October 2009 - virt - Fedora mailing-lists

libvirtd crashes host when autostarting
by Rich Mahn 06 Jan '10

06 Jan '10

Summary: My host crashes when I use the autostart function in libvirtd. This command sequence works fine: service libvirtd stop virsh autostart mydomain --disable service lib virtd start virsh start mydomain This command sequence will crash the host: service libvirtd stop virsh autostart mydomain service libvirtd start I can't find any error messages related to this. The host system freezes, the keyboard starts blinking the caps-lock and scroll-lock, and as best I can tell, the kernel has stopped running. If I starting up the host system with the VM autostarted, I sometimes get messages that look like kernel errors on the screen, but I don't find them in any log files so far. They seem to be related to networking. Here's some configuration information--I don't know really what all is needed, but this, I guess is a start: Host and VM both Fedora 11 with the lastest updates. Host machine in Intel based with the virtual support turned on. In the /var/log/libvirtd/qemu directory, the command line to start the virtual machine is different when started from autostart. The difference is in this portion: -net tap,fd=11,vlan=0 for failure (autostart) and -net tap,fd=18,vlan=0 for success I am using a bridged connection. Here's the command line that's used, split over lines for some degree of readability: LC_ALL=C PATH=/sbin:/usr/sbin:/bin:/usr/bin /usr/bin/qemu-kvm -S -M pc -m 1024 -smp 1 -name mydomain -uuid 5cde5a23-2abe-fcdd-04dc-5ae3a46ab51c -monitor pty -pidfile /var/run/libvirt/qemu//mydomain.pid -boot c -drive file=,if=ide,media=cdrom,index=2 -drive file=/dev/mapper/vg_host-lv_mydomain,if=virtio,index=0,boot=on -drive file=/dev/mapper/vg_host-lv_mydomain_data,if=virtio,index=1 -net nic,macaddr=54:52:00:67:91:e9,vlan=0,model=virtio -net tap,fd=11,vlan=0 -serial pty -parallel none -usb -usbdevice tablet -vnc 127.0.0.1:0

6 9

disabling ksm by default
by Izik Eidus 11 Nov '09

11 Nov '09

Hi, I saw that Fedora 12 will have ksm tunning script that control the ksm speed / kernel pages allocation. The only problem that I have is - as far as I remember ksm is by default enabled and only with 2000 kernel pages... What I am worried about is that users wont use the ksm tunning script and would just run ksm with this 2000 kernel pages - the result would be that ksm will probably merge just the zero pages (that could be alot of memory) and the user might not know that much more memory can be saved... Is it possible to at least make ksm disabled by default? so the users will have to run the ksm tunning script when they want to start ksm? And if we set it to disabled by default, cant we set the initialized values into more realistic value? (like 1/4 of the memory in the mainline kernels? ) Thanks.

8 19

duplicated console and details buttons
by Gianluca Cecchi 30 Oct '09

30 Oct '09

You get the "console" and "details" buttons/functions both inside the toolbar and as labels just under the toolbar. It is ok that if you uncheck "view toolbar" you still get the labels, but it seems to me sort of redundant... they are right one below the same... see attached image Just my opinion. Gianluca

2 4

Fedora virt status
by Mark McLoughlin 30 Oct '09

30 Oct '09

Fedora 12 ========= Fedora 12 Beta slipped by a week, but was released on Oct 20: https://fedoraproject.org/wiki/F12_Beta_Announcement We're now in the final countdown to F12 GA. The first release candidate will be composed next week and, all going well, we should have a release on November 17. Thankfully, the virt blocker list is now clear, but if you're looking to help with making Fedora 12 even better, there's no better place to start than the F12 target tracker bug: https://bugzilla.redhat.com/showdependencytree.cgi?id=F12VirtTarget&hide_re… There's over 100 bugs there that need your help! Witty Tagline ============= Dan Berrange has done a great job of coming up with a "witty tagline" for each of the virtualization feature sets in previous Fedora releases: https://fedoraproject.org/wiki/Virtualization/History Fedora 11: The walled garden Fedora 10: Management at a distance Fedora 9: Farewell to old friends Fedora 8: Protection from the bad guys Fedora 7: The new kid on the block Fedora Core 6: Virtualization grows up Fedora Core 5: The future is now Fedora Core 4: Glimpse of the future We need a tagline which describes our eclectic mix of features in Fedora 12. Any ideas? Fedora Weekly News ================== Dale Bewley has contributed another bunch of virtualization sections to Fedora Weekly News: https://fedoraproject.org/wiki/FWN/Issue196#Virtualization https://fedoraproject.org/wiki/FWN/Issue197#Virtualization https://fedoraproject.org/wiki/FWN/Issue199#Virtualization ABRT and qemu-kvm ================= It's quite common that we have to ask bug reports to collect a stack trace for a qemu-kvm segfault. Is ABRT the future? https://bugzilla.redhat.com/show_bug.cgi?id=528754#c1 Try this: $> yum install --enablerepo=rawhide-debuginfo qemu-debuginfo $> yum install -y abrt abrt-cli abrt-addon-ccpp abrt-plugin-logger $> service abrtd restart then when qemu-kvm crashes, you should be able to do e.g.: $> abrt-cli --get-list $> echo n | abrt-cli --report $uuid > t.log Anyone got ideas for a better way to use ABRT to catch virt related bugs? We could do with adding instructions to: http://fedoraproject.org/wiki/How_to_debug_Virtualization_problems Got Windows Guests? =================== Rich Jones has added yet another weird and fantastical feature to libguestfs ... and he needs your help to check that it works: http://www.redhat.com/archives/fedora-virt/2009-October/msg00128.html If you have any Windows guests, then you can help Fedora to support Windows guests better by spending a few minutes testing the Windows Registry feature we just added to libguestfs 1.0.75. Basically, the new virt-win-reg tool can be used to display entries from the registry in a Windows guest. How cool is that? Bugzilla ======== DOOM-O-METER: 198 bugs open four weeks ago, only 186 now! Ongoing Bugs ============ == kernel == https://bugzilla.redhat.com/515741 2.6.30 kernel stopped supporting xattrs on hugetlbfs John Cooper is working on getting a patch together for F-12. https://bugzilla.redhat.com/524808 swiotlb should be enabled when VT-d setup fails dwmw2 points out that if VT-d setup fails (because e.g. a bad BIOS is detected), then we should make sure that swiotlb is still enabled. Chris Wright has been working on the fix. All are agreed that this shouldn't be considered a release blocker. https://bugzilla.redhat.com/530389 allow userspace to adjust kvmclock offset Glauber has patches queued upstream to add new kvmclock ioctls and wants them backported to Fedora 12. https://bugzilla.redhat.com/531270 TG3, kvm, ipv6 & tso data corruption bug Rik van Riel discovered this TSO6 related issue with the tg3 driver which is causing traffic from his KVM guest to be corrupted. https://bugzilla.redhat.com/528245 starting a KVM guest before wireless NIC is brought up results in "nf_conntrack: table full, dropping packet." Paul Lambert reports that his machine locks up on boot if he sets his KVM guest to autostart. It looks like it may be related to the wireless connection not being available at that point, but we're still pretty far from narrowing the issue down. == qemu == https://bugzilla.redhat.com/528754 qemu-kvm segfault caused by -soundhw es1370 Gene Czarcinski reported this qemu segfault and managed to obtain a stack trace for it. https://bugzilla.redhat.com/528134 qemu should set O_CLOEXEC on all file descriptors glibc is now forking a ptchown program under certain circumstances, and SELinux issues showed that we are leaking file descriptors into forked processes. Uli claims we should set O_CLOEXEC but Dan, Markus and Kevin disagree. https://bugzilla.redhat.com/472236 KVM guests cannot boot PXE "local" This bug was originally reported against Fedora 10, but apparently is still a problem in Fedora 12 with gPXE. https://bugzilla.redhat.com/528764 qemu's SDL backend doesn't release the console framebuffer on exit Matt Booth discovered this annoyance with qemu's SDL backend. == libvirt == https://bugzilla.redhat.com/527736 libvirt can't create storage pool volumes on a FAT32 hard disk - 'cannot set file owner ... Operation not permitted' Reported by Richard Hughes, looks like libvirt's fs storage backend just isn't handling this case correctly. https://bugzilla.redhat.com/529363 virsh restore fails - failing to re-label the save file ? sVirt related virsh restore failure. == virtinst == https://bugzilla.redhat.com/528947 virtinst: Creating LVM volume on QEMU during VM creation fails A virtinst traceback during cloning, fixed by Cole upstream but needing backporting. == virt-manager == https://bugzilla.redhat.com/530210 virt-manager setfacl() fails on filesystems not mounted with the 'acl' option Our solution for "qemu can't access my homedir" was to use setfacl to give it access. We're scuppered in the case where users have created the filesystem themselves without the 'acl' option listed in the defaults. We need a better error message for this case. https://bugzilla.redhat.com/530948 virt-manager: switching VM to fullscreen pins the window to 1st virtual desktop Stefan Assman is seeing very strange behaviour when trying to use virt-manager in fullscreen mode. https://bugzilla.redhat.com/526952 virt-manager locks up frequently using remote connection Problem reported by both Matt Booth and Chris Wright. https://bugzilla.redhat.com/527714 Entering a non-numeric capacity creates a volume with 10000 MB Another minor virt-manager quirk. https://bugzilla.redhat.com/530448 RFE: virt-manager should provide a means to refresh storage pools virsh has a 'refresh-pool' operation, but virt-manager doesn't expose it. https://bugzilla.redhat.com/529956 16x16 icon in Virtual Machine Details window https://bugzilla.redhat.com/529954 virt-manager dialog buttons do not follow the HIG https://bugzilla.redhat.com/529955 virt-manager host details window does not honor button_have_icons setting https://bugzilla.redhat.com/529657 virt-manager accidentally showing tabs in VM details https://bugzilla.redhat.com/530096 virt-manager host details window does not need a menu Another bunch of reports of virt-manager UI warts from Michael Monreal. https://bugzilla.redhat.com/529217 virt-manager storage pool window appears behind dialog This annoyance in Fedora 11 reported by James Laska is fixed upstream, but needs backporting. https://bugzilla.redhat.com/530178 virt-manager right-click menu for connections should have a delete item Yet another UI quirk, this time reported by Gianluca Cecchi. https://bugzilla.redhat.com/529842 virt-manager: Storage browser should change window title when choosing ISO "Choose Volume" was causing some confusion which could be avoided if it was changed to "Choose ISO". == xen == https://bugzilla.redhat.com/523941 kernel 2.6.31-1[24].fc12 doesn't boot in xen PV guest on 64b host Andrew Jones is busy debugging this and has successfully bisected the problem down to a specific upstream commit. Some debate in the bug as to whether we'd actually slip the release for it. In the end, it turned out to be fixed by 2.6.31.5 for unknown reasons. https://bugzilla.redhat.com/531311 2.6.32 boot fails as Xen PV guest with stackprotector Andrew also uncovered this issue with the Fedora 13 kernel. https://bugzilla.redhat.com/526627 F11 Xen DomU unstable (2.6.30) Reporters seeing consistent Fedora 11 Xen guest lockups, seemingly dependant on the clock source. https://bugzilla.redhat.com/521800 kernel backtrace: possible recursive locking detected on Xen domU A strange recursive locking warning under Xen which apparently doesn't make much sense. Resolved Bugs ============= == misc == https://bugzilla.redhat.com/530015 Add libvirt to Virtualization group Somehow when we split libvirt-client from libvirt, we failed to make sure that libvirt was still being pulled in by the Virtualization package group. https://bugzilla.redhat.com/523914 Mouse does not move in PV Xen guest under RHEL-5.4 Wrangling of this complex blocker bug continued. Peter Hutterer (the evdev maintainer) came up with a way to allow evdev to be configured so it works like before. The fixes are now in rawhide. == kernel == https://bugzilla.redhat.com/524229 Local migration of kvm guest fails in Fedora12 Alpha Marcelo thought this was a pvmmu issue fixed in 2.6.31, but apparently not. After some more testing Marcelo now suspects it's a kvmclock related issue fixed by Glauber upstream. This is now fixed in rawhide. https://bugzilla.redhat.com/526869 Add virtio_blk support cache flush (VIRTIO_BLK_F_FLUSH) Justin has pulled this 2.6.32 patch into Fedora 12. https://bugzilla.redhat.com/525890 Backport virtio patches for optimised virtio-net We've decided to wait until 2.6.32 for these patches. == qemu == https://bugzilla.redhat.com/531419 qemu issue with non-virtio NICs receiving heavy traffic volumes Reported by Scott Tsai on the qemu-devel upstream list, patch was pulled into Fedora 12. https://bugzilla.redhat.com/524734 KVM guest ext3 errors at shutdown when using virtio and a qcow2 backing file Kevin debugged this, posted a patch upstream suitable for stable and we pulled it into F-12. https://bugzilla.redhat.com/528901 qemu-kvm : msrs[] array in kvm_arch_save_regs() too small / may cause stack corruption Ulrich Obergfell provided this excellent analysis of a savevm segfault in Fedora 11. Eduardo sent a fix upstream. https://bugzilla.redhat.com/503156 qemu VNC :: xterm inside VM shows garbled text The version of qemu in Fedora 11 appears to have a broken VNC CopyRect encoding implementation and this has been seriously annoying some users. In the absence of a confirmed fix, we've just disabled the encoding for now. https://bugzilla.redhat.com/527653 /etc/init.d/ksm: line 69: [: -eq: unary operator expected Minor bug with the ksm initscript now fixed in F-12. https://bugzilla.redhat.com/527087 qemu-common postinstall scriptlet uses getent before it's installed Bruce Jerrick points out this packaging error. Fixed in F-12 now. https://bugzilla.redhat.com/526381 qcow2 performance bad under i/o load CLOSED NOTABUG since these were compressed qcow2 images. https://bugzilla.redhat.com/530059 VDE support in qemu? CLOSED CANTFIX until someone packages VDE. https://bugzilla.redhat.com/530136 guestfwd option doesn't allow supplementary chardev options like server,nowait Rich Jones points out a problem with the -net user,guestfwd= option, but since libguestfs doesn't need it anymore and the new -chardev option upstream solves the problem, this is now marked CLOSED UPSTREAM. https://bugzilla.redhat.com/528678 qemu-img does not correctly roundtrip data for VPC image format This minor issue was caught by the libvirt-TCK test suite. Kevin reports that this is NOTABUG, but rather a limitation with the vpc file format. https://bugzilla.redhat.com/526776 Add rtl8209 to gpxe-roms-qemu https://bugzilla.redhat.com/526781 Add rtl8209 to etherboot-zroms-kvm https://bugzilla.redhat.com/526243 Guest PXE booting doesn't work when using ne2k_pci NIC model https://bugzilla.redhat.com/526777 Guest PXE booting doesn't work when using ne2k_pci NIC model Fixed this in F-11 and F-12. == libvirt == https://bugzilla.redhat.com/523158 virsh restore causes libvirtd segfault virExec() Dan Berrange identified the recent change that caused this regression. Fixed now in Fedora 12. https://bugzilla.redhat.com/528162 Fix memory leak in libvirtd in F12 An upstream fix which needed backporting to F12. https://bugzilla.redhat.com/526769 libvirt logrotate settings should roll over weekly instead of daily https://bugzilla.redhat.com/531030 log file sized increased by "compression" Two requests to fix logrotate's handling of libvirt's qemu log files. Fixed for F-12. https://bugzilla.redhat.com/517157 libvirt fails to start guest on NFS even when sebool virt_use_nfs is on We never got around to backporting the fix for this to F-11 and other users had been reporting it. Now fixed in a F-11 update. https://bugzilla.redhat.com/524732 org.libvirt.manage policy kit denial This was due to there being no PolicyKit authentication agent running in the session. == virt-manager == https://bugzilla.redhat.com/461962 virt-manager should support logical pool enumeration Creating a lvm volume group using virt-manager is confusing, but Cole has fixed that upstream now. https://bugzilla.redhat.com/440591 virt-manager: Add 'double click to connect' hint to connection row tooltip Another UI annoyance fixed by Cole upstream. https://bugzilla.redhat.com/526111 virt-manager add storage volume allows an empty name field https://bugzilla.redhat.com/526077 virt-manager toolbar buttons should have tooltips https://bugzilla.redhat.com/524083 virt-manager storage "Allocation" field can be set higher than "Max Capacity" Fixed by virt-manager-0.8.0-7.fc12 == xen == https://bugzilla.redhat.com/521806 Unowned directory /usr/lib64/fs https://bugzilla.redhat.com/496135 missing dependency on PyXML Justin Forbes fixed both of these F-12 xen packaging bugs. https://bugzilla.redhat.com/531429 libvirt failed to disconnect and double free crashes on F12/rawhide with Xen Pasi Karkkainen requested a couple of libvirt xen driver fixes from 0.7.2 be backported to Fedora 12.

1 0

ANNOUNCE: Rawhide virt repo for F11 users
by Mark McLoughlin 30 Oct '09

30 Oct '09

Hey, The idea[1] was discussed here before, so I'll keep this short. We've set up a repository for people running Fedora 11 who would like to test the rawhide/F12 virt packages. To use it, do e.g. $> cat > /etc/yum.repos.d/fedora-virt-preview.repo << EOF [rawvirt] name=Virtualization Rawhide for Fedora 11 baseurl=http://markmc.fedorapeople.org/virt-preview/f11/$basearch/ enabled=1 gpgcheck=0 EOF $> yum update At the moment, it contains the F-12 versions of libvirt and qemu, but as F-12 development continues, it will contain more. I'll send periodic mails to the list detailing the latest updates. Also, this is very much a work-in-progress. The TODO list includes: - get createrepo installed on fedorapeople.org - include debuginfo packages in the repo (need more quota) - find a better location than markmc.fedorapeople.org - enable package builders to upload directly to the repo - automatically do preview build/upload for devel/ builds - allow building against preview packages; e.g. for language bindings Comments most welcome. Help with the TODO list is even more welcome :-) Thanks, Mark. [1] - https://www.redhat.com/archives/fedora-virt/2009-April/msg00154.html

9 76

Strange pygrub issue
by Orion Poplawski 30 Oct '09

30 Oct '09

I've got a strange problem with one of my Fedora 11 guests on CentOS 5.4 host. pygrub does not report the current grub menu, but an old (previous) one. So the system boots to an old kernel, despite everything seeming to be up to date in the guest. Any ideas? -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion(a)cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com

2 1

Trying to move VM efficiently, grub-install, guestfish
by Dennis Jacobfeuerborn 29 Oct '09

29 Oct '09

Hi, What I'm currently experimenting with is how to move a VM between two hosts efficiently. Specifically we are talking Linux guest with virtual disk sizes of 150G of which only about 10-20G are used. Doing a plain "dd" takes a lot of time and is kind of wasteful given that 90% of the logical volume used doesn't contain relevant data. So right now I'm trying to re-create the lvm/filesystem structure from scratch on the target system and then copying the data using tar from the source system. This looks good so far but I cannot start the guest on the target because the boot-sector is missing. I tried using guestfish as described here: http://rwmj.wordpress.com/2009/05/11/rescuing-a-failed-vm-install-with-gues… but when I issue the "grub-install / /dev/sda" all I get is: "libguestfs: error: grub_install: grub-install: /dev/sda does not have any corresponding BIOS drive." I also tried copying the MBR using "dd bs=448 count=1 ..." but that only results in the output of "GRUB" on the target system and then nothing happens. What is the best way to copy/initialize the MBR so the target system boots again? Another method I intend to test is to copy the entire volume with "dd" while the VM is still running then shutting the VM down and using "rsync" to do the final fixup on the logical volumes. That of course assumes that rsync actually works on device nodes which I'm not so sure about. Has anyone experience with this kind of thing? (Just to be clear I'm not talking about cloning the VM requiring making changes to the files in the VM but just simply moving the VM to another host with as little down-time as possible.) Regards, Dennis

2 3

Got Windows guests?
by Richard W.M. Jones 29 Oct '09

29 Oct '09

Hello fellow Fedora, libvirt and libguestfs users, If you have any Windows guests, then you can help Fedora to support Windows guests better by spending a few minutes testing the Windows Registry feature we just added to libguestfs 1.0.75. You will need: - A Windows NT/200x/XP/Vista/7/... guest - Fedora 12 or Fedora Rawhide host - libguestfs-tools >= 1.0.75 (from updates or http://koji.fedoraproject.org/koji/packageinfo?packageID=8391 ) - a few minutes of your time The tests: (1) Run the virt-win-reg commands shown in the following web page, where "MyWinGuest" should be replaced with the name of the Windows guest as known to libvirt: http://libguestfs.org/virt-win-reg.1.html#examples Do the commands run without any errors? Does the output look sensible? If you have several Windows guests, please try as many different sorts as possible! (2) Download the registry binary files and try to convert them to XML: guestfish -i MyWinGuest --ro <<'EOF' download win:\windows\system32\config\software software download win:\windows\system32\config\system system download win:\windows\system32\config\sam sam download win:\windows\system32\config\security security EOF hivexml software > software.xml hivexml system > system.xml hivexml sam > sam.xml hivexml security > security.xml Do those commands run without error? If there's an error, try adding the hivexml -k option. Does the XML look complete? (Try running the XML through tidy -xml -indent -quiet < foo.xml | less ) I hope you don't find any bugs, but if you do: Send a reply to this message, or report a bug in Bugzilla: https://bugzilla.redhat.com/enter_bug.cgi?component=libguestfs&product=Virt… It's useful to include the following details: HIVEX_DEBUG=1 hivexml regfile 2>&1 > log.out The Registry file itself that is failing (but note that Registry files can contain sensitive data). It's also useful to have positive feedback ("it worked!"). Thanks for any testing you can give, and if you have any other suggestions for handling Windows guests from Fedora, please let me know. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-df lists disk usage of guests without needing to install any software inside the virtual machine. Supports Linux and Windows. http://et.redhat.com/~rjones/virt-df/

1 0

F-12 host, can't acquire DHCP lease in guest
by Andrew Overholt 29 Oct '09

29 Oct '09

Hi, I installed rawhide/F-12 x86_64 (network using NetworkManager) and installed and started libvirtd. I then created both F-12 i686 and F-11 i686 guests which went fine. However, neither guest is able to acquire a DHCP lease. Is there some sort of configuration that I didn't do? I don't recall having to do anything special with F-11 as a host. Relevant information below. I can file a bug if that's preferred. Thanks, Andrew ------------------------------------------------------------------------ # virsh net-list --all Name State Autostart ----------------------------------------- default active yes # brctl show bridge name bridge id STP enabled interfaces virbr0 8000.6e8e6e10bdb6 yes vnet0 vnet1 # sysctl net.bridge.bridge-nf-call-iptables net.bridge.bridge-nf-call-iptables = 1 # iptables -L -v -n Chain INPUT (policy ACCEPT 0 packets, 0 bytes) pkts bytes target prot opt in out source destination 2129K 2744M ACCEPT all -- * * 0.0.0.0/0 0.0.0.0/0 state RELATED,ESTABLISHED 3 1070 ACCEPT icmp -- * * 0.0.0.0/0 0.0.0.0/0 7 420 ACCEPT all -- lo * 0.0.0.0/0 0.0.0.0/0 0 0 ACCEPT tcp -- * * 0.0.0.0/0 0.0.0.0/0 state NEW tcp dpt:22 2669 335K REJECT all -- * * 0.0.0.0/0 0.0.0.0/0 reject-with icmp-host-prohibited Chain FORWARD (policy ACCEPT 0 packets, 0 bytes) pkts bytes target prot opt in out source destination 9 2952 ACCEPT all -- * * 0.0.0.0/0 0.0.0.0/0 PHYSDEV match --physdev-is-bridged 0 0 REJECT all -- * * 0.0.0.0/0 0.0.0.0/0 reject-with icmp-host-prohibited Chain OUTPUT (policy ACCEPT 1439K packets, 1128M bytes) pkts bytes target prot opt in out source destination # ps -ef | grep dnsmasq nobody 3148 1 0 09:43 ? 00:00:00 /usr/sbin/dnsmasq --strict-order --bind-interfaces --pid-file=/var/run/libvirt/network/default.pid --conf-file= --listen-address 192.168.122.1 --except-interface lo --dhcp-range 192.168.122.2,192.168.122.254 root 20983 20857 0 13:23 pts/5 00:00:00 grep dnsmasq # ifconfig -a eth0 Link encap:Ethernet HWaddr 00:23:54:9E:57:84 inet addr:10.15.16.87 Bcast:10.15.16.255 Mask:255.255.255.0 inet6 addr: fe80::223:54ff:fe9e:5784/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:2414348 errors:0 dropped:0 overruns:0 frame:0 TX packets:2144264 errors:0 dropped:0 overruns:0 carrier:2 collisions:0 txqueuelen:1000 RX bytes:2888713375 (2.6 GiB) TX bytes:1065980118 (1016.5 MiB) Interrupt:29 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:58553 errors:0 dropped:0 overruns:0 frame:0 TX packets:58553 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:256208705 (244.3 MiB) TX bytes:256208705 (244.3 MiB) virbr0 Link encap:Ethernet HWaddr 6E:8E:6E:10:BD:B6 inet addr:192.168.122.1 Bcast:192.168.122.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:85 errors:0 dropped:0 overruns:0 frame:0 TX packets:30 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:19960 (19.4 KiB) TX bytes:6015 (5.8 KiB) vnet0 Link encap:Ethernet HWaddr 6E:8E:6E:10:BD:B6 inet6 addr: fe80::6c8e:6eff:fe10:bdb6/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:14 errors:0 dropped:0 overruns:0 frame:0 TX packets:3410 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:500 RX bytes:3204 (3.1 KiB) TX bytes:180086 (175.8 KiB) vnet1 Link encap:Ethernet HWaddr 92:AE:24:5C:77:7E inet6 addr: fe80::90ae:24ff:fe5c:777e/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:15 errors:0 dropped:0 overruns:0 frame:0 TX packets:3211 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:500 RX bytes:3546 (3.4 KiB) TX bytes:167128 (163.2 KiB) # cat /proc/sys/net/ipv4/ip_forward 1 I've reloaded the libvirtd service, restarted it, rebooted the guests, etc.

2 2

Porting existing virtual machines to newer Fedora
by Gianluca Sforna 27 Oct '09

27 Oct '09

I'm installing Fedora 12 (full install, no upgrade) on my laptop where I have some virtual machines configured via F11's virt-manager. After installation I'd like to recover from a backup the existing virtual machines; is there any documentation on the proper/recommended/safer method for doing the job? TIA G. -- Gianluca Sforna http://morefedora.blogspot.com http://www.linkedin.com/in/gianlucasforna

3 3

Re:[fedora-virt] Request for confirmation/assistance
by Gianluca Cecchi 26 Oct '09

26 Oct '09

On Fri, 23 Oct 2009 11:26:33 -0400 Gene Czarcinski wrote: > The problem appears to be related to "sound" which does not work currently for > a guest and which I do not use. Hello, sorry Gene, can I jump in and ask for having a way to disable sound (aka not having a sound card through the hardware pool components) for a guest during its creation/installation? When you use Qemu/KVM to host several servers, I think that usually you don't want sound at all..... Thanks and let me know if I have to post a bug as enhancement.... if this is not already planned as a feature. Gianluca

1 0

State of KVM paravirtualization
by Giovanni Tirloni 26 Oct '09

26 Oct '09

Hello, I'm trying to understand where KVM stands regarding paravirtualization. I've seen the patches from 2007 by Ingo Molnar (http://people.redhat.com/mingo/kvm-paravirt-patches/) but couldn't find anything newer. Are they integrated to the mainstream kernel? The reason I'm looking at that is because I've a server at home which doesn't have hardware-assisted virtualization support and running qemu in full virtualization mode is a bit slow. I've tried Xen and it sure is fast but I would like to experiment with KVM. Thank you, -- Giovanni.

5 6

Using the virtio drivers in Windows XP setup
by Richard Hughes 23 Oct '09

23 Oct '09

This may be useful to somebody. To use the virtio drivers in windows XP setup, you have to do the following: * create a 1.44Mb image file * mount it by loopback * format it with vfat * copy the Install/Xp/x86/viostor.sys, Install/Xp/x86/wnet.inf, and the attached txtsetup.oem file to the root of the mounted image * umount the loop device * attach the floppy image as a floppy storage element in the VM's details pane * boot the VM * remember to press F6 when booting the windows xp setup and select the VirtoIO device. The txtsetup.oem was just cobbled together after reading various sources online, and hacking values out of the supplied inf file. Richard. p.s. I'm still unable to install XP using ide, scsi or virtio drivers as it gives the message "Setup was unable to format the partition. The disk maybe damaged." -- any ideas welcome.

4 3

Request for confirmation/assistance
by Gene Czarcinski 23 Oct '09

23 Oct '09

I have a qemu-kvm problem which causes a qemu-kvm guest to "suddenly" crash: https://bugzilla.redhat.com/show_bug.cgi?id=528754 This has happened to me at least 5 times on a F12-rawhide host running qemu-kvm-0.11.0-6.fc12.x86_64 The guests (which crashed) have been F11 and F12 with (at the time) all updates applied. The problem appears to be related to "sound" which does not work currently for a guest and which I do not use. I do not know what action on my part "causes" the problem. When the problem occurs, there have been (IIRC) two guests running (F11 and F12). When the problem happens, the guest console suddenly becomes a white screen with "Guest not running". There are messages in /var/log/messages which say that qemu-kvm has had a segfault. OK, has anyone else seen anything like this? So far, there has been insufficient information to give the developers, etc. any idea of where the problem is. I suspect it would be helpful if someone else had the problem. Gene

1 0

virbr0 messing with iptables rules?
by Kenni Lund 23 Oct '09

23 Oct '09

Hi I just did a full system update on my F11 server, but after a reboot, new rules were appended to my iptables setup. The iptables and ip6tables services are both disabled: # chkconfig |grep ip.*tables ip6tables 0:off 1:off 2:off 3:off 4:off 5:off 6:off iptables 0:off 1:off 2:off 3:off 4:off 5:off 6:off I set my iptables rules in a custom firewall script in /etc/rc.local, which starts out by flushing all rules. Eg. if I run the script manually after boot, it will "fix" the issue. The extra rules appended are: ----- Chain INPUT (policy DROP) target prot opt source destination ACCEPT udp -- anywhere anywhere udp dpt:domain ACCEPT tcp -- anywhere anywhere tcp dpt:domain ACCEPT udp -- anywhere anywhere udp dpt:bootps ACCEPT tcp -- anywhere anywhere tcp dpt:bootps Chain FORWARD (policy DROP) target prot opt source destination ACCEPT all -- anywhere 192.168.122.0/24 state RELATED,ESTABLISHED ACCEPT all -- 192.168.122.0/24 anywhere ACCEPT all -- anywhere anywhere REJECT all -- anywhere anywhere reject-with icmp-port-unreachable REJECT all -- anywhere anywhere reject-with icmp-port-unreachable ----- I suppose that these rules are related to libvirt, since I have a virbr0 interface with the IP-address 192.168.122.1. Apparently the update changed something, so these firewall rules are appended after /etc/rc.local runs my custom firewall script. What is the correct solution to this? In general, isn't it a bad design decision to have a service mess with the iptables rules, instead of doing this through the iptables/ip6tables services? I would not expect other services to mess with my rules, when I explicitly disabled the build-in iptables services. I only use bridged networking for my virtual machines, so I don't suppose that I need the virbr0 interface after all. Can I disable it somewhere? Best Regards Kenni Lund

2 1

Re: [fedora-virt] f12-beta geust install fails on f11
by Dale Bewley 22 Oct '09

22 Oct '09

----- "Cole Robinson" <crobinso(a)redhat.com> wrote: > On 10/22/2009 01:13 AM, Dale Bewley wrote: > > Host is f11 with virt-manager-0.7.0-7.fc11.x86_64 > > > > Using the guest creation wizard, I tried to install f12-beta. I > pointed at a http mirror and allowed virt-manager to detect the OS and > variant, but it failed with: > > > > Unable to complete install: 'OS variant 'fedora12-beta; does not > exist in our dictionary for OS type 'linux'' > Can you try pulling virtinst from updates testing?: > > yum --enablerepo=updates-testing update python-virtinst > > There was a relevant fix that I believe hasn't been pushed to stable > yet. Please report the results here. > Yes that seems to work. It detects f12-beta as OS Generic. I'll go give it some Bodhi karma. Thanks, Cole.

2 1

f12-beta geust install fails on f11
by Dale Bewley 22 Oct '09

22 Oct '09

Host is f11 with virt-manager-0.7.0-7.fc11.x86_64 Using the guest creation wizard, I tried to install f12-beta. I pointed at a http mirror and allowed virt-manager to detect the OS and variant, but it failed with: Unable to complete install: 'OS variant 'fedora12-beta; does not exist in our dictionary for OS type 'linux'' I went back and selected F11, and it works fine so far of course. Would it make sense if the auto-detect could sense alpha and beta releases then punt to a default such as the most recent release of the distribution or some other common denominator? -- Dale Bewley - Unix Administrator - Shields Library - UC Davis GPG: 0xB098A0F3 0D5A 9AEB 43F4 F84C 7EFD 1753 064D 2583 B098 A0F3

2 1

virt-manager and ssh with non root user
by Gianluca Cecchi 21 Oct '09

21 Oct '09

Hello, using F11 + rawvirt repo. I would like to connect from my virt-manager from this F11 to a RH EL 5.4 xen Hypervisor. Is it possible? On the RH EL host, root access via ssh is prevented and I don't want to use a CA certficate at the moment. Also I cannot export DISPLAY env so I cannot directly ssh to the host and run its virt-manager program.... I tried to edit /etc/libvirtd/libvirtd.conf on rh el 5.4 host adding unix_sock_group = "mygroup" and then trying to connect with one user of this group. Is this setting correct as I intended it? When in virt-manager of F11 I try to create new connection, it seems I'm only able to create a session of type xen+ssh://root@myrhel54host/ and not of type xen+ssh://myuser@myrhel54host Any hints? Also, as in virt-manager this connection fails to create (root prevented from ssh directly as said above), I would like to delete it... but I didn't find an option for this; In virt-manager only "connect" and "details" options are available.... these are messages in virsh session when setting LIBVIRT_DEBUG=1 virsh # connect xen+ssh://unix@myrhel54host 17:23:44.911: debug : virConnectClose:1291 : conn=0x65de70 17:23:44.911: debug : virUnrefConnect:257 : unref connection 0x65de70 1 17:23:44.911: debug : remoteIO:7421 : Do proc=2 serial=3 length=28 wait=(nil) 17:23:44.911: debug : remoteIO:7483 : We have the buck 2 0x67db20 0x67db20 17:23:44.912: debug : remoteIODecodeMessageLength:7032 : Got length, now need 56 total (52 more) 17:23:44.912: debug : remoteIOEventLoop:7347 : Giving up the buck 2 0x67db20 (nil) 17:23:44.912: debug : remoteIO:7514 : All done with our call 2 (nil) 0x67db20 17:23:44.912: debug : virReleaseConnect:214 : release connection 0x65de70 17:23:44.912: debug : virConnectOpenAuth:1273 : name=xen+ssh://unix@myrhel54host, auth=0x7fb212bc40a0, flags=0 17:23:44.912: debug : do_open:1042 : name "xen+ssh://unix@myrhel54host" to URI components: scheme xen+ssh opaque (null) authority (null) server myrhel54host user unix port 0 path (null) 17:23:44.912: debug : do_open:1052 : trying driver 0 (Test) ... 17:23:44.912: debug : do_open:1058 : driver 0 Test returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 1 (Xen) ... 17:23:44.912: debug : do_open:1058 : driver 1 Xen returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 2 (OPENVZ) ... 17:23:44.912: debug : do_open:1058 : driver 2 OPENVZ returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 3 (PHYP) ... 17:23:44.912: debug : do_open:1058 : driver 3 PHYP returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 4 (VBOX) ... 17:23:44.912: debug : do_open:1058 : driver 4 VBOX returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 5 (ESX) ... 17:23:44.912: debug : do_open:1058 : driver 5 ESX returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 6 (remote) ... 17:23:44.912: debug : doRemoteOpen:535 : proceeding with name = xen:// 17:23:44.912: debug : virExecWithHook:609 : ssh -l unix myrhel54host nc -U /var/run/libvirt/libvirt-sock 17:23:44.917: debug : remoteIO:7421 : Do proc=66 serial=0 length=28 wait=(nil) 17:23:44.918: debug : remoteIO:7483 : We have the buck 66 0x6bde80 0x6bde80 17:23:44.924: debug : remoteIOEventLoop:7368 : Giving up the buck due to I/O error 66 0x6bde80 (nil) 17:23:44.924: debug : do_open:1058 : driver 6 remote returned ERROR 17:23:44.924: debug : virUnrefConnect:257 : unref connection 0x65e260 1 17:23:44.924: debug : virReleaseConnect:214 : release connection 0x65e260 error: Failed to connect to the hypervisor error: cannot recv data: Connection reset by peer Thanks, Gianluca

3 3

Re: [fedora-virt] virtio block driver performance versus ide
by Richard W.M. Jones 21 Oct '09

21 Oct '09

On Wed, Oct 21, 2009 at 08:58:10AM -0600, Thomas S Hatch wrote: > Wow, the rotational change greatly speeds up my vms, as much as 15 fold. > > At the risk of sounding like a complete n00b, is there a way to persistently > change the settings in /sys? I am of course familiar with the sysctl.conf > for /proc/sys. Or is this something that should be changed at the driver > level? It has apparently already been changed (see the new kernels in that Bugzilla entry). In any case, you can't make the change permanent, but you can add the commands to /etc/rc.local so they get executed at each guest boot which has just about the same effect. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones libguestfs lets you edit virtual machines. Supports shell scripting, bindings from many languages. http://et.redhat.com/~rjones/libguestfs/ See what it can do: http://et.redhat.com/~rjones/libguestfs/recipes.html

1 0

virtio block driver performance versus ide
by Dale Bewley 21 Oct '09

21 Oct '09

I assumed the virtio block driver to be faster for a kvm guest than ide, but I'm seeing quite the opposite. Is this expected? I normally create guests with a wrapper script around virt-install which creates guests configured to use IDE disks: f11-db1.xml- <emulator>/usr/bin/qemu-kvm</emulator> f11-db1.xml: <disk type='file' device='disk'> f11-db1.xml- <source file='/var/lib/libvirt/images//f11-db1.img'/> f11-db1.xml- <target dev='hda' bus='ide'/> f11-db1.xml: </disk> I am currently creating a guest with the virt-manager wizard which apparently uses virtio by default: f11-archivist.xml- <emulator>/usr/bin/qemu-kvm</emulator> f11-archivist.xml: <disk type='file' device='disk'> f11-archivist.xml- <source file='/var/lib/libvirt/images//f11-archivist.img'/> f11-archivist.xml- <target dev='vda' bus='virtio'/> f11-archivist.xml: </disk> The filesystem formatting and installation are going very very slowly. Host is 2.6.29.6-213.fc11.x86_64 4x quad core. Guests are also fc11 x86_64. BTW, I've updated to virt-manager-0.7.0-7, libvirt-0.6.2-18, qemu-kvm-0.10.6-6, and kernel-2.6.30.8-64 just before creating this guest but have not restarted libvirtd or rebooted the host. -- Dale Bewley - Unix Administrator - Shields Library - UC Davis GPG: 0xB098A0F3 0D5A 9AEB 43F4 F84C 7EFD 1753 064D 2583 B098 A0F3

5 5

f12 guest memory problems?
by Gianluca Cecchi 21 Oct '09

21 Oct '09

Hello, F11 x86_64 + rawvirt with latest updates I have a f12 near beta x86_64 kvm guest where now I'm doing "yum update" in it and basically I'm only in gui session without doing anything else Some strange things: - in virt-manager set 400Mb for VM (both the allocated, change, max allocation values... btw how to use them???) my guest sees only about 330Mb [gcecchi@localhost ~]$ free total used free shared buffers cached Mem: 344212 252944 91268 0 4268 26908 -/+ buffers/cache: 221768 122444 Swap: 950264 179092 771172 With dmesg on guest I see something like this: Memory: 329452k/409536k ...... 79696k reserved The 79696k reserved seems to fill the gap... is this a feature in f12 or based on it being a guest? - As you see, the guest is swapping about 180Mb, and actually it is very slowly responsive. Shouldn't it use some of buffers/cache before swapping so much? Is this expected during a yum update (only 34 packages for an overall of 72MB)? Or problem with F12 itself or virtualization layer? Thanks Gianluca

1 0

Re: Oops
by Mark McLoughlin 21 Oct '09

21 Oct '09

Hi Gene, On Tue, 2009-10-20 at 19:39 -0400, Gene Czarcinski wrote: > F12 beta was available so I downloaded the DVD iso ... that was a waste of > bandwidth since it was identical to F12-Beta RC2. > > Anyway, I had already installed a couple of copies of RC2 into qemu-kvm guests > so it was time to install this on bare-metal. > > My test system already had F12-rawhide installed with all current updates but, > afterall, this is a test of the install ... so I did another install ... my > test system supports four, separate bootable systems (combinations of /boot > partitions and LVM logical volumes for root). > > OK, boot it up and start configuring it, updating (I did a DVD-only install), > etc. Well, one of the things this system does is run qemu-kvm so I had > selected Virtualization during the install. My guest disk images are all on a > separate logical volume which I mount on /.var/lib/libvirt/images but the > system definition/configuration files are separate. When I went to copy them > from the old system I found the was no place to copy them to ... /etc/libvirt/ > was missing. Further checking showed that the libvirt rpm was not installed > ... this is easy enought to fix but it should not have happened ... it should > have been installed when I selected virtualization. Oops, indeed. I've filed this here: https://bugzilla.redhat.com/530015 Thanks for finding this, it's a very serious issue that could have crept by unnoticed. Cheers, Mark.

1 0

#529711 - qemu: could not open serial device 'pty'
by Gianluca Cecchi 20 Oct '09

20 Oct '09

On Mon, 19 Oct 2009 20:15:51 +0200 Gianluca Varisco wrote: > Hi guys, > > Am I the only one having this issue with qemu/KVM and rawhide? > > https://bugzilla.redhat.com/show_bug.cgi?id=529711 <https://bugzilla.redhat.com/show_bug.cgi?id=529711> > > (It is not a SELinux-related problem) > > Cheers, I would advise to also follow what written at link: https://fedoraproject.org/wiki/Reporting_virtualization_bugs In particular contents of virt-manager.log and posting output when setting LIBVIRT_DEBUG=1 btw I presume that also the version of virt-manager is the latest one, correct? HIH, Gianluca

1 0

#529711 - qemu: could not open serial device 'pty'
by Gianluca Varisco 19 Oct '09

19 Oct '09

Hi guys, Am I the only one having this issue with qemu/KVM and rawhide? https://bugzilla.redhat.com/show_bug.cgi?id=529711 (It is not a SELinux-related problem) Cheers, -- Gianluca Varisco, RHCE | Office: +39 02 9737 4652 Red Hat Italia | Fax: +39 02 669 3111 Via Antonio Da Recanate 1 | Mobile: +39 333 574 0934 20124 Milano | eMail: gvarisco(a)redhat.com

1 0

'eject' on a KVM guest
by Matthew Booth 19 Oct '09

19 Oct '09

What does the eject (to eject a CD) command do to a KVM guest? It succeeds, and the result is that the device can no longer be mounted. However, looking at the domain XML reveals it hasn't disconnected the backing file. Matt -- Matthew Booth, RHCA, RHCSS Red Hat Engineering, Virtualisation Team M: +44 (0)7977 267231 GPG ID: D33C3490 GPG FPR: 3733 612D 2D05 5458 8A8A 1600 3441 EA19 D33C 3490

3 4

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

virt October 2009