- Devel - libvirt List Archives

[libvirt] [PATCH 0/2]: Two small fixes
by Luiz Capitulino 12 Apr '10

12 Apr '10

Two small issues I found while testing libvirt under QMP, there is also a third I won't fix today: dommemstat doesn't work under QMP.

2 5

[libvirt] [PATCH 0/2] Enhance checking of supported flags in drivers
by Jiri Denemark 12 Apr '10

12 Apr '10

I think we need a simple and consistent way of checking whether flags passed to API entry points in drivers are valid/supported or not. So far some entry points don't check flags at all and some checks for 0 when no flags are expected/supported. The first patch introduces a macro which can be called with a whitelist of supported flags (more info about it inside the patch). I don't particularily like introducing a new header file but non of existing header files looked like a good place to stick this macro in. If you think such a place exist, I'll be very happy to adapt the patch. The second patch changes all API calls introduced since the last release to use the macro for checking valid flags. I'm not sure if we can change existing API calls to do the same as it could break some badly written apps which pass invalid flags. My opinion is that we may often do that as such apps don't use the API in a documented/valid way. But even if my thinking is correct, we don't want to change it before 0.8.0. Jirka Jiri Denemark (2): Introduce virCheckFlags for consistent flags checking Use virCheckFlags for APIs added in 0.8.0 src/Makefile.am | 3 +- src/esx/esx_driver.c | 44 ++++++++++++------------- src/nwfilter/nwfilter_driver.c | 5 ++- src/qemu/qemu_driver.c | 69 +++++++++++++++++++--------------------- src/storage/storage_driver.c | 7 +--- src/util/checks.h | 37 +++++++++++++++++++++ src/vbox/vbox_tmpl.c | 42 +++++++++++++++++------- src/xen/xend_internal.c | 5 +++ 8 files changed, 133 insertions(+), 79 deletions(-) create mode 100644 src/util/checks.h

2 3

[libvirt] [PATCH] nwfilter: Fix memory leak on daemon init and shutdown
by Stefan Berger 12 Apr '10

12 Apr '10

This patch fixes a memory leak on daemon init and shutdown. The module was initialized twice and not shut down. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com>

2 2

[libvirt] libvirt Error Domain didn't show up
by Roche, Christian (EXT-Other - MA/Rabat) 12 Apr '10

12 Apr '10

Hi there, I'm trying to use kvm on CentOS 5.4 64-bit to virtualize Windows Server 2003. I've been able to install and start the system using 1GB of memory, however when I try to increase the amount of memory available to the VM to 4GB by editing /etc/libvirt/qemu/windows2003.xml, I get the following message when starting the VM through virt-manager: --------------------------------- Error starting domain: internal error Domain windows2003 didn't show up Traceback (most recent call last): File "/usr/share/virt-manager/virtManager/engine.py", line 498, in run_domain vm.startup() File "/usr/share/virt-manager/virtManager/domain.py", line 573, in startup self.vm.create() File "/usr/lib64/python2.4/site-packages/libvirt.py", line 287, in create if ret == -1: raise libvirtError ('virDomainCreate() failed', dom=self) libvirtError: internal error Domain windows2003 didn't show up --------------------------------- Here is the content of /var/log/libvirt/qemu/windows2003.log : --------------------------------- LC_ALL=C PATH=/sbin:/usr/sbin:/bin:/usr/bin /usr/bin/qemu-system-x86_64 -S -M pc -no-kqemu -m 4096 -smp 4 -monitor pty -pidfile /var/run/libvirt/qemu//windows2003.pid -localtime -boot c -hda /dev/mapper/Raid5-Windows -cdrom /var/lib/libvirt/images/viostor.iso -net nic,macaddr=54:52:00:07:a9:b6,vlan=0 -net tap,fd=16,script=,vlan=0,ifname=vnet0 -serial pty -parallel none -usb -usbdevice tablet -vnc 127.0.0.1:0 -k fr QEMU PC emulator version 0.9.0, Copyright (c) 2003-2007 Fabrice Bellard usage: qemu [options] [disk_image] 'disk_image' is a raw hard image image for IDE hard disk 0 Standard options: -M machine select emulated machine (-M ? for list) -fda/-fdb file use 'file' as floppy disk 0/1 image -hda/-hdb file use 'file' as IDE hard disk 0/1 image blah blah blah --------------------------------- It looks like qemu is unhappy about the memory parameter. Here is my config: >uname -a Linux silver 2.6.18-164.15.1.el5 #1 SMP Wed Mar 17 11:30:06 EDT 2010 x86_64 x86_64 x86_64 GNU/Linux >grep virt /var/log/rpmpkgs libvirt-0.6.3-20.1.el5_4.x86_64.rpm libvirt-python-0.6.3-20.1.el5_4.x86_64.rpm python-virtinst-0.400.3-5.el5.noarch.rpm virt-manager-0.6.1-8.el5.x86_64.rpm Any idea ? Thanks, Christian Roche NSN/Rabat +212 6 61 44 11 71

1 0

Re: [libvirt] Domain not stable while created...
by Sankamesh 12 Apr '10

12 Apr '10

Thanks Mr. Stefan Will that crash mainly be due to the kernel ? Since am not sure about that,else I should try with different image. When I use via a config file using "xm create libvirt-ttylinux.conf ", then the domain runs stable.... But when I use the virsh -c xen:/// define libvirt-ttylinux.xml" and "virsh start ttyimage, it crashes after sometime. The ibvirt-ttylinux.conf file :: -------------------------------------------------------------------------------------------------------------------------------- kernel = "/boot/vmlinuz-2.6.18-164.2.1.el5xen" ramdisk= "/boot/initrd-2.6.18-164.2.1.el5xen.img" memory = 64 name = "ttylinux" disk = ['file:/usr/local/ttylinux-xen-libvirt/ttylinux-xen.img,sda1,w'] root = "/dev/sda1 ro" vif = [''] -------------------------------------------------------------------------------------------------------------------------------- I give the same Xen kernel parameter here in xml file, and use virsh to create domain --------------------------------------------------------------------------------------------------------------------------------------- <?xml version="1.0" encoding="utf-8"?> <domain type='xen'> <name>ttyimage</name> <os> <type>linux</type> <kernel>/boot/vmlinuz-2.6.18-164.2.1.el5xen</kernel> <initrd>/boot/initrd-2.6.18-164.2.1.el5xen.img</initrd> </os> <memory>65536</memory> <vcpu>1</vcpu> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <graphics type='vnc' port='5900'/> <disk type='file'> <source file='/usr/local/ttylinux-xen-libvirt/ttylinux-xen.img' /> <target dev='sda1' /> </disk> <interface type='bridge'> <source bridge='virbr0' /> <mac address='00:1d:60:ec:ae:1c' /> <target dev='testnimb-0' /> </interface> </devices> </domain> --------------------------------------------------------------------------------------------------------------------------------------- My Boot grub file data:: --------------------------------------------------------------------------------------------------------------------------------------title Scientific Linux (2.6.18-164.2.1.el5xen) root (hd0,0) kernel /xen.gz-2.6.18-164.2.1.el5 module /vmlinuz-2.6.18-164.2.1.el5xen ro root=LABEL=/1 rhgb quiet module /initrd-2.6.18-164.2.1.el5xen.img -------------------------------------------------------------------------------------------------------------------------------------- Why is this difference between the xm (when i destroy the image,domain is also destroyed) and virsh (domain is not destroyed while destroying the image) . I doubt that am making some mistake with the xml file .... Is that true? Kindly give some suggestions Thanks ________________________________ libvir-list-bounces(a)redhat.com wrote on 04/10/2010 09:47:17 AM: > libvir-list > > Hello, > > Thanks Alex > > I created a domain using "virsh -c xen:/// define libvirt-ttylinux.xml" and > "virsh start ttyimage" for a simple ttyimage. > > Problem I face is that the image shutsdown immedialtely once virsh > start guest is typed. > I checked the Xend Logs > ---------------------------------------------------------------------------------------------------------- > ---[2010-04-11 07:08:02 xend 3155] DEBUG (DevController:116) > DevController: writing {'frontend-id': '6', 'domain': 'ttyimage', [...] > [2010-04-11 07:08:02 xend 3155] DEBUG (vfbif:11) Spawn: ['/usr/lib/ > xen/bin/qemu-dm', '-M', 'xenpv', '-d', '6', '-domain-name', > 'ttyimage', '-vnc', '127.0.0.1:0', '-vncunused', '-k', 'en-us'] > [2010-04-11 07:08:02 xend.XendDomainInfo 3155] DEBUG > (XendDomainInfo:992) Storing VM details: {'shadow_memory': '0', > 'uuid': 'd5e916b0-bc2d-01f9-8156-4808c754f67a', 'on_reboot': > 'restart', 'start_time': '1270937282.29', 'on_poweroff': 'destroy', > 'name': 'ttyimage', 'xend/restart_count': '0', 'vcpus': '1', > 'vcpu_avail': '1', 'memory': '64', 'on_crash': 'destroy', 'image': > '(linux (kernel /boot/vmlinuz-2.6.18-164.2.1.el5xen))', 'maxmem': '64'} > [2010-04-11 07:08:02 xend.XendDomainInfo 3155] DEBUG > (XendDomainInfo:1027) Storing domain details: {'console/ring-ref': > '65674', 'console/port': '2', 'name': 'ttyimage', 'console/limit': > '1048576', 'vm': '/vm/d5e916b0-bc2d-01f9-8156-4808c754f67a', > 'domid': '6', 'cpu/0/availability': 'online', 'memory/target': > '65536', 'store/ring-ref': '65675', 'store/port': '1'} > [2010-04-11 07:08:02 xend.XendDomainInfo 3155] DEBUG > (XendDomainInfo:1249) XendDomainInfo.handleShutdownWatch > [2010-04-11 07:08:02 xend 3155] DEBUG (DevController:158) Waiting > for devices vif. > [2010-04-11 07:08:02 xend 3155] DEBUG (DevController:164) Waiting for 0. > [2010-04-11 07:08:02 xend 3155] DEBUG (DevController:509) > hotplugStatusCallback /local/domain/0/backend/vif/6/0/hotplug-status. > [2010-04-11 07:08:51 xend 3155] INFO (XendDomain:387) Domain > ttyimage (6) unpaused. > [2010-04-11 07:08:54 xend.XendDomainInfo 3155] WARNING > (XendDomainInfo:1177) Domain has crashed: name=ttyimage id=6. You domain crashed here... Is the VM image and kernel in that image known to work? Regards, Stefan Send free SMS to your Friends on Mobile from your Yahoo! Messenger. Download Now! http://messenger.yahoo.com/download.php

2 1

[libvirt] about qemu monitor mode.
by Osier Yang 12 Apr '10

12 Apr '10

Hello there libvirt compiles JSON mode as the default qemu monitor mode if yajl-devel is available. As a result of this solution, the user has no way to change into the text mode. And because the qemu QMP always lacks commands currently, such as "savevm", it will prevent us creating the snapshot of domain, especially for us who want to do some testing. What I think is could we add the configuration item in qemu.conf? so that the user can choose which mode to use, such as: monitor_mode=json #monitor_mode=text Regards osier

2 1

[libvirt] libvirt: C++ or Java
by Gokhale, Shashibhushan 11 Apr '10

11 Apr '10

Hello, I am planning to develop a management application for managing Redhat Xen and Redhat KVM hypervisors. I am evaluating the Java bindings versus the C++ version of libvirt. As part of evaluation, I could not find meaningful answers to the questions below. Hence, I thought that people on this mailing list would be the most appropriate to comment on them. Can you please let me know your views? 1. Are all functions that are available in C API available in Java? 2. Are Java bindings added along with every enhancement to C API? 3. If answer to above question is negative, typically how long does it takes for the addition of Java bindings for the enhancement? 4. Most of libvirt based third party management applications use which of Java bindings or C API? 5. What are your good/bad experiences while using either of C API or Java bindings? 6. What is your concluding personal recommendation? And finally, since libvirt is a Redhat Emerging Technology project, does Redhat or any other third party service providers provide paid assistance, support? Thanks, Shashi

1 0

[libvirt] Domain not stable while created...
by Sankamesh 10 Apr '10

10 Apr '10

Hello, I am using Scientific Linux 5.4 and xen. I want to deploy a sample ttyimage by creating a domain using virsh create. My configuration file is : ----------------------------------------------------------------------------------------------------- <?xml version="1.0" encoding="utf-8"?> <domain type='xen'> <name>ttyimage</name> <os> <type>linux</type> <kernel>/boot/vmlinuz-2.6.18-164.2.1.el5xen</kernel> </os> <memory>65536</memory> <vcpu>1</vcpu> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <graphics type='vnc' port='5900'/> <disk type='file'> <source file='/usr/local/ttylinux-xen-libvirt/ttylinux-xen.img' /> <target dev='sda' /> </disk> <interface type='bridge'> <source bridge='virbr0' /> <mac address='00:1d:60:ec:ae:1c' /> <target dev='testnimb-0' /> </interface> </devices> </domain> ------------------------------------------------------------------------------------------------------- when I use "virsh -c "xen:///" create libvirt-ttylinux.xml " to create a domain, the domain is crated and i can view the image (in paused mode) using the virt-manager. and when I resume the image it ,disappears (domain gets destroyed) after nearly 10 seconds..... Further when I use vncviewer localhost::5900 when the image is running (during the 10 sec) ,i can view the console. But it closes once the domain is destroyed. Can someone explain why the image is not stable??? how can I make the domain to be stable???? Thanks in advance The INTERNET now has a personality. YOURS! See your Yahoo! Homepage. http://in.yahoo.com/

3 3

Re: [libvirt] [RFC] Unify KVM kernel-space and user-space code into a single project
by Antoine Martin 09 Apr '10

09 Apr '10

Hi, I am moving this thread here as this seems more appropriate. Sorry it has taken so long.. Here are 2 things that really get in the way of moving my existing installations to libvirt: * I tend to store much meta data with each VM instance: it can be things like ownership (contact details as text), monitoring info (sms phone numbers), backup (list of paths), firewall rules (custom syntax, with failover rules, etc), etc. At the moment, these extra bits of information consist of just a few optional lines of shell in each VM's definition file. I can extend these whenever I need, enumerate the VMs using the standard mechanism and trigger my specific actions as needed (firewall rules, backup or whatever). I see no way of doing this with libvirt. But please correct me if I am wrong. * not everything is exposed via libvirt: virsh can retrieve vncdisplay but libvirt (or at least the python bindings) does not. How come? This happens to be one thing I need for writing a libvirt backend for my virtual desktop software. Cheers Antoine Antoine Martin wrote: > Hi Daniel, > > I'll take a look and get back to you asap. > > Cheers > Antoine > > Daniel P. Berrange wrote: >> On Tue, Mar 23, 2010 at 03:00:28AM +0700, Antoine Martin wrote: >>> On 03/23/2010 02:15 AM, Anthony Liguori wrote: >>>> On 03/22/2010 12:55 PM, Avi Kivity wrote: >>>>>> Lets look at the ${HOME}/.qemu/qmp/ enumeration method suggested by >>>>>> Anthony. >>>>>> There's numerous ways that this can break: >>>>> I don't like it either. We have libvirt for enumerating guests. >>>> We're stuck in a rut with libvirt and I think a lot of the >>>> dissatisfaction with qemu is rooted in that. It's not libvirt that's >>>> the probably, but the relationship between qemu and libvirt. >>> +1 >>> The obvious reason why so many people still use shell scripts rather >>> than libvirt is because if it just doesn't provide what they need. >>> Every time I've looked at it (and I've been looking for a better >>> solution for many years), it seems that it would have provided most of >>> the things I needed, but the remaining bits were unsolvable. >> If you happen to remember what missing features prevented you choosing >> libvirt, that would be invaluable information for us, to see if there >> are quick wins that will help out. We got very useful feedback when >> recently asking people this same question > >> http://rwmj.wordpress.com/2010/01/07/quick-quiz-what-stops-you-from-using-l… > >> Allowing arbitrary passthrough of QEMU commands/args will solve some of >> these issues, but certainly far from solving all of them. eg guest cut+ >> paste, host side control of guest screen resolution, easier x509/TLS >> configuration for remote management, soft reboot, Windows desktop support >> for virt-manager, host network interface management/setup, etc > >> Regards, >> Daniel

5 13

[libvirt] [PATCH] nwfilter: Process DHCP option to determine whether packet is a DHCP_OFFER
by Stefan Berger 09 Apr '10

09 Apr '10

I mistakenly took the op field in the DHCP message as the DHCP_OFFER type. Rather than basing the decision to read the VM's IP address on that field, process the appended DHCP options where option 53 indicates the actual type of the packet. I am also reading the broadcast address of the VM, but don't use it so far. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com> --- src/nwfilter/nwfilter_learnipaddr.c | 61 +++++++++++++++++++++++++++++++++--- 1 file changed, 56 insertions(+), 5 deletions(-) Index: libvirt-acl/src/nwfilter/nwfilter_learnipaddr.c =================================================================== --- libvirt-acl.orig/src/nwfilter/nwfilter_learnipaddr.c +++ libvirt-acl/src/nwfilter/nwfilter_learnipaddr.c @@ -64,6 +64,13 @@ struct f_arphdr { } ATTRIBUTE_PACKED; +struct dhcp_option { + uint8_t code; + uint8_t len; + uint8_t value[0]; /* length varies */ +} ATTRIBUTE_PACKED; + + /* structure representing DHCP message */ struct dhcp { uint8_t op; @@ -78,9 +85,12 @@ struct dhcp { uint32_t siaddr; uint32_t giaddr; uint8_t chaddr[16]; - /* omitted */ + uint8_t zeroes[192]; + uint32_t magic; + struct dhcp_option options[0]; } ATTRIBUTE_PACKED; +#define DHCP_MSGT_DHCPOFFER 2 struct ether_vlan_header { @@ -212,6 +222,40 @@ virNWFilterGetIpAddrForIfname(const char #ifdef HAVE_LIBPCAP +static void +procDHCPOpts(struct dhcp *dhcp, int dhcp_opts_len, + uint32_t *vmaddr, uint32_t *bcastaddr, + enum howDetect *howDetected) { + struct dhcp_option *dhcpopt = &dhcp->options[0]; + + while (dhcp_opts_len >= 2) { + + switch (dhcpopt->code) { + + case 28: /* Broadcast address */ + if (dhcp_opts_len >= 6) { + uint32_t *tmp = (uint32_t *)&dhcpopt->value; + (*bcastaddr) = ntohl(*tmp); + } + break; + + case 53: /* Message type */ + if (dhcp_opts_len >= 3) { + uint8_t *val = (uint8_t *)&dhcpopt->value; + switch (*val) { + case DHCP_MSGT_DHCPOFFER: + *vmaddr = dhcp->yiaddr; + *howDetected = DETECT_DHCP; + break; + } + } + } + dhcp_opts_len -= (2 + dhcpopt->len); + dhcpopt = (struct dhcp_option*)((char *)dhcpopt + 2 + dhcpopt->len); + } +} + + /** * learnIPAddressThread * arg: pointer to virNWFilterIPAddrLearnReq structure @@ -236,12 +280,13 @@ learnIPAddressThread(void *arg) struct ether_header *ether_hdr; struct ether_vlan_header *vlan_hdr; virNWFilterIPAddrLearnReqPtr req = arg; - uint32_t vmaddr = 0; + uint32_t vmaddr = 0, bcastaddr = 0; unsigned int ethHdrSize; char *listen_if = (strlen(req->linkdev) != 0) ? req->linkdev : req->ifname; int to_ms = (strlen(req->linkdev) != 0) ? 1000 : 0; + int dhcp_opts_len; char macaddr[VIR_MAC_STRING_BUFLEN]; virBuffer buf = VIR_BUFFER_INITIALIZER; char *filter= NULL; @@ -382,12 +427,18 @@ learnIPAddressThread(void *arg) sizeof(struct dhcp)) { struct dhcp *dhcp = (struct dhcp *) ((char *)udphdr + sizeof(udphdr)); - if (dhcp->op == 2 /* DHCP OFFER */ && + if (dhcp->op == 2 /* BOOTREPLY */ && !memcmp(&dhcp->chaddr[0], req->macaddr, 6)) { - vmaddr = dhcp->yiaddr; - howDetected = DETECT_DHCP; + dhcp_opts_len = header.len - + (ethHdrSize + iphdr->ihl * 4 + + sizeof(struct udphdr) + + sizeof(struct dhcp)); + procDHCPOpts(dhcp, dhcp_opts_len, + &vmaddr, + &bcastaddr, + &howDetected); } } }

2 1