August 2011 - Devel - libvirt List Archives

[libvirt] ALERT Virtualization Automatic Builder
by berrange＠redhat.com 25 Aug '11

25 Aug '11

Overall status: failed Start date: Thu Aug 25 2011 Start time: 11:05:02 UTC / 07:05:02 EDT Build counter: 1314270302 Build timestamp: 1314270302 URL: http://builder.virt-tools.org/index.html Module: libvirt Status: failed URL: http://builder.virt-tools.org/module-libvirt.html

1 0

[libvirt] Libvirt 0.9.4rc2 + qemu 0.15 : shutdown domain
by Radek Hladik 24 Aug '11

24 Aug '11

Hi again, while experimenting with new libvirt and qemu I discovered different behavior than with my previous versions. When I shutdown machine with shutdown (virsh) or virDomainShutdown (via php) the domain is shutdown but it ends up paused instead of destroyed. The OS is shut down, CPU probably stopped and I can connect to VNC. I would suppose its because of the new reboot feature and the -no-shutdown option for qemu. I am curious whether it is a bug or expected behavior. It would make perfect sense to be able to "semi"shutdown machine like this. But I was not able to reboot the machine from this state by any means. The best I could do was to resume (unpause) the CPU. Radek

1 0

[libvirt] Libvirt 0.9.4rc2 + qemu 0.15 VNC/TLS not working
by Radek Hladik 24 Aug '11

24 Aug '11

Hi all, I am trying new versions of software for my libvirt/qemu farm. I upgraded libvirt to 0.9.3 and I had problems with loading certificates for libvirt itself (probably GNUTLS related). Upgrade to 0.9.4rc2 solved the problem. Now I am trying to upgrade qemu (from 0.12.5) to 0.15. I compiled the SRPM from rawhide for Fedora 13 (OS on my farms) and I copied the binary to one farm, so I can choose qemu version by <emulator> element in guest XML. I had problem with booting off virtio disk, but upgrade of seabios to 0.6.2 solved it (I had to make a wrapper for qemu to be able to add -bios option though). But main issue that I am not able to solve is that VNC is not using TLS. I've found out these: * qemu 0.12.5 is working without problems * there is no error in libvirt or qemu machine log * If I copy the command from libvirt log, remove the network and monitor definition and start it under root, TLS works as expected * the vnc definition on command line is quite simple: -vnc 0.0.0.0:0,password,tls,x509=/etc/pki/libvirt/pki-vnc -k en-us I am thinking whether there is not a problem with monitor setting something after the machine starts. Libvirt does the same with password, so maybe it does something with TLS. Unfortunately I do not know whether there is any option how to debug the monitor libvirt<->qemu communication. And I have one other question. Is there any way how to interact directly with monitor? Qemu has new guest agent that seems to provide some nice functions and expose them via monitor and I would like to test it. Radek

3 3

[libvirt] Automated builds of libvirt
by Daniel P. Berrange 24 Aug '11

24 Aug '11

You may have seen references to the libvirt automated builder with various patches sent to the list. Previously this was running on one of my own private machines, but I've now got it in a dedicated virtual machine, and have it rsyncing its results to a publically visible site: http://builder.virt-tools.org/ It is currently configure to run every 4 hours to build latest GIT master for all the apps listed. It won't rebuild if the source code hasn't changed since the last build, unless one of the dependant modules has changed. If it finds a build failure, it will spam libvir-list with an ALERT message The build logs for the failure can be found on the libvirt status page http://builder.virt-tools.org/module-libvirt.html It will only alert on the first failed build. Until a later build attempt is successfull it will avoid sending further alerts. This should avoid any out of control spamming of the list. Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|

1 0

[libvirt] [PATCH 2/3] qemu: When reserving all functions on a slot don't reserve function 0 on the next slot.
by Shradha Shah 24 Aug '11

24 Aug '11

Signed-off-by: Steve Hodgson <shodgson(a)solarflare.com> --- src/qemu/qemu_command.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index dbfc7d9..12adcc0 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -878,7 +878,7 @@ int qemuDomainPCIAddressReserveSlot(qemuDomainPCIAddressSetPtr addrs, { int function; - for (function = 0; function <= QEMU_PCI_ADDRESS_LAST_FUNCTION; function++) { + for (function = 0; function < QEMU_PCI_ADDRESS_LAST_FUNCTION; function++) { if (qemuDomainPCIAddressReserveFunction(addrs, slot, function) < 0) goto cleanup; } @@ -956,7 +956,7 @@ int qemuDomainPCIAddressReleaseSlot(qemuDomainPCIAddressSetPtr addrs, int slot) dev.addr.pci.bus = 0; dev.addr.pci.slot = slot; - for (*function = 0; *function <= QEMU_PCI_ADDRESS_LAST_FUNCTION; (*function)++) { + for (*function = 0; *function < QEMU_PCI_ADDRESS_LAST_FUNCTION; (*function)++) { addr = qemuPCIAddressAsString(&dev); if (!addr) return -1; -- 1.7.4.4

2 1

[libvirt] [PATCH 1/3] pci: Fix callers of pciDeviceListSteal() when there's more than one device
by Shradha Shah 24 Aug '11

24 Aug '11

pciDeviceListSteal(pcidevs, dev) removes dev from pcidevs reducing the length of pcidevs, so moving onto what was the next dev is wrong. Instead callers should pop entry 0 repeatedly until pcidevs is empty. Signed-off-by: Steve Hodgson <shodgson(a)solarflare.com> Signed-off-by: Shradha Shah <sshah(a)solarflare.com> --- src/qemu/qemu_hostdev.c | 13 ++++++------- 1 files changed, 6 insertions(+), 7 deletions(-) diff --git a/src/qemu/qemu_hostdev.c b/src/qemu/qemu_hostdev.c index 7f5ad51..6f77717 100644 --- a/src/qemu/qemu_hostdev.c +++ b/src/qemu/qemu_hostdev.c @@ -74,7 +74,6 @@ int qemuUpdateActivePciHostdevs(struct qemud_driver *driver, virDomainDefPtr def) { pciDeviceList *pcidevs; - int i; int ret = -1; if (!def->nhostdevs) @@ -83,8 +82,8 @@ int qemuUpdateActivePciHostdevs(struct qemud_driver *driver, if (!(pcidevs = qemuGetPciHostDeviceList(def->hostdevs, def->nhostdevs))) return -1; - for (i = 0; i < pciDeviceListCount(pcidevs); i++) { - pciDevice *dev = pciDeviceListGet(pcidevs, i); + while (pciDeviceListCount(pcidevs) > 0) { + pciDevice *dev = pciDeviceListGet(pcidevs, 0); pciDeviceListSteal(pcidevs, dev); if (pciDeviceListAdd(driver->activePciHostdevs, dev) < 0) { pciFreeDevice(dev); @@ -152,8 +151,8 @@ int qemuPrepareHostdevPCIDevices(struct qemud_driver *driver, } /* Now steal all the devices from pcidevs */ - for (i = 0; i < pciDeviceListCount(pcidevs); i++) { - pciDevice *dev = pciDeviceListGet(pcidevs, i); + while (pciDeviceListCount(pcidevs) > 0) { + pciDevice *dev = pciDeviceListGet(pcidevs, 0); pciDeviceListSteal(pcidevs, dev); } @@ -164,8 +163,8 @@ inactivedevs: /* Only steal all the devices from driver->activePciHostdevs. We will * free them in pciDeviceListFree(). */ - for (i = 0; i < pciDeviceListCount(pcidevs); i++) { - pciDevice *dev = pciDeviceListGet(pcidevs, i); + while (pciDeviceListCount(pcidevs) > 0) { + pciDevice *dev = pciDeviceListGet(pcidevs, 0); pciDeviceListSteal(driver->activePciHostdevs, dev); } -- 1.7.4.4

2 1

[libvirt] [PATCH v2] daemon: Move TLS initialization to virInitialize
by Michal Privoznik 24 Aug '11

24 Aug '11

My previous patch 74c75671331d284e1f777f9692b72e9737520bf0 introduced a regression by removing TLS initialization from client. --- diff to v1: -remove virNetTLSDeinit() daemon/libvirtd.c | 2 -- src/libvirt.c | 3 +++ src/rpc/virnettlscontext.c | 8 -------- src/rpc/virnettlscontext.h | 1 - 4 files changed, 3 insertions(+), 11 deletions(-) diff --git a/daemon/libvirtd.c b/daemon/libvirtd.c index 5969a82..423c3d7 100644 --- a/daemon/libvirtd.c +++ b/daemon/libvirtd.c @@ -1516,7 +1516,6 @@ int main(int argc, char **argv) { virHookCall(VIR_HOOK_DRIVER_DAEMON, "-", VIR_HOOK_DAEMON_OP_START, 0, "start", NULL); - virNetTLSInit(); if (daemonSetupNetworking(srv, config, sock_file, sock_file_ro, ipsock, privileged) < 0) { @@ -1555,7 +1554,6 @@ cleanup: virNetServerProgramFree(qemuProgram); virNetServerClose(srv); virNetServerFree(srv); - virNetTLSDeinit(); if (statuswrite != -1) { if (ret != 0) { /* Tell parent of daemon what failed */ diff --git a/src/libvirt.c b/src/libvirt.c index c8af3e1..fdf3923 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -40,6 +40,7 @@ #include "memory.h" #include "configmake.h" #include "intprops.h" +#include "rpc/virnettlscontext.h" #ifndef WITH_DRIVER_MODULES # ifdef WITH_TEST @@ -409,6 +410,8 @@ virInitialize(void) virLogSetFromEnv(); + virNetTLSInit(); + VIR_DEBUG("register drivers"); #if HAVE_WINSOCK2_H diff --git a/src/rpc/virnettlscontext.c b/src/rpc/virnettlscontext.c index 8482eaf..e901b63 100644 --- a/src/rpc/virnettlscontext.c +++ b/src/rpc/virnettlscontext.c @@ -1410,11 +1410,3 @@ void virNetTLSInit(void) { gnutls_global_init(); } - -/* - * See virNetTLSInit - */ -void virNetTLSDeinit(void) -{ - gnutls_global_deinit(); -} diff --git a/src/rpc/virnettlscontext.h b/src/rpc/virnettlscontext.h index 99f31b9..fdfce6d 100644 --- a/src/rpc/virnettlscontext.h +++ b/src/rpc/virnettlscontext.h @@ -31,7 +31,6 @@ typedef virNetTLSSession *virNetTLSSessionPtr; void virNetTLSInit(void); -void virNetTLSDeinit(void); virNetTLSContextPtr virNetTLSContextNewServerPath(const char *pkipath, bool tryUserPkiPath, -- 1.7.3.4

2 1

[libvirt] [PATCH] util: only fchown newly created files in virFileOpenAs
by Laine Stump 24 Aug '11

24 Aug '11

virFileOpenAs takes desired uid:gid as arguments, and not only uses them for a fork/setuid/setgid when retrying failed open operations, but additionally always forces the opened file to be owned by the given uid:gid. One example of the problems this causes is that, when restoring a domain from a file that is owned by the qemu user, opening the file chowns it to root. if dynamic_ownership=1 this is coincidentally expected, but if dynamic_ownership=0, no existing file should ever have its ownership changed. This patch adds an extra check before calling fchown() - it only does it if O_CREAT was passed to virFileOpenAs() in the openflags. --- src/util/util.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/util/util.c b/src/util/util.c index b278165..9556bdb 100644 --- a/src/util/util.c +++ b/src/util/util.c @@ -697,6 +697,7 @@ virFileOpenAsNoFork(const char *path, int openflags, mode_t mode, goto error; } if (((st.st_uid != uid) || (st.st_gid != gid)) + && (openflags & O_CREAT) && (fchown(fd, uid, gid) < 0)) { ret = -errno; virReportSystemError(errno, _("cannot chown '%s' to (%u, %u)"), -- 1.7.3.4

2 1

[libvirt] [PATCH 3/3] If VIR_MIGRATE_PERSIST_DEF is set then make the live configuration transient
by Shradha Shah 24 Aug '11

24 Aug '11

Make the migrated VM config as transient. Ensures transient hotplug operations do not persist past shutdown. Currently the running VM config is transient but that of a migrated VM is persistent. Signed-off-by: Steve Hodgson <shodgson(a)solarflare.com> Signed-off-by: Shradha Shah <sshah(a)solarflare.com> --- src/qemu/qemu_migration.c | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index a84faf6..aa458c3 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -2509,6 +2509,9 @@ qemuMigrationFinish(struct qemud_driver *driver, goto endjob; } + /* Make the VM transient again to match qemuProcessStart */ + virDomainObjSetDefTransient(driver->caps, vm, true); + event = virDomainEventNewFromObj(vm, VIR_DOMAIN_EVENT_DEFINED, newVM ? -- 1.7.4.4

1 0

[libvirt] [PATCH] daemon: Move TLS initialization to virInitialize
by Michal Privoznik 24 Aug '11

24 Aug '11

My previous patch 74c75671331d284e1f777f9692b72e9737520bf0 introduced a regression by removing TLS initialization from client. --- daemon/libvirtd.c | 1 - src/libvirt.c | 3 +++ 2 files changed, 3 insertions(+), 1 deletions(-) diff --git a/daemon/libvirtd.c b/daemon/libvirtd.c index 5969a82..8f04a99 100644 --- a/daemon/libvirtd.c +++ b/daemon/libvirtd.c @@ -1516,7 +1516,6 @@ int main(int argc, char **argv) { virHookCall(VIR_HOOK_DRIVER_DAEMON, "-", VIR_HOOK_DAEMON_OP_START, 0, "start", NULL); - virNetTLSInit(); if (daemonSetupNetworking(srv, config, sock_file, sock_file_ro, ipsock, privileged) < 0) { diff --git a/src/libvirt.c b/src/libvirt.c index c8af3e1..fdf3923 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -40,6 +40,7 @@ #include "memory.h" #include "configmake.h" #include "intprops.h" +#include "rpc/virnettlscontext.h" #ifndef WITH_DRIVER_MODULES # ifdef WITH_TEST @@ -409,6 +410,8 @@ virInitialize(void) virLogSetFromEnv(); + virNetTLSInit(); + VIR_DEBUG("register drivers"); #if HAVE_WINSOCK2_H -- 1.7.3.4

3 4