- Devel - libvirt List Archives

[Libvir] [PATCH] Topology fix for "no cpus"
by beth kon 15 Oct '07

15 Oct '07

I was able to test on a 128-way NUMA box and found a bug. My code did not handle the case of no cpus being associated with a node. I decided to represent (pretty straightforward decision :-) no cpus as follows in the xml... <cell id='2'> <cpus num='0'> </cpus> </cell> Here is the patch... Signed-off-by: Beth Kon <eak(a)us.ibm.com> -- Elizabeth Kon (Beth) IBM Linux Technology Center Open Hypervisor Team email: eak(a)us.ibm.com

3 3

[Libvir] PATCH: Fix const-ness for attach/detach device APIs
by Daniel P. Berrange 15 Oct '07

15 Oct '07

virDomainAttachDevice and virDomainDetachDevice both take a char * for the XML description, but this is mistakenly not declared to be const. This patch fixes the public header files & all the internal drivers. NB, yes this is in the public API, no it won't break any apps since these are input params only. Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|

3 2

[Libvir] Core dump while executing virsh in RHEL5 .
by Veerendra 15 Oct '07

15 Oct '07

Hi all, I have installed <mailto:libvir-list@redhat.com>libvirt-0.3.3-1, on my RHEL 5 64 bit machine(KerVer 2.6.18-8.el5xen). But I am getting core-dump , while executing virsh command. Could anyone please help me to solve this? #gdb /usr/bin/virsh core Core was generated by `virsh'. Program terminated with signal 11, Segmentation fault. #0 0x000000393686ca27 in malloc_consolidate () from /lib64/libc.so.6 (gdb) where #0 0x000000393686ca27 in malloc_consolidate () from /lib64/libc.so.6 #1 0x000000393686eea2 in _int_malloc () from /lib64/libc.so.6 #2 0x00000039368706dd in malloc () from /lib64/libc.so.6 #3 0x000000393685eb4a in __fopen_internal () from /lib64/libc.so.6 #4 0x000000393682cb5f in read_alias_file () from /lib64/libc.so.6 #5 0x000000393682d09e in _nl_expand_alias () from /lib64/libc.so.6 #6 0x000000393682b93e in _nl_find_domain () from /lib64/libc.so.6 #7 0x000000393682b2ff in __dcigettext () from /lib64/libc.so.6 #8 0x000000393687530c in strerror_r () from /lib64/libc.so.6 #9 0x000000393687514e in strerror () from /lib64/libc.so.6 #10 0x00002aaaaab018f2 in __virConfReadFile () from /usr/lib64/libvirt.so.0 #11 0x00002aaaaab021ec in __virConfReadFile () from /usr/lib64/libvirt.so.0 #12 0x00002aaaaaadd44b in virInitialize () from /usr/lib64/libvirt.so.0 #13 0x000000000040a65e in ?? () #14 0x000000393681d8a4 in __libc_start_main () from /lib64/libc.so.6 #15 0x00000000004033b9 in ?? () #16 0x00007fffd2df95c8 in ?? () #17 0x0000000000000000 in ?? () I have installed the following rpms from libvirt.org/libvirt # rpm -qa | grep virt libvirt-devel-0.3.3-1 libvirt-0.3.3-1 libvirt-debuginfo-0.3.3-1// * With Regards Veerendra C*

4 9

[Libvir] PATCH: Allow libvirtd configuration for QEMU VNC server
by Daniel P. Berrange 13 Oct '07

13 Oct '07

The QEMU VNC server spawned by the QEMU driver in libvirt is hardcoded to start off on 127.0.0.1, unless the person creating a guest overides this in the XML passed to libvirt. If wanting to setup off-host VNC access, it is much much more convenient to be able to set the system wide default to be 0.0.0.0 than to specify 0.0.0.0 for each VM created. In addition, it is desirable to be able to configure use of TLS and x509 certificates for the VNC servers system wide. In Xen world this is already possible through the /etc/xen/xend-config.sxp configuration file. In QEMU world, libvirtd takes the place of XenD. The /etc/libvirt/libvirtd.conf file though is for the daemon as a whole. There is no config file for the QEMU driver in libvirt itself. So this patch extends the QEMU driver to be able to load /etc/libvirt/qemu.conf and configure a handful of options. - vnc_listen = "0.0.0.0" - the address for VNC to listen on. Defaults to 127.0.0.1 for security sake - vnc_tls = 1 - turn on use of TLS extension - vnc_tls_x509_verify - request client certificates for auth - vnc_tls_cert_dir - the location of the VNC server certs. Defaults to /etc/pki/libvirt-vnc The code for building command line args for the -vnc flag to QEMU takes care to look at these options & add approprivate syntax. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|

2 2

[Libvir] how to setup usb/serial ports with libvirt & kvm (Fedora 7)
by John Davey 11 Oct '07

11 Oct '07

I am experimenting with the following Redhat 9 kvm on Fedora 7. How do I add a USB device with Libvirt? Specifically I'd like to add a usb-serial converter for logging data. "Man qemu-kvm" shows -usbdevice devname to add a USB device but I cannot find out how to do this with libvirt. Regards John <domain type='kvm'> <name>Rho</name> <memory>458752</memory> <currentMemory>458752</currentMemory> <vcpu>1</vcpu> <os> <type>hvm</type> <boot dev='hd'/> </os> <features> <acpi/> </features> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>restart</on_crash> <devices> <emulator>/usr/bin/qemu-kvm</emulator> <disk type='file' device='disk'> <source file='/var/lib/xen/images/rho.img'/> <target dev='hda'/> </disk> <interface type='bridge'> <mac address='06:00:CF:DE:01:A2'/> <source bridge='eth0'/> <target dev='rho0'/> </interface> <input type='mouse' bus='ps2'/> <graphics type='vnc' port='-1' listen='127.0.0.1'/> </devices> </domain>

2 2

[Libvir] [PATCH] qemu: reset migration source if restore fails
by Jim Paris 10 Oct '07

10 Oct '07

With the latest KVM I'm having some issues with save/restore, but I think it's a kvm issue and not libvirt. The problems did expose a libvirt bug, though; patch below. -jim --- A failed restore of a configured VM would leave the migration source set incorrectly, preventing the VM from being started normally. Always clear migration source, then deal with failures. src/qemu_driver.c | 12 ++++++------ 1 files changed, 6 insertions(+), 6 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index 87b66ad..bff17b0 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -2083,6 +2083,7 @@ static int qemudDomainRestore(virConnectPtr conn, struct qemud_vm_def *def; struct qemud_vm *vm; int fd; + int ret; char *xml; struct qemud_save_header header; @@ -2161,18 +2162,17 @@ static int qemudDomainRestore(virConnectPtr conn, /* Set the migration source and start it up. */ snprintf(vm->migrateFrom, sizeof(vm->migrateFrom), "stdio"); vm->stdin = fd; - - if (qemudStartVMDaemon(conn, driver, vm) < 0) { + ret = qemudStartVMDaemon(conn, driver, vm); + close(fd); + vm->migrateFrom[0] = '\0'; + vm->stdin = -1; + if (ret < 0) { qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, "failed to start VM"); if (!vm->configFile[0]) qemudRemoveInactiveVM(driver, vm); - close(fd); return -1; } - close(fd); - vm->migrateFrom[0] = '\0'; - vm->stdin = -1; /* If it was running before, resume it now. */ if (header.was_running) { -- 1.5.3.rc4

3 2

[Libvir] Win Pro 64 dies in second half of windows guest installation
by Lawrence Nitz 10 Oct '07

10 Oct '07

When I set up a guest VM using virtp-install on a Fedor a 7 system, the process gets as far as reading the raw device (/dev/scd0) for the install disk, writes the basic windows data onto the disk, and then reboots to complete the Windows installation. At this point the installation hangs, with a diagnostic that says the file cannot be found. The same error occurs whether the install is from the hard disk image or from the cd-rom. The install gets half done, and has nothing to read when the install program restarts itself. In the Virt manager the edit command allows me to edit the machine, but not to add back the cd drive, only add network cards and hard disk partitions. Any help by a more experienced virt-install user would be much appreciated. Larry Nitz.

1 0

[Libvir] Problems with defineXML in python..
by Mark Dehus 10 Oct '07

10 Oct '07

I am having some problems getting libvirt under python to work properly. I can setup new domains by using the createLinux() function, but when I attempt to use defineXML() it fails with no reason given other then "virDomainDefineXML() failed". I am using the same XML file for both functions. This is on a Ubuntu 7.04 server running python 2.5 & libvirt 0.1.8 (from the ubuntu universe repo). Here is my test code: import libvirt import sys import xml import random try: conn = libvirt.open(None) except Exception, e: print e if conn == None: print 'Failed to open connection to the hypervisor' sys.exit(1) create = True if create: xmlfile = open('webserv.xml') xmldesc = xmlfile.read() xmlfile.close() try: # webserv = conn.createLinux(xmldesc, 0) webserv = conn.defineXML(xmldesc) except Exception, e: print 'Failed to create webserv: %s' % e sys.exit(1) if webserv is None: print 'whoops this shouldnt happen!' conn.create(webserv) try: print "Domain 0: id %d running %s" % (webserv.ID(), webserv.OSType()) print webserv.info() print webserv.XMLDesc(0) except Exception, e: print e #webserv.shutdown() And the XML file: <domain type='xen' > <name>webserv</name> <os> <type>linux</type> <kernel>/boot/vmlinuz-2.6.19-4-generic-amd64</kernel> <initrd>/boot/initrd.img-2.6.19-4-generic-amd64</initrd> <root>/dev/hda1</root> </os> <memory>1024</memory> <vcpu>1</vcpu> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>restart</on_crash> <devices> <interface type='bridge'> <source bridge='xenbr0'/> <mac address='00:16:3e:24:27:a5'/> <script path='vif-bridge'/> </interface> <graphics type='vnc' port='5902'/> <disk type='file' device='disk'> <driver name='tap' type='aio'/> <source file='/virtual/images/webserv/webserv.img'/> <target dev='hda1'/> </disk> <console tty='/dev/pts/4'/> </devices> </domain> Am I missing something simple, or is there a bug? Thanks!

2 1

[Libvir] [PATCH] Topology fix for "no cpus"
by beth kon 08 Oct '07

08 Oct '07

Here is a resubmission of this patch, corrected. -- Elizabeth Kon (Beth) IBM Linux Technology Center Open Hypervisor Team email: eak(a)us.ibm.com

2 1

[Libvir] [PATCH] Found minor bug in topology code
by beth kon 05 Oct '07

05 Oct '07

I found a problem with an error path... I was cutting off incrementing of a value in the loop then checking if it was too big after the loop, so the way it was structured, it could never happen. Here is the fix. -- Elizabeth Kon (Beth) IBM Linux Technology Center Open Hypervisor Team email: eak(a)us.ibm.com

2 1