided via AoE. I have already set up the initrd and everything and the disk works perfectly. The system comes up and xend starts. But when I try to create a new domU: [root@cpu01 ~]# /usr/sbin/xm create /etc/xen/auto/temp1 Using config file "/etc/xen/auto/temp1". Error: Device 0 (vif) could not be connected. Backend device not found. Someone on IRC suggested that the netloop and netbk modules were not installed. But it looks like they are compiled into the kernel that comes with xen-3.0-testing.hg, the XEN section of the kernel''s .config: CONFIG_XEN_PRIVILEGED_GUEST=y # CONFIG_XEN_UNPRIVILEGED_GUEST is not set CONFIG_XEN_BACKEND=y CONFIG_XEN_PCIDEV_BACKEND=y # CONFIG_XEN_PCIDEV_BACKEND_VPCI is not set CONFIG_XEN_PCIDEV_BACKEND_PASS=y # CONFIG_XEN_PCIDEV_BE_DEBUG is not set CONFIG_XEN_BLKDEV_BACKEND=y # CONFIG_XEN_BLKDEV_TAP_BE is not set CONFIG_XEN_NETDEV_BACKEND=y # CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER is not set CONFIG_XEN_NETDEV_LOOPBACK=y # CONFIG_XEN_TPMDEV_BACKEND is not set CONFIG_XEN_BLKDEV_FRONTEND=y CONFIG_XEN_NETDEV_FRONTEND=y # CONFIG_XEN_BLKDEV_TAP is not set # CONFIG_XEN_TPMDEV_FRONTEND is not set CONFIG_XEN_SCRUB_PAGES=y CONFIG_XEN_DISABLE_SERIAL=y CONFIG_XEN_SYSFS=m CONFIG_HAVE_ARCH_ALLOC_SKB=y CONFIG_HAVE_ARCH_DEV_ALLOC_SKB=y CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y CONFIG_NO_IDLE_HZ=y So it can''t be that the module isn''t installed right? And this is definitely the kernel I am running. Googling has also suggested that the bridge module could be part of the problem but that module is loaded also. lsmod shows it as the first module listed: Module Size Used by bridge 52256 0 When I issue the xm create command this appears in the messages file: Aug 17 15:35:36 vmware logger: /etc/xen/scripts/vif-bridge: Could not find bridge, and none was specified Aug 17 15:35:37 vmware net.agent[4344]: remove event not handled Aug 17 15:35:39 vmware logger: /etc/xen/scripts/vif-bridge: Could not find bridge, and none was specified The /var/log/xend.log of an attempt to create a new domain can be found here: http://pastebin.ca/136812 Anyone know what the problem could be? I have been working on this project for a couple of months and learned all about initrd''s, AoE, and lots of other stuff and I am *so* close to getting this whole cluster working! Really looking forward to putting this into production. Thanks for any help you can provide. -- Tracy R Reed http://ultraviolet.org _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Tracy R Reed
2006-Aug-17 22:53 UTC
Re: [Xen-users] Error: Device 0 (vif) could not be connected. Backend device not found.
Tracy R Reed wrote: > [root@cpu01 ~]# /usr/sbin/xm create /etc/xen/auto/temp1 > Using config file "/etc/xen/auto/temp1". > Error: Device 0 (vif) could not be connected. Backend device not > found. So I had the bright idea of deciding I would see what happens if I just comment out the vif = [ '''' ] line in the xen config file and create the domU. [root@cpu01 linux-2.6.16.13-xen]# /usr/sbin/xm create /etc/xen/auto/temp1 Using config file "/etc/xen/auto/temp1". Started domain temp1 [root@cpu01 linux-2.6.16.13-xen]# And then the box hung solid. No network, no console. The console is blanked and it will not unblank. Even hitting scroll lock on the keyboard does nothing. It has done this a couple of times before when doing various different things like when I restart xend. I am quite puzzled as to why it was so relatively easy to set up my home machine with xen and it has been stable for a year but I can''t get this cluster to work. -- Tracy R Reed http://ultraviolet.org _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Tracy R Reed
2006-Aug-17 23:06 UTC
Re: [Xen-users] Error: Device 0 (vif) could not be connected. Backend device not found.
Tracy R Reed wrote:> And then the box hung solid. No network, no console. The console is > blanked and it will not unblank. Even hitting scroll lock on theAnd finally, one more datapoint in case it matters: /proc/cpuinfo shows us the following: processor : 0 vendor_id : AuthenticAMD cpu family : 15 model : 33 model name : Dual Core AMD Opteron(tm) Processor 270 stepping : 2 cpu MHz : 1992.062 cache size : 1024 KB physical id : 0 siblings : 1 core id : 0 cpu cores : 1 fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu tsc msr pae mce cx8 apic mtrr mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow pni lahf_lm cmp_legacy bogomips : 3985.47 TLB size : 1024 4K pages clflush size : 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management: ts fid vid ttp And that is all. Just the one cpu. Even though it is a dual core system. The kernel''s .config contains: CONFIG_SMP=y so I would think it should detect two cpu''s. The motherboard is a Tyan S2882-D Thunder K8SD Pro. It has an Intel eepro100 ethernet adaptor and two Broadcom Tigon3 gig-e ethernet adaptors. I currently only have an ethernet cable plugged into the eepro100 (uses the e100 driver) and it is eth0. Just in case that has some bearing on the bridge/network situation. -- Tracy R Reed http://ultraviolet.org A: Because we read from top to bottom, left to right Q: Why should I start my reply below the quoted text _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Tracy R Reed
2006-Aug-18 00:16 UTC
Re: [Xen-users] Error: Device 0 (vif) could not be connected. Backend device not found.
Tracy R Reed wrote:> And that is all. Just the one cpu. Even though it is a dual core system. > The kernel''s .config contains:Ok, this issue is resolved which brings us to a new issue. Turns out my pxeboot config file (named by the mac addy of the machine) contained: DEFAULT mboot.c32 xen.gz dom0_mem=258048 nosmp noacpi --- vmlinuz-2.6.16.13-xen console=tty0 noreboot root=/de\v/aoe/e0.0 --- initrd-2.6.16.13-xen So at some point I put (or more likely copied an example with it already there) the nosmp on there. So now I have two cpu''s. And I successfully created a domU and it started executing the CentOS 4.3 initrd for my kickstart install which is great! Might there be some sort of bug involving xen running an smp cpu configured with nosmp? It seemed to hang pretty reliably. I don''t have time to explore this further but perhaps I will come back to this later on. So I did an xm kill temp1 on the domU and got a kernel oops. In the process of attempting to download ksymoops I noticed that the gateway was not set so set the gateway and rebooted the box. Normally I don''t reboot just to set the gateway but I had some problems due to local netmask weirdness and just wanted to start all over. When the machine was coming back up it got as far as: Starting xend: Bridge firewalling registered And then another solid hang. It now hangs every time the machine boots and tries to start xend. I will try to figure out how to get some more debugging output from it. -- Tracy R Reed http://ultraviolet.org A: Because we read from top to bottom, left to right Q: Why should I start my reply below the quoted text _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users