Today, I've been having a heck of a time keeping a Debian VM running on a Debian host. Anytime I use virtual SCSI drives, or do lots of disk I/O with virtual SCSI or IDE drives, or do a lot network I/O using virtual e1000 NICs, my guest VM hangs, and the kvm process in the host starts spitting out stack dumps to the console and /var/log/messages. Not sure if it's related to the 2.6.25 kernel, the kvm modules from said kernel, kvm-69, or the Debian setup.
Host Hardware: Tyan h2000M motherboard 2x dual-core Opteron 2220 CPUs at 2.8 GHz 8 GB DDR2-667 ECC SDRAM (2x 2 GB per CPU socket) 3Ware 9650SE-16ML RAID Controller 12x 500 GB SATA-II HDs in 1 large RAID6 array Host Software: Debian Lenny (64-bit) Linux kernel 2.6.25 from Debian Sid kvm-69 from repos Guest Software: Debian Lenny (64-bit) Linux kernel 2.6.25 from Debian Sid just the minimal install from CD Guest commandline: /usr/bin/kvm -name webmail -daemonize -localtime -usb -usbdevice tablet -smp 2 -m 3072 -vnc :05 -pidfile /var/run/kvm/webmail.pid -net nic,macaddr=00:16:3e:00:00:05,model=e1000 -net tap,ifname=tap05 -boot c -drive index=0,media=disk,if=ide,file=/dev/mapper/vol0-webmail--boot -drive index=1,media=disk,if=ide,file=/dev/mapper/vol0-webmail--storage If I issue a "shutdown -r now" in the guest, I can almost always trigger the stack dump in the host. If I do a lot of disk I/O (restore dar archives, for instance) in the guest, I can always trigger the stack dump (although the time it takes to cause it varies). If I do a lot of network I/O in the guest (transferring 10-20 GB of dar archives, for instance), I can sometimes trigger the stack dump in the host. If I use virtual scsi drives, it happens everytime I install a guest using a virtual CD-ROM. Stack dump output in the host looks like: Jun 6 13:29:38 cube kernel: [20272.148988] ------------[ cut here ]------------ Jun 6 13:29:38 cube kernel: [20272.148990] WARNING: at arch/x86/kvm/x86.c:161 kvm_queue_exception_e+0x26/0x47 [kvm]() Jun 6 13:29:38 cube kernel: [20272.148992] Modules linked in: tun nfsd auth_rpcgss exportfs sbs sbshc ac container battery video output wmi nfs lockd nfs_acl sunrpc bridge bonding ext3 jbd mbcache kvm_amd kvm loop psmouse serio_raw k8temp pcspkr i2c_piix4 i2c_core button shpchp pci_hotplug evdev xfs dm_mirror dm_snapshot dm_mod ide_cd_mod cdrom ata_generic libata dock sd_mod serverworks floppy e1000e 3w_9xxx scsi_mod tg3 ide_pci_generic e100 mii ehci_hcd ohci_hcd ide_core thermal processor fan Jun 6 13:29:38 cube kernel: [20272.149012] Pid: 4710, comm: kvm Tainted: G M 2.6.25-2-amd64 #1 Jun 6 13:29:38 cube kernel: [20272.149014] Jun 6 13:29:38 cube kernel: [20272.149015] Call Trace: Jun 6 13:29:38 cube kernel: [20272.149018] [<ffffffff80234cc5>] warn_on_slowpath+0x51/0x63 Jun 6 13:29:38 cube kernel: [20272.149028] [<ffffffff88231987>] :kvm:gfn_to_hva+0x1c/0x41 Jun 6 13:29:38 cube kernel: [20272.149038] [<ffffffff88231b14>] :kvm:kvm_read_guest_page+0x34/0x46 Jun 6 13:29:38 cube kernel: [20272.149047] [<ffffffff88231b7d>] :kvm:kvm_read_guest+0x57/0x6c Jun 6 13:29:38 cube kernel: [20272.149058] [<ffffffff88238d02>] :kvm:paging64_walk_addr+0xd9/0x2b3 Jun 6 13:29:38 cube kernel: [20272.149062] [<ffffffff8041ee85>] __down_read+0x12/0xa1 Jun 6 13:29:38 cube kernel: [20272.149071] [<ffffffff882347f5>] :kvm:kvm_queue_exception_e+0x26/0x47 Jun 6 13:29:38 cube kernel: [20272.149080] [<ffffffff88239281>] :kvm:paging64_page_fault+0x90/0x37b Jun 6 13:29:38 cube kernel: [20272.149085] [<ffffffff8824f085>] :kvm_amd:svm_vcpu_load+0x4f/0x76 Jun 6 13:29:38 cube kernel: [20272.149088] [<ffffffff8022a54c>] hrtick_start_fair+0xfb/0x143 Jun 6 13:29:38 cube kernel: [20272.149095] [<ffffffff8041de83>] thread_return+0x6b/0xac Jun 6 13:29:38 cube kernel: [20272.149107] [<ffffffff882383a7>] :kvm:kvm_mmu_page_fault+0x19/0x81 Jun 6 13:29:38 cube kernel: [20272.149118] [<ffffffff88236010>] :kvm:kvm_arch_vcpu_ioctl_run+0x3d4/0x529 Jun 6 13:29:38 cube kernel: [20272.149128] [<ffffffff882323fa>] :kvm:kvm_vcpu_ioctl+0xda/0x2dd Jun 6 13:29:38 cube kernel: [20272.149132] [<ffffffff80249395>] hrtimer_start+0x112/0x134 Jun 6 13:29:38 cube kernel: [20272.149137] [<ffffffff8020be39>] sysret_signal+0x2b/0x45 Jun 6 13:29:38 cube kernel: [20272.149139] [<ffffffff8023020d>] hrtick_set+0x9e/0xf7 Jun 6 13:29:38 cube kernel: [20272.149143] [<ffffffff8024fc76>] futex_wake+0x78/0x101 Jun 6 13:29:38 cube kernel: [20272.149149] [<ffffffff8023e814>] __dequeue_signal+0x19/0x15a Jun 6 13:29:38 cube kernel: [20272.149154] [<ffffffff8023e574>] recalc_sigpending+0xe/0x38 Jun 6 13:29:38 cube kernel: [20272.149156] [<ffffffff8023f7bd>] dequeue_signal+0x8d/0x113 Jun 6 13:29:38 cube kernel: [20272.149161] [<ffffffff8024054e>] sys_rt_sigtimedwait+0xf1/0x25f Jun 6 13:29:38 cube kernel: [20272.149165] [<ffffffff802a5a61>] vfs_ioctl+0x21/0x6b Jun 6 13:29:38 cube kernel: [20272.149168] [<ffffffff802a5cf3>] do_vfs_ioctl+0x248/0x261 Jun 6 13:29:38 cube kernel: [20272.149172] [<ffffffff802a5d5d>] sys_ioctl+0x51/0x70 Jun 6 13:29:38 cube kernel: [20272.149176] [<ffffffff8020bd8a>] system_call_after_swapgs+0x8a/0x8f Jun 6 13:29:38 cube kernel: [20272.149181] Jun 6 13:29:38 cube kernel: [20272.149182] ---[ end trace a091c5726725b2f1 ]--- And it just repeats continuously until I forcibly kill the kvm process, and even then it can continue outputting to the logs for up to 5 minutes. Not sure where to even start looking on this one. Open to any and all suggestions. -- Freddie Cash [EMAIL PROTECTED] -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
