- Devel - libvirt List Archives

[Libvir] [patch] virsh setvcpus range check for negative value case
by Atsushi SAKAI 15 Aug '07

15 Aug '07

Hi, This patch adds virsh setvcpus range check for negative value case. for example to the inactive domain virsh setvcpus -1 sets vcpus=4294967295 And cannot boot the inactive domain. Signed-off-by: Atsushi SAKAI <sakaia(a)jp.fujitsu.com> Thanks Atsushi SAKAI

2 2

[Libvir] [PATCH] fix UUID lookup when using the proxy
by Daniel Veillard 14 Aug '07

14 Aug '07

Chasing a bug at Red Hat I noticed that the code never copied the UUID onto the packet sent to the proxy, which of course makes things harder. Finding why this still worked on 32bits hosts is left as an exercise for interested readers ! (of course I wrote that code on a 32 bits machine at the time) One line patch enclosed, Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard(a)redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/

2 1

[Libvir] fix SYNOPSIS of virsh --help schedinfo
by Atsushi SAKAI 14 Aug '07

14 Aug '07

fix SYNOPSIS of virsh --help schedinfo. from sched to schedinfo Thanks Atsushi SAKAI

2 1

[Libvir] several typo fixes on man virsh
by Atsushi SAKAI 14 Aug '07

14 Aug '07

Hi, several typo fixes on man virsh (uses ispell) interract -> interact suports -> supports privledges -> privileges synoposis -> synopsis documetnation -> documentation runable -> runnable Ouput -> Output guaruntee -> guarantee processus -> processes (2 typo) Detatch -> Detach dvices -> devices instanciated -> instantiated Thanks Atsushi SAKAI

2 1

[Libvir] [PATCH 0/7] QEMU/KVM save/restore support, take 3
by Jim Paris 14 Aug '07

14 Aug '07

Here's take 3 of the QEMU/KVM save/restore support. Thanks for your input. Changes since last time: - Remove escape sequence filtering, it's not necessary. - Clean up stdin handling in virExec, use -1 to signify unused - Add signal-safe read/write wrappers that handle EINTR and use them. - Add version and padding to image header, and check version on restore. - Include null-termination in XML data & length - Show name of conflicting domain in error message Everything seems to work well in my tests. I've run into a few rare cases where the migration doesn't work correctly (causing segfaults in the guest, or kvm to crash), but it's not libvirt's fault, and libvirt handles the failures well. (I suspect it's related to http://article.gmane.org/gmane.comp.emulators.kvm.devel/5583) Thanks, -jim

3 17

[Libvir] save/restore support for KVM
by Jim Paris 13 Aug '07

13 Aug '07

Hi, I've implemented save/restore support for KVM using the live migration feature. First, a few patches to KVM to fix bugs: Subject: [PATCH 1/3] qemu: fix freed pointer dereference http://article.gmane.org/gmane.comp.emulators.kvm.devel/5572 Subject: [PATCH 2/3] qemu: don't start a new migration if one is already in progress http://article.gmane.org/gmane.comp.emulators.kvm.devel/5575 Subject: [PATCH 3/3] qemu: reset buffer pointers after CR/LF http://article.gmane.org/gmane.comp.emulators.kvm.devel/5573 (If compatibility with old KVM is wanted, it might be possible to work around the kvm bugs in other ways, but I'm not sure). Then, another patch to KVM to support inbound migration from a filename. It already supports migration from stdin, but adding this seemed easier from a libvirt perspective. Subject: [PATCH] qemu: accept filename for incoming migration http://article.gmane.org/gmane.comp.emulators.kvm.devel/5590 Finally, the libvirt side. Some notes: - Arbitrary decisions about VM status: I pause the VM before starting migration, and destroy it once it's finished. Neither is required by KVM but I'd be concerned about the disk image state otherwise. Also, after resuming, the VM is still paused. I don't know how Xen behaves in this regard but the behavior here is easy enough to change. - I append the domain's UUID at the end of the migration image. This doesn't affect KVM at all (it ignores the extra data). Does that seem reasonable? It's unclear how the saved image is supposed to get associated with a particular VM configuration without doing something like this. - I added the migrateFrom field to the qemu_vm struct. Dunno if that's the best place. Could also add put it in qemu_vm_def, or add parameters to qemudStartVMDaemon, etc.. Patch against current CVS is below. Signed-off-by: Jim Paris <jim(a)jtan.com> -jim Index: src/qemu_conf.c =================================================================== RCS file: /data/cvs/libvirt/src/qemu_conf.c,v retrieving revision 1.10 diff -u -r1.10 qemu_conf.c --- src/qemu_conf.c 7 Aug 2007 13:02:35 -0000 1.10 +++ src/qemu_conf.c 9 Aug 2007 21:15:10 -0000 @@ -1518,7 +1518,8 @@ (vm->def->os.initrd[0] ? 2 : 0) + /* initrd */ (vm->def->os.cmdline[0] ? 2 : 0) + /* cmdline */ (vm->def->graphicsType == QEMUD_GRAPHICS_VNC ? 2 : - (vm->def->graphicsType == QEMUD_GRAPHICS_SDL ? 0 : 1)); /* graphics */ + (vm->def->graphicsType == QEMUD_GRAPHICS_SDL ? 0 : 1)) + /* graphics */ + (vm->migrateFrom[0] ? 3 : 0); /* migrateFrom */ snprintf(memory, sizeof(memory), "%d", vm->def->memory/1024); snprintf(vcpus, sizeof(vcpus), "%d", vm->def->vcpus); @@ -1767,6 +1768,15 @@ /* SDL is the default. no args needed */ } + if (vm->migrateFrom[0]) { + if (!((*argv)[++n] = strdup("-S"))) + goto no_memory; + if (!((*argv)[++n] = strdup("-incoming"))) + goto no_memory; + if (!((*argv)[++n] = strdup(vm->migrateFrom))) + goto no_memory; + } + (*argv)[++n] = NULL; return 0; Index: src/qemu_conf.h =================================================================== RCS file: /data/cvs/libvirt/src/qemu_conf.h,v retrieving revision 1.6 diff -u -r1.6 qemu_conf.h --- src/qemu_conf.h 30 Jul 2007 09:59:06 -0000 1.6 +++ src/qemu_conf.h 9 Aug 2007 21:15:10 -0000 @@ -213,6 +213,7 @@ char configFile[PATH_MAX]; char autostartLink[PATH_MAX]; + char migrateFrom[PATH_MAX]; struct qemud_vm_def *def; /* The current definition */ struct qemud_vm_def *newDef; /* New definition to activate at shutdown */ Index: src/qemu_driver.c =================================================================== RCS file: /data/cvs/libvirt/src/qemu_driver.c,v retrieving revision 1.14 diff -u -r1.14 qemu_driver.c --- src/qemu_driver.c 30 Jul 2007 09:59:06 -0000 1.14 +++ src/qemu_driver.c 9 Aug 2007 21:15:10 -0000 @@ -656,7 +656,7 @@ if (virExecNonBlock(conn, argv, &vm->pid, &vm->stdout, &vm->stderr) == 0) { vm->id = driver->nextvmid++; - vm->state = VIR_DOMAIN_RUNNING; + vm->state = vm->migrateFrom[0] ? VIR_DOMAIN_PAUSED : VIR_DOMAIN_RUNNING; driver->ninactivevms--; driver->nactivevms++; @@ -1852,29 +1852,151 @@ return 0; } +#define QEMUD_SAVE_MAGIC "LibvirtQemudUUID" -static int qemudDomainSave(virDomainPtr dom, - const char *path ATTRIBUTE_UNUSED) { +static int qemudDomainSave(virDomainPtr dom, + const char *path) { struct qemud_driver *driver = (struct qemud_driver *)dom->conn->privateData; struct qemud_vm *vm = qemudFindVMByID(driver, dom->id); + char *command, *info; + int fd; + if (!vm) { - qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_DOMAIN, "no domain with matching id %d", dom->id); + qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_DOMAIN, + "no domain with matching id %d", dom->id); return -1; } + if (!qemudIsActiveVM(vm)) { - qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, "domain is not running"); + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + "domain is not running"); return -1; } - qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, "save is not supported"); - return -1; + + if (strchr(path, '\'') || strchr(path, '\\') ) { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + "invalid filename"); + return -1; + } + + /* Pause */ + if (qemudDomainSuspend(dom) != 0) { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + "failed to pause domain"); + return -1; + } + + /* Migrate to file. */ + if (asprintf (&command, "migrate \"exec:dd of='%s' 2>/dev/null\"\n", + path) == -1) { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + "out of memory"); + return -1; + } + + if (qemudMonitorCommand(driver, vm, command, &info) < 0) { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + "migrate operation failed"); + free(command); + return -1; + } + + free(info); + free(command); + + /* Append the UUID, which we'll need in order to restore the domain. */ + if ((fd = open(path, O_APPEND|O_WRONLY)) < 0) { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + "failed to open '%s'", path); + return -1; + } + + if (write(fd, QEMUD_SAVE_MAGIC, sizeof(QEMUD_SAVE_MAGIC)) != + sizeof(QEMUD_SAVE_MAGIC)) { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + "failed to write magic"); + close(fd); + return -1; + } + + if (write(fd, dom->uuid, VIR_UUID_BUFLEN) != VIR_UUID_BUFLEN) { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + "failed to write uuid"); + close(fd); + return -1; + } + close(fd); + + /* Shut it down */ + qemudShutdownVMDaemon(dom->conn, driver, vm); + if (!vm->configFile[0]) + qemudRemoveInactiveVM(driver, vm); + + return 0; } static int qemudDomainRestore(virConnectPtr conn, - const char *path ATTRIBUTE_UNUSED) { - /*struct qemud_driver *driver = (struct qemud_driver *)conn->privateData;*/ - qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, "restore is not supported"); - return -1; + const char *path) { + struct qemud_driver *driver = (struct qemud_driver *)conn->privateData; + unsigned char uuid[VIR_UUID_BUFLEN]; + unsigned char magic[sizeof(QEMUD_SAVE_MAGIC)]; + virDomainPtr dom; + struct qemud_vm *vm; + int fd; + + /* Find the UUID */ + if ((fd = open(path, O_RDONLY)) < 0) { + qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, + "cannot read domain image"); + return -1; + } + if (lseek(fd, 0 - (sizeof(magic) + sizeof(uuid)), SEEK_END) == -1) { + qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, + "failed to seek to end of domain image"); + close(fd); + return -1; + } + if (read(fd, magic, sizeof(magic)) != sizeof(magic)) { + qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, + "failed to read domain image"); + close(fd); + return -1; + } + if (memcmp(magic, QEMUD_SAVE_MAGIC, sizeof(magic))) { + qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, + "image magic is not correct"); + close(fd); + return -1; + } + if (read(fd, uuid, sizeof(uuid)) != sizeof(uuid)) { + qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, + "failed to read domain UUID from image"); + close(fd); + return -1; + } + close(fd); + + /* Find the domain & vm */ + dom = qemudDomainLookupByUUID(conn, uuid); + if (!dom) { + qemudReportError(conn, NULL, NULL, VIR_ERR_INVALID_DOMAIN, + "no domain with matching uuid"); + return -1; + } + + vm = qemudFindVMByUUID(driver, dom->uuid); + if (!vm) { + qemudReportError(conn, dom, NULL, VIR_ERR_INVALID_DOMAIN, + "no vm with matching uuid"); + return -1; + } + + /* Set the migration source and start it up. */ + snprintf(vm->migrateFrom, sizeof(vm->migrateFrom), "file://%s", path); + vm->migrateFrom[sizeof(vm->migrateFrom)-1] = '\0'; + + return qemudStartVMDaemon(dom->conn, driver, vm); } @@ -1931,6 +2053,7 @@ return -1; } + vm->migrateFrom[0] = '\0'; return qemudStartVMDaemon(dom->conn, driver, vm); }

4 30

[Libvir] [PATCH] Block device and network stats
by Richard W.M. Jones 13 Aug '07

13 Aug '07

This is the implementation (currently Xen, local only). It includes a couple of virsh commands if you want to play with looking at the stats. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

3 6

[Libvir] PATCH: Fix paravirt FB with xm driver
by Daniel P. Berrange 10 Aug '07

10 Aug '07

The code for dealing with the paravirt framebuffer was missing code to deal with an explicit vncdisplay=N setting - only working for vncunused=1. It was also not flexible enough in processing the 2 different styles of FB config. 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 5

[Libvir] PATCH: Fix virtual neworks with xm driver
by Daniel P. Berrange 10 Aug '07

10 Aug '07

The virtual networks stuff was never added to the xm driver for Xen 3.0.3 or earlier. This means that if adding a virtual network the bridge device won't get defined in the config. This patch addreses this, and also fixes a tiny mem leak in the equivalent code in the xend driver. 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 4

[Libvir] PATCH: Improve performance on Xen
by Daniel P. Berrange 10 Aug '07

10 Aug '07

The Xen implementations of virDomainLookupByID virDomainLookupByUUID virDomainLookupByName virDomainGetOSType all have sub-optimal performance since they speak to XenD. The lookupXXX functions all basically require 3 pieces of info in the end (name,id,uuid). They each get given one piece & have to lookup the other piece. Every running domain has to have an entry in XenStore on /local/domain/N where 'N' is the id. Under this location there is always a 'name' field. So if we have the id we can efficiently get the name. I just realized that the getdomaininfo hypercall struct contains the uuid and id. So for the ByID and ByUUID calls we can get all the info we need with a hypercall and a read of xenstore. This is very efficient compared to hitting XenD. As of Xen 3.1.0 hypervisor the flags in the getdomaininfo hypercall struct also mark whether the guest is HVM vs paraivrt, so we can also implement the GetOSType api with a single hypercall. That just leaves the ByName impl. The only way I can think of doing this is to scan /local/domain/N in xenstore until we find it. I'm going to try this, but I'm not convinced it'll be any significantly than talking to XenD. I may be surprised though. Any for the 3 improvements I have done, the 'virsh dominfo' call has improved when using either ID or UUID: Create a set of UUIDs & IDs # for i in `seq 1 500` ; do echo dominfo 73 ; done > ids.txt # for i in `seq 1 500` ; do echo dominfo 8f07fe28-753f-2729-d76d-bdbd892f949a ; done > uuids.txt Now before measurements: # time ./virsh < ids.txt > /dev/null real 0m17.317s user 0m0.144s sys 0m0.164s # time ./virsh < uuids.txt > /dev/null real 0m14.432s user 0m0.236s sys 0m0.460s Against the after measurements: # time ./virsh < demo.txt > /dev/null real 0m0.259s user 0m0.100s sys 0m0.072s # time ./virsh < demo.txt > /dev/null real 0m0.482s user 0m0.324s sys 0m0.108s A pretty good speedup really considering how frequently these calls are made if you're monitoring domains frequently. 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 5

[Libvir] Proposal: Block device and network stats
by Richard W.M. Jones 10 Aug '07

10 Aug '07

With the very timely question that has been raised about block and network device stats, I'm posting my proposed interface. I almost have this working for the Xen case on my test machine. The idea of this interface is to be optimised for the following case: * stats need to be fetched frequently (eg. once per second) * the layout of devices doesn't change often (ie. adding or removing devices from domains is very infrequent) * most domains have only a single block and single network device With the above assumptions, the idea is that you would use the API like this: (1) When the program starts up or (infrequently) is notified of a change or when a new domain appears, the program calls virDomainGetXMLDesc and parses out the //domain/devices/interface and //domain/devices/disk fields to get the list of network interfaces and block devices. So for each domain, you'll have two lists like this: dom[1]["blockdevs"] = ["xvda"] dom[1]["interfaces"] = ["virbr0"] (2) Frequently (eg. once per second) the program calls (in the above case): virDomainBlockStats (dom1, "xvda", &bstats, sizeof bstats); virDomainInterfaceStats (dom1, "virbr0", &istats, sizeof istats); (3) Since stats are cumulative the program must do its own subtraction and calculation in order to display throughput per second. The implementation goes directly from the name "xvda" to the backend device (/sys/devices/xen-backend/- [type]-[domid]-[major:minor]/statistics) for block devices, and slightly more complicatedly to a particular line in /proc/net/dev for network interfaces. (Note in the Xen case, the name "virbr0" is little more than a placeholder to mean "the zeroth interface for domain d"). In particular the current implementation doesn't cache anything. This should all work fine in the Linux / Xen case. libxenstat gives us sample code that we can copy for the Solaris / Xen case, but it would need testing from someone with access to such a machine. I don't think qemu supports stats at all. Initially we won't support stats for tap devices because there needs to be an upstream patch to support this (http://lists.xensource.com/archives/html/xen-changelog/2007-02/msg00278.html) The extra size parameter will allow us to extend the stats structures in future, maintaining binary backwards compatibility with existing clients. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

3 7

[Libvir] Xen physical host statistics
by Mindaugas Kiznis 10 Aug '07

10 Aug '07

Hello, - I am trying to make an application, which would connect to a remote xen host and fetch some information about total (physical) CPU usage, total memory usage, network and disk I/O. I have managed to do a remote statistics grabber for CPU and memory usage using libvirt. I couldn't find any source of information about libvirt ability to return I/O numbers. Also haven't found any other solution yet. Any suggestions? Thank you in advance! Best regards, Mindaugas Kiznis

3 4

[Libvir] PATCH: UUID code cleanup
by Daniel P. Berrange 09 Aug '07

09 Aug '07

Browsing the source I noticed we have many differents constants for UUID length, many ways of turning a UUID into a string, 2 ways of turning a string into a UUID (one with some scary signed <-> unsigned casting). This is just a result of the way the code evolved & I figured it could do with cleaning up a little. So I added a virUUIDFormat(const unsigned char uuid, char *uuidstr) method to the uuid.c, removed all the other equivalent code throughout and kiled all the duplicate constants. So now everything UUID related is in the one place uuid.c, while the constants are in libvirt/libvirt.h Its a surprisingly big patch, but it should not have any functional change with one exception. The XenD driver used to generate XML <uuid>...</uuid> tags without any '-' embedded in them. All other places use the '-' and technically the RFC requires them. Our parser is flexible enough to cope with, or without '-'. diffstat ~/libvirt-uuid-cleanup.patch proxy/Makefile.am | 2 - qemud/internal.h | 2 - src/libvirt.c | 34 +++------------------------- src/openvz_conf.c | 14 +++++------ src/openvz_conf.h | 3 -- src/qemu_conf.c | 20 ++++++---------- src/qemu_conf.h | 5 +--- src/test.c | 18 +++++---------- src/uuid.c | 57 +++++++++++++++++++++++++++++++++-------------- src/uuid.h | 13 ++++------ src/xend_internal.c | 62 +++++++++++++++++++--------------------------------- src/xm_internal.c | 31 +++++--------------------- src/xml.c | 61 --------------------------------------------------- src/xml.h | 1 14 files changed, 104 insertions(+), 219 deletions(-) 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 3

[Libvir] ANNOUNCE: Virt-top - a top-like utility for displaying virtualization stats
by Richard W.M. Jones 09 Aug '07

09 Aug '07

I'm pleased to announce the first release of virt-top, which is a top-like utility for displaying virtualization stats. It aims to look and feel very much like regular 'top', so as to be as familiar as possible for systems administrators. You can also use it as a pleasant replacement for xentop. It uses libvirt, so can display stats across a variety of different hypervisors and virtualization systems (not just Xen, although that is where the testing has gone so far). http://et.redhat.com/~rjones/virt-top/ The license is a combination of LGPL (for the library) and GPL (for the virt-top program). Current status -------------- You can view domains and use familiar keys like 'P'/'M'/... to sort by processor/memory/..., and 'd'/'s' to set the delay between updates. Also some common top command-line options are implemented. The man page is here: http://et.redhat.com/~rjones/virt-top/virt-top.txt There are a variety of source and binary RPMs available for Fedora users. I don't yet have a working Debian/Ubuntu package, but will have a go at making one tomorrow. The next thing I'll be working on is showing virtual and physical CPU usage of guests. After that I'm hoping to discuss extensions to libvirt to make other interesting statistics available to virt-top, in particular disk and network I/O stats. There are a few data collection artifacts which need to be investigated. In particular, %CPU sometimes goes over 100%. Obviously accurate data collection is an important goal for this tool. Memory usage is good: typical 'RES' (in regular top) for virt-top is just under 2.5 MB, and I've had it running for hours at a time without memory usage increasing, which seems to indicate that there aren't any major memory leaks. Development ----------- The program is currently very small: exactly 500 lines of code! If you want to dive in and send me patches they are most welcome, but remember that I'm trying to make this utility act as much like 'top' as possible, so if 'top' does it in a particular way, then I'd prefer virt-top to do the same thing. (May not apply to obscure top functionality, if top does something silly). If you're a systems administrator, not a programmer, and you'd like virt-top to have some particular feature, then please let me know, and I will be happy to review it and code it for you. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

4 10

[Libvir] libvirt daemon UNIX socket auth with PolicyKit
by Daniel P. Berrange 08 Aug '07

08 Aug '07

Currently our authentication model for local connections is using the basic UNIX file permissions, possibly with a setuid helper (in Xen case only). It can be summarized as - If app using libvirt is running as root => full access - Else => read only access The latter is enforced by fact that in Xen case libvirt_proxy only has impl for a handful of read only APIs, or in non-Xen case that the UNIX domain socket for the daemon /var/run/libvirt/libvirt-sock is mode 0700, while /var/run/libvirt/libvirt-sock-ro is 0777 & the daemon enforces based on which socket the client connects to. This is good because it allows non-root to at least monitor guest state while requiring root authentication for actually changing state. This is bad because it requires any app which wants to change state to run as root. ie we are required to launch virt-manager as root to gain ability to manage local guests. Problem with this include: - running the entire PyGTK & GTK & X codebase as root is undesirable - no integration with the DBus desktop session (gnome-vfs integartion) - no integration with the GNOME keyring (for VNC server passwords) - redundant (&dangerous) if all you want to do is manage remote libvirt hosts In summary what I really need for virt-manager is - Always run as non-root - Authenticate for local guest management (ie read+write) UNIX domain sockets already provide a way for each end to identify the PID and UID of the other end. This enables the libvirt daemon to determine the identity of the application on the other end. With this information the daemon merely needs to check this identity against some access control policy rules. Where to get/define these rules though ? Enter PolicyKit. http://lists.freedesktop.org/archives/hal/2006-March/004770.html http://lists.freedesktop.org/archives/hal/2007-June/008815.html PolicyKit is a freedesktop.org project to replace all the distro specific 'do admin stuff as root' hacks like consolehelper[1], sudo, gksudo with a standard mechanism. Most importantly this mechanism does not require that the application in question run as root, or ever gain root privileges. In addition this is completely configurable by the administrator in a application independant manner. How would the daemon use PolicyKit for authenticating clients on UNIX domain sockets - libvirtd defines two actions it can check called 'libvirt-local-monitor (read only monitoring of state), and 'libvirt-local-manage' (full read-write management). - Both libvirt-sock and libvirt-sock-ro are mode 0777 (ie all users) - A connection arrives on the UNIX domain socket either libvirt-sock-ro or libvirt-sock - libvirtd use SO_PEERCRED to get the PID of the client - If the connection is on libvirt-sock-ro 'action' is 'libvirt-local-monitor' Else if connection is on libvirt-sock 'action' is 'libvirt-local-manage' - Libvirt asks policy kit if PID is allowed to perform the 'action' -> Yes, connection proceeds -> No, connection is dropped NB, having two separate UNIX domain sockets is strictly speaking redundant now - the daemon could simply check each 'action' in turn - but keeping the separate sockets simplifies the code, because it minimises the diffs when PolicyKit is not enabled by the 'configure' script. How is policy determined by PolicyKit ? - libvirt RPM provides /usr/share/PolicyKit/policy/libvirt.policy - This file defines the two actions it supports 'libvirt-local-monitor' and 'libvirt-local-manage'. - For each action we define a default policy for an active desktop session, and an inactive desktop session. The sample policy is to allow 'libvirt-local-monitor' for any local desktop session, but only allow 'libvirt-local-manage' for the active session with the additional requirement that the user authenticate with PolicyKit using the root password. This basically replicates the existing security semantics, without requiring that we run virt-manager itself as root. This accomplishes the core goal. There are other benefits to PolicyKit though - the administrator can override the application provided default Policy. For example, could add a rule that says "if the user fred is logged into the local X session, allow them to do 'libvirt-local-manage' without any further prompts for root password" Since PolicyKit is a general framework, that rule could be generalized beyond just virt-manager "if the user fred is logged into the local X session, allow them to do any defined action without any further prompts for root password" Or going the other way the administrator can lock things down requiring a root password even for read only monitoring of VMs. Or a halfway house of requiring the user's own password (sudo equiv, instead of su equiva). If course this doesn't just apply to virt-manager either. This transparently 'just works' for virsh too & any other apps using libvirt - eg if policy grants 'libvirt-local-manage' then you can do full management whether virsh is root or not. This all talks about enforcement of policy from the daemon side. There is one outstanding question - if the policy so requires, where/how do we ask the user to enter their credentials. PolicyKit provides two mechanisms for this: - A command line tool polkit-grant which uses pam to authenticate you as root, or using your own user password. If successful you will have been granted the permission to perform that action for the remainder of the session. - A DBus service for triggering display of some UI for authenticating. There is a GNOME impl of this service & likely a KDE impl too. eg with virsh start off without any priveleges $ virsh --connect qemu:///system net-list libvir: Remote error : Connection reset by peer error: failed to connect to the hypervisor error: no valid connection $ polkit-grant --action libvirtd-local-manage2 Authentication is required. Password: Keep this privilege for the session? [no/session]? $ ./src/virsh --connect qemu:///system net-list --all Name State Autostart ----------------------------------------- default active yes One option is to have the libvirt library spawn polkit-grant if not finding any $DISPLAY, and call the DBus service (for UI prompt) if finding any. The other option is to punt this code upto the application - eg virt-manager always call the DBus service, while virsh always call polkit-grant. Finally, PolicyKit is new in Fedora 8 if you wish to try this out. The configure script is setup so that it automagically detects presence of PolicyKit and only turns it on at compile time if its available. The fallback is to continue current behaviour of restricting based on the socket permissions. http://fedoraproject.org/wiki/Releases/FeaturePolicyKit Finally, finally, if we enable PolicyKit support, then we should disable the Xen specific libvirt_proxy setuid binary, since it would subvert the policy for libvirt-local-monitor. IMHO we should just kill the setuid binary altogether, since we already require that the libvirt daemon be running for the networking APIs & this provides read-only access already which works for non-Xen drivers too. Attaching the patch implementing the proof of concept $ diffstat ~/libvirt-policykit.patch configure.in | 60 ++++++++++++++++++-- qemud/Makefile.am | 12 +++- qemud/internal.h | 23 ++++++- qemud/libvirtd.policy | 41 ++++++++++++++ qemud/qemud.c | 145 ++++++++++++++++++++++++++++++++++++++++++++------ qemud/remote.c | 6 +- 6 files changed, 258 insertions(+), 29 deletions(-) Oh, finally, finally, finally. PolicyKit is pretty new so I'd understand if there's some resistance to including it in an official libvirt release at this point. I think the patch is small enough that we could carry it as add-on in the Fedora RPMs only to start with untill its proved to be maintainable long term - I really need this for Fedora 8 to make virt-manager work sanely - ie not need to run as root. 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 -=|

4 9

[Libvir] [PATCH] A variety of fixes for Debian, old versions of GnuTLS
by Richard W.M. Jones 07 Aug '07

07 Aug '07

Attached is a patch to get libvirt working on Debian/stable again. (1) Add ./configure option --disable-stack-protector. The stack protector feature is still enabled by default, but it causes no end of problems when compiling on Debian systems, apparently because of bugs in their gcc. So Debian users would be well advised to try disabling it if they have problems building. (2) Adds support for GnuTLS 1.0.x by detecting this version and enabling some simple compatibility macros. (3) In qemudDispatchClientEvent, a prototype shadowed a global function name. A simple name change to a parameter fixes it. (4) Debian (and likely non-Linux systems also) don't have the SIOCBRADDBR and related ioctls. If these are missing, then I've changed this to runtime errors. This means that these systems won't be able to start the network daemon, but other (non-network) features of libvirt continue to work. (5) Removed a warning in src/qemu_conf.c. Rich. PS. You need to drop the file gnutls_1_0_compat.h into src/ subdirectory. Sorry this file isn't in the proper patch but I don't have write access to CVS from the testing machine. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

3 3

[Libvir] [PATCH] Fix libvirtd --config / -f option
by Richard W.M. Jones 07 Aug '07

07 Aug '07

As well as fixing the --config/-f option for libvirtd, this also changes an abort() into a meaningful error message. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

3 4

[Libvir] ANNOUNCE: OCaml bindings 0.3.1.0 released
by Richard W.M. Jones 02 Aug '07

02 Aug '07

http://et.redhat.com/~rjones/ocaml-libvirt/ This removes use of the dangerous vir{Domain,Network}GetConnect functions and makes the relationship between domains, networks and connections explicit to the garbage collector. I've also added support for domain migration. mlvirtmanager displays the hostname when connected to multiple remote hosts (see screenshot attached). Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

1 0

[Libvir] [PATCH] check file format in virsh attach/detach-device
by Masayuki Sunou 02 Aug '07

02 Aug '07

Hi virsh attach/detach-devich does not check file format now. This patch checks config file format is XML or not. (in virsh attach/detach-device) If file format is not XML, just return the error with message. Signed-off-by: Masayuki Sunou <fj1826dm(a)aa.jp.fujitsu.com> Atsushi SAKAI <sakaia(a)jp.fujitsu.com> Thanks, Masayuki Sunou.

3 4

[Libvir] Proposal: Check availability of driver calls (repost)
by Richard W.M. Jones 01 Aug '07

01 Aug '07

-- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

3 9

[Libvir] Exporting calls just to libvirtd
by Richard W.M. Jones 31 Jul '07

31 Jul '07

With the two patches[1][2] I recently posted, the remote driver has started to make calls more directly into the drivers. Calls like driver->domainMigratePrepare in [2] are exposed to the remote driver, but shouldn't be exported for public use. However because libvirtd links to regular libvirt.so, the calls actually do need to be exported, so we export them as __virDomainMigratePrepare (etc) with the double underscore and exclusion from <libvirt.h> indicating that users of the published libvirt API should not call them. It seems we could avoid exporting them entirely with one of these methods: (a) Link libvirtd to the static libvirt.a. -- Probably violates a load of distro policy and removes some supposed benefits of dynamic linking. (b) Build a separate shared library with a different dynamic symbol table. -- Also removes one benefit of dynamic linking, because the two libraries are different despite having the same code in them. Any ideas if some advanced feature of ELF offers a way out here? Rich. [1]https://www.redhat.com/archives/libvir-list/2007-July/msg00444.html [2]https://www.redhat.com/archives/libvir-list/2007-July/msg00357.html -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

1 0

[Libvir] PATCH: Check for bogus default architecture
by Daniel P. Berrange 31 Jul '07

31 Jul '07

There was a missing check in the QEMU driver for a bogus architecture in the XML which could cause a crash. This wasn't a problem except for the fact that another bug in my recent virt-install 0.200.0 release passed a bogus architecture on 32-bit platforms :-( 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] PATCH: Overhaul the test driver
by Daniel P. Berrange 31 Jul '07

31 Jul '07

The test driver has suffered from quite alot of bitrot. Its got all sorts of checks in the driver entry points already done by libvirt.c, its got private data, redundantly pointing to private data, it doesn't implement any of the networking APIs, is missing alot of the newer domain APIs. So I've given it a thorough overhaul to remove alot of redundant code, clean up more, and add entry points for all the APIs except for VCPU pinning ones. The addition of the networking APIs means we no longer get errors thrown when trying to run tests without the libvirt daemon running, and means apps can actually do tests for their networking capabilities. docs/testnetdef.xml | 12 docs/testnetpriv.xml | 11 docs/testnode.xml | 2 src/driver.h | 1 src/libvirt.c | 17 src/qemu_driver.c | 1 src/remote_internal.c | 1 src/test.c | 1892 +++++++++++++++++++++++++++++++------------------- tests/virshtest.c | 465 +++--------- 9 files changed, 1382 insertions(+), 1020 deletions(-) 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] PATCH Fix a couple of mem leaks
by Daniel P. Berrange 31 Jul '07

31 Jul '07

I ran 'make valgrind' for the first time in too long and found a hanful of memory leaks in code exercised by the test suite. The attached patch fixes them all. I'm sure there are more lurking though - the test suite is mainly focused on exercising & validating correctness of parsing routines - its not really tried doing any validation of failure scenaarios. So I fear there's probably a number of error code paths where we don't cleanup properly, and/or don't reject invalid data. Its an area which should be addressed by someone with some copious free time ;-) 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 4

[Libvir] PATCH 0/2: Capabilities and features
by David Lutterkort 30 Jul '07

30 Jul '07

Hi, these two patches extend the features libvirt reports in the capabilities XML; in particular, features are now reported for the QEMU driver, and features for both Xen and QEMU include information on ACPI/APIC support. Tools that rely on feature reporting, in general, need to know whether a certain feature is on or off by default, and whether that can be toggled, because users will be asking for one of three things for their VM: * Feature X (e.g., ACPI) must be on/present in the VM * Feature X must not be on/present in the VM * It doesn't matter whether feature X is there or not For example, in QEMU, ACPI is on by default, but can be turned off. The APIC on the other hand, is always there and can't be turned off. That is now reported as <features> <acpi default="on" toggle="yes"/> <apic default="on" toggle="no"/> </features> It's unfortunate that the format for PAE (<pae/> vs. <nonpae/>) reporting doesn't fit into this scheme. I didn't want to start with having tags <X/> and <nonX/> for each feature; that means that parsers need to special case PAE. It's probably too late, but it would be nice if PAE was reported in a similar format, with <pae> correponding to <pae default="on" toggle="no"/>, <nonpae/> corresponding to <pae default="yes" toggle="no"/>, and having both together corresponding to <pae default="yes" toggle="yes"/>, but that's a separate discussion from these patches. David

4 12

[Libvir] [PATCH] virsh Range check in Credit Scheduler
by Atsushi SAKAI 30 Jul '07

30 Jul '07

Hi, Add Range check in Credit Scheduler. Signed-off-by: Atsushi SAKAI <sakaia(a)jp.fujitsu.com> Thanks Atsushi SAKAI

6 20

[Libvir] [PATCH] add a ttyconsole command
by Daniel Veillard 26 Jul '07

26 Jul '07

It was pointed out that we have a vncdisplay command but none for extracting the TTY console device: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=239687 the trivial associated patch adds it as well as the doc in the man page Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard(a)redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/

3 4

[Libvir] PATCH: Separate QEMU impl of nodeinfo API
by Daniel P. Berrange 25 Jul '07

25 Jul '07

While fixing the QEMU nodeinfo API to correctly deal with case where CPU sockets have sparse numbering (eg sockets 0 & 3 are populated), I realized that OpenVZ doesn't have a nodeinfo API, and its requirements are basically identical to the QEMU driver's. So this patch moves the impl of the nodeinfo API into a nodeinfo.c file, and makes both the QEMU and OpenVZ driver call out to this shared impl. I also put #ifdef __linux__ around the impl since code reading /proc/cpuinfo is never going to work on any non-Linux platform. For non linux I just return -1 which'll get treated as not-implemented. If QEMU driver is ported to work on Solaris, the nodeinfo.c file can be easily extended for their custom impl. Finally I'm adding a testcase with a bunch of example /proc/cpuinfo files to validate correctness 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 4

[Libvir] [PATCH] Python bindings: Use virDomain/NetworkGetConnect instead of storing connection in domain/network object
by Richard W.M. Jones 24 Jul '07

24 Jul '07

Previously[1] I patched libvirt to add the virDomainGetConnect and virNetworkGetConnect functions. The rationale for these is explained here[2]. The attached patch makes the Python bindings use these functions instead of storing the connection object explicitly. Rich. [1]https://www.redhat.com/archives/libvir-list/2007-June/msg00365.html [2]https://www.redhat.com/archives/libvir-list/2007-June/msg00352.html

3 8

[Libvir] Release of libvirt-0.3.1
by Daniel Veillard 24 Jul '07

24 Jul '07

Quite a few things were fixed since 0.3.0, and it was looking like a new release at this point would be a good idea. Availbale at ftp://libvirt.org/libvirt/ * Documentation: - index to remote page - script to test certificates - IPv6 remote support docs (Daniel Berrange) - document VIRSH_DEFAULT_CONNECT_URI in virsh man page (David Lutterkort) - Relax-NG early grammar for the network XML (David Lutterkort) * Bug fixes: - leaks in disk XML parsing (Masayuki Sunou) - hypervisor alignment call problems on PPC64 (Christian Ehrhardt) - dead client registration in daemon event loop (Daniel Berrange) - double free in error handling (Daniel Berrange) - close on exec for log file descriptors in the daemon (Daniel Berrange) - avoid caching problem in remote daemon (Daniel Berrange) - avoid crash after QEmu domain failure (Daniel Berrange) * Improvements: - checks of x509 certificates and keys (Daniel Berrange) - error reports in the daemon (Daniel Berrange) - checking of Ethernet MAC addresses in XML configs (Masayuki Sunou) - support for a new clock switch between UTC and localtime (Daniel Berrange) - early version of OpenVZ support (Shuveb Hussain) - support for input devices on PS/2 and USB buses (Daniel Berrange) - more tests especially the QEmu support (Daniel Berrange) - range check in credit scheduler (with Saori Fukuta and Atsushi Sakai) - add support for listen VNC parameter un QEmu and fix command line arg (Daniel Berrange) * Cleanups: - debug tracing (Richard Jones) - removal of --with-qemud-pid-file (Richard Jones) - remove unused virDeviceMode - new util module for code shared between drivers (Shuveb Hussain) - xen header location detection (Richard Jones) thanks to everybody who helped for this release, with bug reports or patches ! Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard(a)redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/

1 0

[Libvir] PATCH: Make QEMU driver honour 'listen' flag for VNC
by Daniel P. Berrange 24 Jul '07

24 Jul '07

In QEMU 0.9.0 or later it is possible to tell QEMU to only listen on a particular IP address. THis patch adapts the code so that it honours the 'listen' attribute on the <graphics> tag if using QEMU >= 0.9.0. It also re-enables the tests for this capability that I temporarily disabled. src/qemu_conf.c | 37 +++++++++++++----- src/qemu_conf.h | 1 tests/qemuxml2argvdata/qemuxml2argv-graphics-sdl.args | 2 tests/qemuxml2argvtest.c | 4 - tests/qemuxml2xmltest.c | 2 5 files changed, 32 insertions(+), 14 deletions(-) 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 5

[Libvir] PATCH: Fix crash in cleanup when VM creation fails
by Daniel P. Berrange 24 Jul '07

24 Jul '07

If using the 'virDomainCreateLinux' call to create a VM, a so called 'transient' domain will be created - ie one without any config file. There is special code in the qemudShutdownVMDaemon method to cleanup the resources associated with such domains, in particuarly free'ing the struct qemud_vm object. Unfortunately in the virDomainCreateLinux codepath this is a problem, because we still need the 'struct qemud_vm' object in certain edge cases, and so the caller has to free it. We currently have a double free() in that scenario. This patch removes the call to qemudFreeVMDaemon from qemudShutdownVMDaemon. Instead it is now always the caller's responsibility to cleanup after transient domains. qemu_driver.c | 39 ++++++++++++++++++++------------------- 1 file changed, 20 insertions(+), 19 deletions(-) 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] Remote daemon & virDomainFree interaction
by Daniel P. Berrange 24 Jul '07

24 Jul '07

In looking at a problem with domain object cleanup in virt-manager I came across a problem in the remote driver, well the internal driver API itself actually. Specifically the implmenetation of virDomainFree() never calls into the driver API - it simply uses virFreeDomain() release the memory associated with the virDomainPtr object. Couple this with the remote driver though, and virDomainPtr objects in the remote daemon never get released, because the virDomainFree call is never propagated over the wire to the server. Its quite easy to see this in practice. Simply add a printf to the impl of virDomainLookupByName which prints out the ref count. Then run either virsh or virt-manager for a while Get info QEMUGuest1 69 c7a5fdbd-edaf-9455-926a-d65c16db1809 Get info QEMUGuest1 70 c7a5fdbd-edaf-9455-926a-d65c16db1809 Get info QEMUGuest1 71 c7a5fdbd-edaf-9455-926a-d65c16db1809 We need to make virDomainFree call into the driver API, and also make sure that the remote driver implements it. 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 4

[Libvir] [PATCH] virDomainMigrate version 4 (for discussion only!)
by Richard W.M. Jones 24 Jul '07

24 Jul '07

This is version 4 of the virDomainMigrate patch. It includes remote support, but no qemu, virsh or python yet. And it needs lots more testing which I intend to do more of tomorrow. Interface --------- The interface is now this: virDomainPtr virDomainMigrate (virDomainPtr domain, virConnectPtr dconn, unsigned long flags, const char *dname, const char *uri, unsigned long resource); The caller may set dname, uri and resource to 0/NULL and forget about them. Or else the caller may set, in particular, uri to allow for more complicated migration strategies (especially for qemu). https://www.redhat.com/archives/libvir-list/2007-July/msg00249.html Driver support -------------- As outlined in the diagram in this email: https://www.redhat.com/archives/libvir-list/2007-July/msg00264.html migration happens in two stages. Firstly we send a "prepare" message to the destination host. The destination host may reply with a cookie. It may also suggest a URI (in the current Xen implementation it just returns gethostname). Secondly we send a "perform" message to the source host. Correspondingly, there are two new driver API functions: typedef int (*virDrvDomainMigratePrepare) (virConnectPtr dconn, char **cookie, int *cookielen, const char *uri_in, char **uri_out, unsigned long flags, const char *dname, unsigned long resource); typedef int (*virDrvDomainMigratePerform) (virDomainPtr domain, const char *cookie, int cookielen, const char *uri, unsigned long flags, const char *dname, unsigned long resource); Remote support -------------- To make this work in the remote case I have had to export two private functions from the API which only remote should call: __virDomainMigratePrepare __virDomainMigratePerform The reason for this is that libvirtd is just linked to the regular libvirt.so, so can only make calls into libvirt through exported symbols in the dynamic symbol table. There are two corresponding wire messages (REMOTE_PROC_DOMAIN_MIGRATE_PREPARE and REMOTE_PROC_DOMAIN_MIGRATE_PERFORM) but they just do dumb argument shuffling, albeit rather complicated because of the number of arguments passed in and out. The complete list of messages which go across the wire during a migration is: client -- prepare --> destination host client <-- prepare reply -- destination host client -- perform --> source host client <-- perform reply -- source host client -- lookupbyname --> destination host client <-- lookupbyname reply -- destination host Xen URIs -------- Xen recognises the following forms of URI: hostname hostname:port tcp://hostname/ tcp://hostname:port/ Capabilities ------------ I have extended capabilities with <migration_features>. For Xen this is: <capabilities> <host> <migration_features> <live/> <uri_transports> <uri_transport>tcp</uri_transport> </uri_transports> </migration_features> Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

4 18

[Libvir] OpenVZ XML format and VPS properties get/set interface [long]
by Shuveb Hussain 24 Jul '07

24 Jul '07

Hi, I started a discussion on OpenVZ XML format a while ago. But let me do it again with more explanation about OpenVZ this time, so that others can understand how it is different and how this can best fit into the libvirt model of doing things. Terminology: Virtual Private Server (VPS), Virtual Environment (VE) and Domain are all the same. OpenVZ is a lot about providing QoS to its users. About 20 carefully chosen parameters regarding various resources such as memory, CPU, disk and network are chosen. These are then used to provide minimum guarantee on any system running OpenVZ. Most of the time, these are limits that can be set per Virtual Private Server(VPS). In Xen or QEMU, if a disk image is available(Xen needs an additional kernel), it is possible to run the domain. Then forget all about it after the domain is shutoff. This is not possible in OpenVZ. When a new VPS/VE/Domain needs to be created, it needs a file system. This needs to be created along with its related configuration files in specific locations. Only after this can it be started. There is a "destroy" command available in OpenVZ, which is different from the destroy in libvirt. This will completely erase the file system and remove the related config file as well. Since there are many configurable parameters, the OpenVZ tools provides 2 sample templates or profiles on which newly created Virtual Environments(VEs) can be based. So, during VPS creation, rather than taking a million parameters, the name of the profile is taken as an argument and the variables in the file are used to create the VE. These values can later on be overridden and also be optionally stored in the VE's private config file to ensure persistence across reboots. Since there are many parameters needed during VE creation, using the profile name is practical. So, in the proposed XML file, I'm using the profile name. OpenVZ has its own config file format. We are storing the UUID there in a comment, since UUIDs are not used by OpenVZ. While a VE is created, the easiest way to do it is using a so called template cache. This is just a tar file of a Linux distro FS that is used to create a new file system for a VE. There are no disk images. The VE root fs resides on the host file system as a bunch of files and directories. A few template caches are usually available, say one based on Debian, one based on Fedora Core and another based on Suse. The user can choose which one to use while creating a new VE. However, the name of the template cache is not stored anywhere once the VE filesystem is created. I think one more comment is needed in the per-VE config file for this, just as we are storing the UUID. Here is a sample template. This one is called vps.basic, comes with the OpenVZ tools: ----------------------------------------------------------------- ONBOOT="no" # UBC parameters (in form of barrier:limit) # Primary parameters AVNUMPROC="40:40" NUMPROC="65:65" NUMTCPSOCK="80:80" NUMOTHERSOCK="80:80" VMGUARPAGES="6144:2147483647" # Secondary parameters KMEMSIZE="2752512:2936012" TCPSNDBUF="319488:524288" TCPRCVBUF="319488:524288" OTHERSOCKBUF="132096:336896" DGRAMRCVBUF="132096:132096" OOMGUARPAGES="6144:2147483647" # Auxiliary parameters LOCKEDPAGES="32:32" SHMPAGES="8192:8192" PRIVVMPAGES="49152:53575" NUMFILE="2048:2048" NUMFLOCK="100:110" NUMPTY="16:16" NUMSIGINFO="256:256" DCACHESIZE="1048576:1097728" PHYSPAGES="0:2147483647" NUMIPTENT="128:128" # Disk quota parameters (in form of softlimit:hardlimit) DISKSPACE="1048576:1153434" DISKINODES="200000:220000" QUOTATIME="0" # CPU fair sheduler parameter CPUUNITS="1000" ------------------------------------------------------------ Here is the proposed XML format: <domain type='openvz'> <name>105</name> <uuid>8509a1d4-1569-4467-8b37-4e433a1ac7b1</uuid> <filesystem> <template>gentoo-20060317-i686-stage3</template> <quota level='first'>10737418240</quota> <quota level='second' uid='500'>5368709120</quota> </filesystem> <profile>vps.basic</profile> <devices> <interface> <ipaddress>192.168.1.105</ipaddress> </interface> </devices> <nameserver>192.168.1.1</nameserver> <hostname>fedora105</hostname> </domain> I don't think the "filesystem" tag can fit logically into "devices", since it has quota and other information. The "template" is the name of the template cache used to create the VE. One of the main reasons many people(especially hosting providers) use OpenVZ is since it can be used to provide service level agreements. There must be a way to set/get various VPS parameters from libvirt. I understand concerns about driver specific code in libvirt based clients like virt-manager. The capabilities paradigm will not fit here, since this is simply about various properties of the VE/domain, not the hardware or the VM capabilities. Please correct me, if I am wrong. So, how to we do it? Thanks and Regards, -- Shuveb Hussain Unix is very user friendly. It is just a little choosy about who its friends are http://www.binarykarma.com

5 14

[Libvir] [PATCH] virDomainMigrate (version 5) (for discussion only, but getting there)
by Richard W.M. Jones 23 Jul '07

23 Jul '07

This is a rolled up patch for virDomainMigrate, including virsh and Python fixes. Changes since last time: - VIR_MIGRATE_LIVE is exported to Python - 'resource' parameter is now named 'bandwidth' - there is now an extra domainMigrateFinish step It needs a bit more testing which I can't do today while I'm at the office. Rich.

2 1

[Libvir] [PATCH] Document VIRSH_DEFAULT_CONNECT_URI
by David Lutterkort 23 Jul '07

23 Jul '07

I keep forgetting the exact name of VIRSH_DEFAULT_CONNECT_URI - patch to add it to the man page attached. David

2 1

[Libvir] FW: [et-mgmt-tools] Release 0.200.0 of virtinst
by Daniel P. Berrange 20 Jul '07

20 Jul '07

FYI a new libvirt based tool for cloning VMs... ----- Forwarded message from "Daniel P. Berrange" <berrange(a)redhat.com> ----- > Date: Thu, 19 Jul 2007 03:16:56 +0100 > From: "Daniel P. Berrange" <berrange(a)redhat.com> > To: et-mgmt-tools(a)redhat.com > Subject: [et-mgmt-tools] Release 0.200.0 of virtinst > > I'm happy to announce the release of virtinst version 0.200.0 > > This release introduces the virt-clone tool which provides the ability > to clone an existing inactive guest. The disk images will be copied, > new MAC address, UUID and name will be given to the guest. Documentation > is improved with the addition of manual pages for both virt-install and > virt-clone. Validation of input parameters has been further enhanced, > and re-factored to allow sharing with virt-manager. The virt-install > tool can be used to boot live CD images, with or without underlying > storage. > > It is available for download from the usual place: > > http://virt-manager.org/download.html > > Or directly: > > http://virt-manager.org/download/sources/virtinst/virtinst-0.200.0.tar.gz > > Thanks to everyone who has participated in development & testing of the > code in this new release! > > Regards, > Dan. ----- End forwarded message ----- 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 1

[Libvir] [PATCH]0/3 OpenVZ driver : Some clean ups
by Shuveb Hussain 19 Jul '07

19 Jul '07

Hi, Here are some cleanups as per discussions on the list: Change log: * Make sure we are running on OpenVZ enebled kernel in the driver open() method * Some minor macro cleanups * Use generic UUID parsing rather than own and scanf() * Made private functions static * New files utils.[c,h] added that have 2 functions virExec and virExecNonBlock * Removed popen and used virExec in openvz_driver.c * qemu_driver.c updated to use new virExecNonBlock() functions instead of qemudExec, which is removed Thanks, -- Shuveb Hussain Unix is very user friendly. It is just a little choosy about who its friends are http://www.binarykarma.com

2 3

[Libvir] Xen reboot does not work with Sys::Virt perl binding
by Keck, Christian (ext) 19 Jul '07

19 Jul '07

Hello guys! I am using the perl bindings (Sys::Virt) to manage my virtual xen hosts. Because it directly maps the C api, I thought it's a good idea to ask you instead of the author of this module. Xen version is 3.0 with libvirt 0.1.9 on a red hat box. while calling the reboot method i get following error: libvirt error code: 0, message: Unknown problem In this case the virtual system does not reboot. But with shutdown method, which throw the same error, the system shuts down correctly. Here is my code: my $vmm = Sys::Virt->new(); my $dom = eval { $vmm->get_domain_by_name( $p->{ name } ) }; $dom->reboot(&Sys::Virt::Domain::REBOOT_RESTART); Is this a known issue? Thank you for your help. Best regards, Christian

2 2

[Libvir] RFC: format for mapping host devices to guest
by Daniel P. Berrange 19 Jul '07

19 Jul '07

I'm considering the problem of mapping host USB devices through to the guest as one of the 2 remaining bits of USB support needed (the other being the mapping disks through as USB mass storage). I have previously mentioned some ideas in this mail, though looking back on this I don't much like them https://www.redhat.com/archives/libvir-list/2007-March/msg00205.html Also, there was just the message about a desire to support mapping of PCI devices. The USB stuff I want todo with both Xen and QEMU. In Xen, USB device passthrough is based on: usbdevice=host:[bus id].[device id] eg usbdevice=host:003.001 Or usbdevice=host:[vendor id]:[product id] eg usbdevice=host:0483:2016 In QEMU, USB device passthrough is pretty much the same with command line args: -usbdevice host:[bus id].[device id] eg -usbdevice host:003.001 Or -usbdevice host:[vendor id]:[product id] eg -usbdevice host:0483:2016 Finally, in Xen, PCI device passthrough is done with pci=['pci device id',....] eg pci=['00:1f.3'] So we need some form of element under the <devices> section per device to map through. At minimum this will need the source device info. If we are to support unplugging of USB devices in QEMU it is neccessary to have the target device id - this is generated by QEMU when attaching the USB device. So to attach a device: <host bus='usb'> <source device="003.001"/> </host> Or <host bus='usb'> <source vendor='0483' product='2016'> </host> The first form is useful for live hotplug - where you know the exact device instance. The second form is useful for permanent config - you don't know what device id it'll have since it may be plugged into the host in any order. When splitting the XML back out, QEMU would fill in the target info giving us either: <host bus='usb'> <source dev='003.001'/> <target dev='001.005'/> </host> Or <host bus='usb'> <source vendor='0483' product='2016'/> <target dev='001.005'/> </host> This is suitable for giving to virDomainDetachDevice() - we use the target dev for the QEMU monitor 'usb-detach' command. Now PCI would be similar - no hotplug to worry about here though, and no need (or availability) or target info. So the XML would simply be <host bus='pci'> <source dev='00:1f.3'/> </host> NB I choose 'host' as the element name since we're basically giving the guest an arbitrary host device. My previous email had thought about calling it 'bus' but we're not really defining a bus in this context and there may very well be some scenario in the future where we do need to represent the concept of a 'bus' in the XML directly. 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 3

[Libvir] [PATCH]3/3 OpenVZ driver : qemu_driver.c changes
by Shuveb Hussain 19 Jul '07

19 Jul '07

-- Shuveb Hussain Unix is very user friendly. It is just a little choosy about who its friends are http://www.binarykarma.com

1 0

[Libvir] [PATCH]2/3 OpenVZ driver : generic utils
by Shuveb Hussain 19 Jul '07

19 Jul '07

-- Shuveb Hussain Unix is very user friendly. It is just a little choosy about who its friends are http://www.binarykarma.com

1 0

[Libvir] [PATCH]1/3 OpenVZ driver : Cleanups
by Shuveb Hussain 19 Jul '07

19 Jul '07

Cleanup patch, also adds utils.c and utils.h to Makefile.am -- Shuveb Hussain Unix is very user friendly. It is just a little choosy about who its friends are http://www.binarykarma.com

1 0

[Libvir] PATCH: Fix accident close of monitor FD & leak of logfile FD to child
by Daniel P. Berrange 19 Jul '07

19 Jul '07

After the QEMU monitor filehandle has been successfully initialized, we then promptly close it ! It seems a 'return' statement accidentally got removed somewhere during previously refactoring. Second, the QEMU logfile does not have the close-on-exec flag set so the logfile is leaked to child processes. qemu_driver.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) 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 1

[Libvir] XM config <-> XML config
by Arik Raffael Funke 19 Jul '07

19 Jul '07

Hi, I am using Fedora 7 with virt-manager to control my xen domains. I would like to delegate a pci device to one of my domU's. I know how this is done in the XM and sxp config files, but how is it done now? I read that libvirt can convert XM <-> XML and though I could export my current config to XM format, edit it and reimport. How is is done with libvirt? Thanks, Arik

2 3

[Libvir] PATCH: Support input devices in XML format
by Daniel P. Berrange 18 Jul '07

18 Jul '07

This patch is the first step towards supporting USB devices in libvirt XML format. As per the original thread some months back[1], I'm the grouping is being done based on device classes, rather than bus types. So this first patch is actually introducing the concept of 'input' devices. This is best illustrated by examples - PS2 mouse - automatically added when yuou turn on <graphics> in an HVM guest. <input type='mouse' bus='ps2'/> - Xen mouse - automatically added when you turn on <graphics> in a Xen paravirt guest (ie the paravirt input device) <input type='mouse' bus='xen'/> - USB mouse - have to actually ask for this. Not all that useful since its still using relative co-ordinates, but here for completeness. <input type='mouse' bus='usb'/> - USB tablet - have to actually ask for this. Very useful. Windows and Linux can both drive this in absolute co-ordinate mode. <input type='tablet' bus='usb'/> If it possible to omit the 'bus' attribute, in which case it'll automatically set up the device on the most appropriate bus for that type of device. So if you say type='mouse', it'll pick a ps2/xen bus. If you say type='tablet' it'll pick a usb bus. In my previous email[1] I suggested a possible 'model' attribute as well. I since discovered that Xen removed support for the summagraphics tablet completely, so adding a model at this time is a waste of time. We can always choose it add it later if we port to another virt manager here we need to distinguish devices on more than just a (type, bus) pair. As with the previous '<clock/>' element, it is intended that virt-install and virt-manager will automatically add a USB tablet if they you tell them you're installing Windows, or a recent Linux guest. It'll also be possible manally add post-install through the virt-manager UI. For the QEMU driver at least, I intend to make it possible to hot add & remove USB devices. This will appear as impl of virAttach/Detch methods for QEMU at least. Xen doesn't support hot add/remove of USB at this time. Finally, before I actually commit this I think it is well overdue to write some unit tests for the QEMU driver. So I'm going to do a test which chceks XML -> data struct -> QEMU command line, and XML -> data struct -> XML sequences of operations. This should let us exercise all the XML related code in QEMU which is a good start & helpful to avoid regressions. Regards, Dan. [1] http://www.redhat.com/archives/libvir-list/2007-March/msg00205.html -- |=- 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 6

[Libvir] Asynchronous notifications of domain start/stop
by Richard W.M. Jones 18 Jul '07

18 Jul '07

I had some private discussions with a developer who wants to start using libvirt, but requires asynch notifications of domain start/stop (and other stuff, but this will do for the moment). Right now he's using XenStore watches. The alternative of polling didn't look very appealing for him. A rough API for this might look like this: typedef int virEventHandle; /* Register a callback to catch domain start events. */ virEventHandle virEventDomainStart (virConnectPtr, void (*f) (virConnectPtr, int domid)); /* Register a callback to catch domain stop events. */ virEventHandle virEventDomainStop (virConnectPtr, void (*f) (virConnectPtr, int domid)); /* Unregister any type of callback. */ int virEventCancel (virConnectPtr, virEventHandle handle); Implementation -------------- In the Xen case, we can use XenStore watches[1]. Callbacks from XenStore can be translated directly into event callbacks. In the qemu case, the daemon knows when the qemu instances that it manages come and go. In the qemu/remote case, asynch notifications are nominally supported by the protocol, but I haven't really looked at all the implications. Seems like we'd need to use MSG_OOB to signal the client process, and translate the SIGURG signal received into a callback. This is all tricky and Unix-specific though so if you have any better ideas ... We would also need to avoid mixing ordinary request/response with the notifications, which OOB takes care of. Rich. [1] http://wiki.xensource.com/xenwiki/XenBusDriver -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

4 17

[Libvir] virDomainConfigureDevice?
by Hugh Brock 18 Jul '07

18 Jul '07

I'm looking at ways to replicate xm block-configure at the libvirt level. xm block-configure is useful in that it allows you to change the back device for an HVM guest while the guest is still running; this permits you for example to change the CDROM without shutting down your guest or poking around in xenstore. Thus we can have an "eject" button on cdrom devices in virt-manager (or "eject" and "load" buttons, I guess), which we really need. The issue of course is that xm block-configure is specifically intended for block devices and takes arguments for the back device and the front device (among others). We would naturally prefer to accept a block of XML just as virDomainAttachDevice does and then parse it, determine if the device in question is of a type we can actually edit, determine if the new backdev is appropriate, and so on. None of this seems terribly hard, but before I go implement it, are there obvious pitfalls I'm not thinking? And given that virDomainAttachDevice is not implemented for qemud in libvirt (is it?), how do we handle the non-xen case? Thanks, --Hugh -- Red Hat Virtualization Group http://redhat.com/virtualization Hugh Brock | virt-manager http://virt-manager.org hbrock(a)redhat.com | virtualization library http://libvirt.org

4 7

[Libvir] [PATCH] cleanup of virDeviceMode
by Daniel Veillard 18 Jul '07

18 Jul '07

Since this was raised yesterday about the purity of the libvirt API w.r.t. Xen specifics, we don't actually use the virDeviceMode nor have any possible use for them at the API level since basically this kind of informations is always rovided as part of an XML and never as an int/enum. So I propose to just remove them from the headers and API, I don't see how an application could use the enum but maybe I missed one, in which case please tell me :-) Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard(a)redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/

2 2