This is the API agreed on in:
https://www.redhat.com/archives/libvir-list/2011-May/msg00026.html
(with a slight name change to use "...begin" rather than
"...start"). This implements transactional changes to the host network
config. When a transaction is begun with ncf_change_begin(), all other
netcf APIs will continue to work as they always have, but a snapshot
of the existing config will be taken, allowing reversion (rollback,
using ncf_change_rollback()) to the exact state of config at the time
ncf_change_begin() was called. Alternately, if it's determined that
the new changes are acceptable, ncf_change_commit() can be called,
which will eliminate the snapshot and make the changes permanent.
As a failsafe measure, if neither ncf_change_commit() or
ncf_change_rollback() is called by the next time the system reboots,
the netcf-transaction initscript will be automatically called to
rollback the changes.
Commit f044376530 replaced openvz_readline with getline and
changed EOF-handling in the openvzGetVPSUUID.
This patch restores original EOF-handling.
Reported by Jean-Baptiste Rouault.
This patch allows to modify interfaces of domain(qemu)
* src/conf/domain_conf.c src/conf/domain_conf.h src/libvirt_private.syms:
(virDomainNetInsert) : Insert a network device to domain definition.
(virDomainNetIndexByMac) : Returns an index of net device in array.
(virDomainNetRemoveByMac): Remove a NIC of passed MAC address.
* src/qemu/qemu_driver.c
(qemuDomainAttachDeviceConfig): add codes for NIC.
(qemuDomainDetachDeviceConfig): add codes for NIC.
Before commit 145d6cb05c (in August 2010) absolute file names
in VMX and domain XML configs were handled correctly. But this got
lost during the refactoring. The test cases didn't highlight this
problem because they have their own set of file name handling
functions. The actual ones require a real connection to an ESX
server. Also the test case functions always worked correctly.
Fix the regression and add a new in-the-wild VMX file that contains
such a problematic absolute path. Even though this test case won't
protect against new regressions.
Reported by lofic (IRC nick)
As reported by Diego Blanco in
https://bugzilla.redhat.com/show_bug.cgi?id=702602
commit f0443765 which replaced openvz_readline to getline(3)
broke OpenVZ driver as it changed semantics of EOF-handling
when parsing OpenVZ configuration.
There're several other issues reported with current OpenVZ driver:
#1: unclear error message when parsing "CPUS=" line
#2: openvz driver goes into crashing loop
#3: "NETIF=" line in configuration is not parsed correctly
#4: aborts even when optional parameter is missing
#5: there's a potential memory leak
This updated patch to fix #[145]. This patch does not fix #[23]
as I haven't verified these yet, but this at least got me to run
OpenVZ on libvirt once again.
Coverity spotted this off-by-one. Thankfully, no one in libvirt
was ever calling virAuditSend with an argument of 3.
* src/util/virtaudit.c (virAuditSend): Use correct comparison.
Originally most of libvirt domain-specific calls were blocking
during a migration.
A new mechanism to allow specific calls (blkstat/blkinfo) to be
executed in such condition has been implemented.
In the long term it'd be desirable to get a more general
solution to mark further APIs as migration safe, without needing
special case code.
* src/qemu/qemu_migration.c: add some additional job signal
flags for doing blkstat/blkinfo during a migration
* src/qemu/qemu_domain.c: add a condition variable that can be
used to efficiently wait for the migration code to clear the
signal flag
* src/qemu/qemu_driver.c: execute blkstat/blkinfo using the
job signal flags during migration
Based on the device attach/detach code from the QEMU driver,
but using the new functions to create the structures associated.
* src/libxl/libxl_driver.c: implements domainAttachDevice,
domainAttachDeviceFlags, domainDetachDevice, domainDetachDeviceFlags
and domainUpdateDeviceFlags
Create 3 new function refactored from previous list ones and
exports them internally to the driver
* src/libxl/libxl_conf.c src/libxl/libxl_conf.h: create libxlMakeDisk,
libxlMakeNic libxlMakeVfb out of the exsting static List functions
and exports them
When modifying the disk devices of a live domain and the domain
configuration, the function qemuDomainAttachDeviceConfig
first sets dev->data->disk to NULL. Later qemuDomainAttachDeviceLive
accesses dev->data.disk and causes a segfault.
* src/qemu/qemu_driver.c: fix qemuDomainModifyDeviceFlags() accordingly
Anything generated that must end up in the tarball must either
have unconditional rules for generation (remote_protocol.c) or
must live in libvirt.git for the case where the person running
'make dist' has disabled the configure options that control the
rebuild of the generated file (remote_protocol-structs).
* src/Makefile.am (remote_protocol-structs): Add a dependency and
document why it must live in git.
($(srcdir)/remote/%_protocol.c, $(srcdir)/remote/%_protocol.c):
Unconditionally generate.
http://lists.gnu.org/archive/html/qemu-devel/2011-05/threads.html#02162
Currently, qemu silently clips any JSON integer in the range
0x8000000000000000 - 0xffffffffffffffff (all numbers in this range
will be clipped to 0x7fffffffffffffff == LLONG_MAX).
To avoid this, pass these as signed 64 bit integers in the QMP
request.
In most cases this affects flags parameters that are unsigned in the
public and driver API but signed in the XDR protocol. Switch the
XDR protocol to unsigned for those.
A counterexample is virNWFilterGetXMLDesc. Its flags parameter is signed
in the public API and XDR protocol, but unsigned in the driver API.
virNodeGetFreeMemory used unsigned long long in the public API but
signed hyper in the XDR protocol. Convert the XDR protocol to use
unsigned hyper.
As explained by Eric before, this doesn't affect the on-the-wire protocol.
Several functions return values by reference parameters. This is realized
by passing the members of remote_CALL_ret by reference to the called
function.
The position of this parameters in the function signature follows some
patterns with some exceptions. This patterns and exceptions are hardcoded
in the generator.
Add an insert@<offset> annotation to the remote_CALL_ret struct members
for functions that return lists to remove some of the hardcoded patterns
and exceptions.
The current virDomainMigrateFinish3 method signature attempts to
distinguish two types of errors, by allowing return with ret== 0,
but ddomain == NULL, to indicate a failure to start the guest.
This is flawed, because when ret == 0, there is no way for the
virErrorPtr details to be sent back to the client.
Change the signature of virDomainMigrateFinish3 so it simply
returns a virDomainPtr, in the same way as virDomainMigrateFinish2
The disk locking code will protect against the only possible
failure mode this doesn't account for (loosing conenctivity to
libvirtd after Finish3 starts the CPUs, but before the client
sees the reply for Finish3).
* src/driver.h, src/libvirt.c, src/libvirt_internal.h: Change
virDomainMigrateFinish3 to return a virDomainPtr instead of int
* src/remote/remote_driver.c, src/remote/remote_protocol.x,
daemon/remote.c, src/qemu/qemu_driver.c, src/qemu/qemu_migration.c:
Update for API change
When doing migration, if an error occurs in Perform, it must not
be overwritten during Finish/Confirm steps. If an error occurs
in Finish, it must not be overwritten in Confirm.
Previous commit a9d12c2444 added
code to qemudDomainMigrateFinish2 to preserve the error. This
is not the right place, because it is not applicable in non-p2p
migration. The src/libvirt.c virDomainMigrateV2/3 methods need
code to preserve errors for non-p2p migration, while the
doPeer2PeerMigrate2 and doPeer2PeerMigrate3 methods contain
code to preverse errors for p2p migration.
Remove the bogus error preservation from qemudDomainMigrateFinish2
and qemudDomainMigrateFinish3.
Fix virDomainMigrateV3 and doPeer2PeerMigrate3 so that they
preserve any error hit during the Finish3 step, before invoking
Confirm3.
Finally if qemuMigrationFinish fails to resume the CPUs, it must
preserve the error before tearing down the VM, so that VM cleanup
doesn't overwrite it.
* src/libvirt.c: Preserve error before invoking Confirm3
* src/qemu/qemu_driver.c: Remove bogus error preservation
code in qemudDomainMigrateFinish2/qemudDomainMigrateFinish3
* src/qemu/qemu_migration.c: Preserve error before invoking Confirm3
and after resume fails in qemuMigrationFinish.
* src/libvirt.c: Add further debug lines in helper APIs for
migration
* src/qemu/qemu_migration.c: Add debug lines for all internal
migration API parameters
Even when failing to start CPUs, the finish method was returning
a success result. Fix this so that the QEMU process is killed
off when finish fails under v3 protocol. Also rename the
killOnFinish boolean to 'v3proto' to make it clearer that this
is a tunable based on the migration protocol version
* src/qemu/qemu_driver.c: Update for API change
* src/qemu/qemu_migration.c, src/qemu/qemu_migration.h: Kill
VM in qemuMigrationFinish if failing to start CPUs
The SPICE seamless migration process requires data to be passed
back from the target host, to the source host via a cookie.
The cookie includes the target host's hostname, but this was not
stored, merely validated. This patch explicitly records the
remote hostname after parsing the cookie, and uses it when
initiating the SPICE migration
* qemu/qemu_migration.c: Fix SPICE seamless migration hostname
Before running perform in peer-2-peer migration, the current
guest state must be recorded, so that non-live migration can
currently unpause a running guest on completion.
* src/qemu/qemu_migration.c: Move check for offline guest
to fix non-live migration
There are two pieces of information which are desirable for
migration, which cannot be supplied by applications
- The explicit QEMU migration URI, while using Peer2Peer
migration
- An override for the target VM XML
This introduces two new public APIs to support these extra
parameters. There is no need for extra wire protocool changes,
since this is supported by the v3 migration enhancements
* include/libvirt/libvirt.h.in,
src/libvirt.c, src/libvirt_public.syms: Add virDomainMigrate2
and virDomainMigrateToURI2
The virDomainMigratePerform3 currently has a single URI parameter
whose meaning varies. It is either
- A QEMU migration URI (normal migration)
- A libvirtd connection URI (peer2peer migration)
Unfortunately when using peer2peer migration, without also
using tunnelled migration, it is possible that both URIs are
required.
This adds a second URI parameter to the virDomainMigratePerform3
method, to cope with this scenario. Each parameter how has a fixed
meaning.
NB, there is no way to actually take advantage of this yet,
since virDomainMigrate/virDomainMigrateToURI do not have any
way to provide the 2 separate URIs
* daemon/remote.c, src/remote/remote_driver.c,
src/remote/remote_protocol.x, src/remote_protocol-structs: Add
the second URI parameter to perform3 message
* src/driver.h, src/libvirt.c, src/libvirt_internal.h: Add
the second URI parameter to Perform3 method
* src/libvirt_internal.h, src/qemu/qemu_migration.c,
src/qemu/qemu_migration.h: Update to handle URIs correctly
This extends the v3 migration protocol such that the
virDomainMigrateBegin3 and virDomainMigratePerform3
methods accept an application supplied XML config for
the target VM.
If the 'xmlin' parameter is NULL, then Begin3 uses the
current guest XML as normal. A driver implementing the
Begin3 method should either reject all non-NULL 'xmlin'
parameters, or strictly validate that the app supplied
XML does not change guest ABI.
The Perform3 method also needed the xmlin parameter to
cope with the Peer2Peer migration sequence.
NB it is not yet possible to use this capability since
neither of the public virDomainMigrate/virDomainMigrateToURI
methods have a way to pass in XML.
* daemon/remote.c, src/remote/remote_driver.c,
src/remote/remote_protocol.x, src/remote_protocol-structs:
Add 'remote_string xmlin' parameter to begin3/perform3
RPC messages
* src/libvirt.c, src/driver.h, src/libvirt_internal.h: Add
'const char *xmlin' parameter to Begin3/Perform3 methods
* src/qemu/qemu_driver.c, src/qemu/qemu_migration.c,
src/qemu/qemu_migration.h: Pass xmlin parameter around
migration methods
Otherwise an attempt to use virConnectOpen or virConnectOpenAuth without
auth pointer results in the driver declining the URI and libvirt falling
back to the remote driver for an esx:// URI.
The cur_vcpus member of struct libxl_domain_build_info was incorrectly
initialized to the number of vcpus, when it should have been interpreted
as a bitmap, where bit X corresponds to online/offline status of vcpuX.
To complicate matters, cur_vcpus is an int, so only 32 vcpus can be
set online. Add a check to ensure vcpus does not exceed this limit.
V2: Eric Blake noted a compilation pitfal when '1 << 32' on an int.
Account for vcpus == 32.
We don't use the gnulib vsnprintf replacement, which means that
on mingw, vsnprintf doesn't support %zn or %lln.
And as it turns out, VIR_GET_VAR_STR was a rather inefficient
reimplementation of virVasprintf logic.
* src/util/logging.c (VIR_GET_VAR_STR): Drop.
(virLogMessage): Inline a simpler version here.
* src/util/virterror.c (VIR_GET_VAR_STR, virRaiseErrorFull):
Likewise.
Reported by Matthias Bolte.
Saving domain to previously created file changes also its ownership.
This is certainly not what users want if some conditions are met:
it is a regular, local file and dynamic_ownership is off.
NB: the enum that uses the string vnet-host (now changed to vhost-net)
is used in XML, but fortunately that hasn't been in an official
release yet, so it can still be fixed.
Since -vnc uses ':' to separate the address from the port, raw
IPv6 addresses need to be escaped like [addr]:port
* src/qemu/qemu_command.c: Escape raw IPv6 addresses with []
* tests/qemuxml2argvdata/qemuxml2argv-graphics-vnc.args,
tests/qemuxml2argvdata/qemuxml2argv-graphics-vnc.xml: Tweak
to test Ipv6 escaping
* docs/schemas/domain.rng: Allow Ipv6 addresses, or hostnames
in <graphics> listen attributes
The qemuMigrationConfirm method shouldn't deal with final VM
cleanup, since it can be called from the peer2peer migration,
which expects to still use the 'vm' object afterwards.
Push the cleanup code out of qemuMigrationConfirm, into its
caller, qemuDomainMigrateConfirm3
* src/qemu/qemu_driver.c: Add VM cleanup code to
qemuDomainMigrateConfirm3
* src/qemu/qemu_migration.c, src/qemu/qemu_migration.h: Remove
job handling cleanup from qemuMigrationConfirm
To allow new mandatory migration cookie data to be introduced,
add support for checking supported feature flags when parsing
migration cookie.
* src/qemu/qemu_migration.c: Feature flag checking in migration
cookie parsing
Two additional places need initgroups call to properly work in an
environment where the UID is allowed to open/create stuff through its
supplementary groups.
This adds a streaming-video=filter|all|off attribute. It is used to change
the behavior of video stream detection in spice, the default is filter (the
default for libvirt is not to specify it - the actual default is defined in
libspice-server.so).
Usage:
<graphics type='spice' autoport='yes'>
<streaming mode='off'/>
</graphics>
Tested with the above and with tests/qemuxml2argvtest.
Signed-off-by: Alon Levy <alevy@redhat.com>
This patch enables filtering of gratuitous ARP packets using the following XML:
<rule action='accept' direction='in' priority='425'>
<arp gratuitous='true'/>
</rule>
This was discussed in:
https://www.redhat.com/archives/libvir-list/2011-May/msg01370.html
The capabilities code only sets the flag to allow use of vhost-net if
kvm is detected (set if the help string contains "(qemu-kvm-" or
"(kvm-"), but actually vhost-net is available in some qemu builds that
don't have kvm in their name, so just checking for ",vhost=" is enough.
When using TLS authentication and operating as the non-root user,
initially attempt to use that specific user's TLS certificates before
attempting to use the system wide TLS certificates.
Signed-off-by: Doug Goldstein <cardoe@cardoe.com>
Otherwise qemu is unable to write to it, with the error:
libvir: QEMU error : internal error unable to execute QEMU command 'memsave': Could not open '/var/cache/libvirt/qemu/qemu.mem.RRNvLv'
The v2 migration protocol had a limit on cookie length that was
too small to be useful for QEMU. Avoid generating cookies with
v2 protocol, so that old libvirtd can still reliably migrate a
guest to new libvirtd uses v2 protocol.
* src/qemu/qemu_driver.c: Avoid migration cookies with v2
migration
When generating a cookie for a guest with no data, the
QEMU_MIGRATION_COOKIE_GRAPHICS flag was set even if no
graphics data was added. Avoid setting the flag unless
it was needed, also add a safety check for mig->graphics
being non-NULL
* src/qemu/qemu_migration.c: Avoid cookie crash for guest
with no graphics
The internal virDomainMigratePeer2Peer and virDomainMigrateDirect
helper methods were not checking whether the target supports the
v3 migration protocol.
* src/libvirt.c: Use v3 migration protocol for p2p/direct
migration if available.
Some bogus apps are generating a VNC/SPICE/RFB listen attribute
with no content. This then causes a failure with the graphics
migration cookie parsing. Blank out the 'listenAddr' parameter
after parsing domain XML if it is the empty string, so the host
default takes over
* src/qemu/qemu_migration.c: Blank out listenAddr parameter
if empty
The on-the-wire protocol is identical; XDR guarantees that
both 'hyper' and 'unsigned hyper' are transmitted as 8 bytes.
* src/remote/remote_protocol.x (remote_get_version_ret)
(remote_get_lib_version_ret): Match public API.
* daemon/remote_generator.pl: Drop special case.
* src/remote_protocol-structs: Reflect updated type.
Clang couldn't quite see that the same condition of
(flags & VIR_DOMAIN_MEM_CONFIG) is used twice, such that
the second block is guaranteed that def was assigned in
the first block.
* src/libxl/libxl_driver.c (libxlDomainSetMemoryFlags): Add a hint
for clang.
Improve invalid argument checks in the size query case. The drivers already
relied on this unchecked behavior.
Relax the implementation of virDomainGet(Memory|Blkio)MemoryParameters
in the drivers and allow to pass more memory than necessary for all
parameters.
Add invalid argument checks for params and nparams to the public API
and remove them from the drivers (e.g. xend).
Add subset handling to libxl and test drivers.
params and nparams are essential and cannot be NULL. Check this in
libvirt.c and remove redundant checks from the drivers (e.g. xend).
Instead of enforcing that nparams must point to exact same value as
returned by virDomainGetSchedulerType relax this to a lower bound
check. This is what some drivers (e.g. xen hypervisor and esx)
already did. Other drivers (e.g. xend) didn't check nparams at all
and assumed that there is enough space in params.
Unify the behavior in all drivers to a lower bound check and update
nparams to the number of valid values in params on success.
Some drivers assumed it can be NULL (e.g. qemu and lxc) and check it
before assigning to it, other drivers assumed it must be non-NULL
(e.g. test and esx) and just assigned to it.
Unify this to nparams being optional and document it.
This error code has existed since the dawn of time, yet the messages it
generates are almost universally busted. Here's a small sampling:
src/conf/domain_conf.c:4889 : XML description for missing root element is not well formed or invalid
src/conf/domain_conf.c:4951 : XML description for unknown device type is not well formed or invalid
src/conf/domain_conf.c:5460 : XML description for maximum vcpus must be an integer is not well formed or invalid
src/conf/domain_conf.c:5468 : XML description for invalid maxvcpus %(count)lu is not well formed or invalid
Fix up the error code to instead be
XML error: <msg>
Adjust the few locations that were using the original correctly (or shouldn't
have been using the error code at all).
v2:
Fix wording of error code without a passed argument
starting with kernel 2.6.38 macvtap supports a 'passthru' mode for
attaching virtual functions of a SRIOV capable network card directly to a VM.
This patch adds the capability to configure such a device.
Signed-off-by: Dirk Herrendoerfer <d.herrendoerfer@herrendoerfer.name>
<sys/syslimits.h> is not standardized, so portable programs should
not need to rely on it. If there really is something that we need
where <sys/syslimits.h> provided the limit but <limits.h> did not,
then that would be a candidate for fixing in gnulib. But this patch
did not turn up any compilation failures on Linux.
* src/internal.h (includes): Drop unused header.
* daemon/libvirtd.h (includes): Likewise.
* configure.ac (AC_CHECK_HEADERS): Likewise.
Based on a report by Matthias Bolte.
POSIX allows sysconf(_SC_GETPW_R_SIZE_MAX) to return -1 if there
is no fixed limit, and requires ERANGE errors to track real size.
Model our behavior after the example in POSIX itself:
http://pubs.opengroup.org/onlinepubs/9699919799/functions/getpwuid_r.html
Also, on error for get*_r functions, errno is undefined, and the
real error was the return value.
* src/util/util.c (virGetUserEnt, virGetUserID, virGetGroupID)
(virSetUIDGID): Cope with sysconf failure or too small buffer.
Reported by Matthias Bolte.
virRunWithHook is now unused, so we can drop it. Tested w/ raw + qcow2
volume creation and copying.
v2:
Use opaque data to skip hook second time around
Simply command building
v3:
Drop explicit FindFileInPath
virStreamNew needs to dispatch the error that virGetStream reports
on failure.
remoteCreateClientStream can fail due to virStreamNew or due to
VIR_ALLOC. Report OOM error for VIR_ALLOC failure to report errors
in all error cases.
Remove OOM error reporting from remoteCreateClientStream callers.
When the session has expired then multiple threads can race while
reestablishing it.
This race condition is not that critical as it requires a special usage
pattern to be triggered. It can only happen when an application doesn't
do API calls for quite some time (the session expires after 30 min
inactivity) and then multiple threads doing simultaneous API calls and
end up doing simultaneous calls to esxVI_EnsureSession.
virsh didn't call virInitialize(), which (among other things)
initializes virLastErr thread local variable. As a result of that, virsh
could just segfault in virEventRegisterDefaultImpl() since that is the
first call that touches (resets) virLastErr.
I have no idea what lucky coincidence made this bug visible but I was
able to reproduce it in 100% cases but only in one specific environment
which included building in sandbox.
Mingw execve() has a broken signature. Disable this
function until gnulib fixes the signature, since we
don't really need this on Win32 anyway.
* src/util/command.c: Disable virCommandExec on Win32
When failing to marshall an XDR message, include the
full program/version/status/proc/type info, to allow
easier debugging & diagnosis of the problem.
* src/remote/remote_driver.c: Improve error when marshalling
fails
By running the doTunnelSendAll code in a separate thread, the
main thread can do qemuMigrationWaitForCompletion as with
normal migration. This in turn ensures that job signals work
correctly and that progress monitoring can be done
* src/qemu/qemu_migration.c: Run tunnelled migration in
separate thread
Cancelling the QEMU migration may cause QEMU to flush pending
data on the migration socket. This may in turn block QEMU if
nothing reads from the other end of the socket. Closing the
socket before cancelling QEMU migration avoids this possible
deadlock.
* src/qemu/qemu_migration.c: Close sockets before cancelling
migration on failure
The 'nbytes' variable was not re-initialized to the
buffer size on each iteration of the tunnelled migration
loop. While saferead() will ensure a full read, except
on EOF, it is clearer to use the real buffer size
* src/qemu/qemu_migration.c: Always read full buffer of data
The qemuMigrationWaitForCompletion method contains a loop which
repeatedly queries QEMU to check migration progress, and also
processes job signals (pause, setspeed, setbandwidth, cancel).
The tunnelled migration loop does not currently support this
functionality, but should. Refactor the code to allow it to
be used with tunnelled migration.
Implement the v3 migration protocol, which has two extra
steps, 'begin' on the source host and 'confirm' on the
source host. All other methods also gain both input and
output cookies to allow bi-directional data passing at
all stages.
The QEMU peer2peer migration method gains another impl
to provide the v3 migration. This finally allows migration
cookies to work with tunnelled migration, which is required
for Spice seamless migration & the lock manager transfer
* src/qemu/qemu_driver.c: Wire up migrate v3 APIs
* src/qemu/qemu_migration.c, src/qemu/qemu_migration.h: Add
begin & confirm methods, and peer2peer impl of v3
Merge the doNonTunnelMigrate2 and doTunnelMigrate2 methods
into one doPeer2PeerMigrate2 method, since they are substantially
the same. With the introduction of v3 migration, this will be
even more important, to avoid massive code duplication.
* src/qemu/qemu_migration.c: Merge tunnel & non-tunnel migration
The v2 migration protocol was accidentally missing out the
finish step, when prepare succeeded, but returned an invalid
URI
* src/libvirt.c: Teardown VM if prepare returns invalid URI
To facilitate the introduction of the v3 migration protocol,
the doTunnelMigrate method is refactored into two pieces. One
piece is intended to mirror the flow of virDomainMigrateVersion2,
while the other is the helper for setting up sockets and processing
the data.
Previously socket setup would be done before the 'prepare' step,
so errors could be dealt with immediately, avoiding need to shut
off the destination QEMU. In the new split, socket setup is done
after the 'prepare' step. This is not a serious problem, since
the control flow already requires calling 'finish' to tear down
the destination QEMU upon several errors.
* src/qemu/qemu_migration.c:
Use the graphics information from the QEMU migration cookie to
issue a 'client_migrate_info' monitor command to QEMU. This causes
the SPICE client to automatically reconnect to the target host
when migration completes
* src/qemu/qemu_migration.c: Set data for SPICE client relocation
before starting migration on src
* src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h,
src/qemu/qemu_monitor_json.c, src/qemu/qemu_monitor_json.h,
src/qemu/qemu_monitor_text.c, src/qemu/qemu_monitor_text.h: Add
new qemuMonitorGraphicsRelocate() command
Extend the QEMU migration cookie structure to allow information
about the destination host graphics setup to be passed by to
the source host. This will enable seamless migration of any
connected graphics clients
* src/qemu/qemu_migration.c: Add graphics info to migration
cookies
* daemon/libvirtd.c: Always initialize gnutls to enable
x509 cert parsing in QEMU
The migration protocol has support for a 'cookie' parameter which
is an opaque array of bytes as far as libvirt is concerned. Drivers
may use this for passing around arbitrary extra data they might
need during migration. The QEMU driver needs to do a few things:
- Pass hostname/uuid to allow strict protection against localhost
migration attempts
- Pass SPICE/VNC server port from the target back to the source to
allow seamless relocation of client sessions
- Pass lock driver state from source to destination
This patch introduces the basic glue for handling cookies
but only includes the host/guest UUID & name.
* src/libvirt_private.syms: Export virXMLParseStrHelper
* src/qemu/qemu_migration.c, src/qemu/qemu_migration.h: Parsing
and formatting of migration cookies
* src/qemu/qemu_driver.c: Pass in cookie parameters where possible
* src/remote/remote_protocol.h, src/remote/remote_protocol.x: Change
cookie max length to 16384 bytes
The qemuMigrationPrepareTunnel method should not unlock the
qemu driver, since that is the caller's job.
* src/qemu/qemu_migration.c: Fix qemuMigrationPrepareTunnel
unlocking of QEMU driver
Migration just seems to go from bad to worse. We already had to
introduce a second migration protocol when adding the QEMU driver,
since the one from Xen was insufficiently flexible to cope with
passing the data the QEMU driver required.
It turns out that this protocol still has some flaws that we
need to address. The current sequence is
* Src: DumpXML
- Generate XML to pass to dst
* Dst: Prepare
- Get ready to accept incoming VM
- Generate optional cookie to pass to src
* Src: Perform
- Start migration and wait for send completion
- Kill off VM if successful, resume if failed
* Dst: Finish
- Wait for recv completion and check status
- Kill off VM if unsuccessful
The problems with this are:
- Since the first step is a generic 'DumpXML' call, we can't
add in other migration specific data. eg, we can't include
any VM lease data from lock manager plugins
- Since the first step is a generic 'DumpXML' call, we can't
emit any 'migration begin' event on the source, or have
any hook that runs right at the start of the process
- Since there is no final step on the source, if the Finish
method fails to receive all migration data & has to kill
the VM, then there's no way to resume the original VM
on the source
This patch attempts to introduce a version 3 that uses the
improved 5 step sequence
* Src: Begin
- Generate XML to pass to dst
- Generate optional cookie to pass to dst
* Dst: Prepare
- Get ready to accept incoming VM
- Generate optional cookie to pass to src
* Src: Perform
- Start migration and wait for send completion
- Generate optional cookie to pass to dst
* Dst: Finish
- Wait for recv completion and check status
- Kill off VM if failed, resume if success
- Generate optional cookie to pass to src
* Src: Confirm
- Kill off VM if success, resume if failed
The API is designed to allow both input and output cookies
in all methods where applicable. This lets us pass around
arbitrary extra driver specific data between src & dst during
migration. Combined with the extra 'Begin' method this lets
us pass lease information from source to dst at the start of
migration
Moving the killing of the source VM out of Perform and
into Confirm, means we can now recover if the dst host
can't successfully Finish receiving migration data.
Change all the driver struct initializers to use the
C99 style, leaving out unused fields. This will make
it possible to add new APIs without changing every
driver. eg change:
qemudDomainResume, /* domainResume */
qemudDomainShutdown, /* domainShutdown */
NULL, /* domainReboot */
qemudDomainDestroy, /* domainDestroy */
to
.domainResume = qemudDomainResume,
.domainShutdown = qemudDomainShutdown,
.domainDestroy = qemudDomainDestroy,
And get rid of any existing C99 style initializersr which
set NULL, eg change
.listPools = vboxStorageListPools,
.numOfDefinedPools = NULL,
.listDefinedPools = NULL,
.findPoolSources = NULL,
.poolLookupByName = vboxStoragePoolLookupByName,
to
.listPools = vboxStorageListPools,
.poolLookupByName = vboxStoragePoolLookupByName,
Fix some driver names:
s/virDrvCPUCompare/virDrvCompareCPU/
s/virDrvCPUBaseline/virDrvBaselineCPU/
s/virDrvQemuDomainMonitorCommand/virDrvDomainQemuMonitorCommand/
s/virDrvSecretNumOfSecrets/virDrvNumOfSecrets/
s/virDrvSecretListSecrets/virDrvListSecrets/
And some driver struct field names:
s/getFreeMemory/nodeGetFreeMemory/
Only in drivers which use virDomainObj, drivers that query hypervisor
for domain status need to be updated separately in case their hypervisor
supports this functionality.
The reason is also saved into domain state XML so if a domain is not
running (i.e., no state XML exists) the reason will be lost by libvirtd
restart. I think this is an acceptable limitation.
This has been present since the introduction of phypAttachDevice
in commit 444fd07a.
* src/phyp/phyp_driver.c (phypAttachDevice): Don't dereference
NULL.
virFDStreamClose used a mutex after it was freed, and failed
to destroy that mutex on its last use.
* src/fdstream.c (virFDStreamFree): Inline into sole caller...
(virFDStreamClose): ...to avoid use-after-free and leak.
Reported by Matthias Bolte.
Several vSphere API methods are called on global objects like the
FileManager, the PerformanceManager or the SearchIndex. The generator
input file allows to mark such methods and the generator generates
such method in a way that automatically handles marked parameter. This
is done by some special macros. Those were manually written and this
patch moves them to the generator.
One functionality change here is that we no longer force enable the event
timeout for every queued event, only enable it for the first event after
the queue has been flushed. This is how other drivers have already done it,
and I haven't encountered problems in practice.
v3:
Adjust for new virDomainEventStateNew argument
The same code for queueing, flushing, and deregistering events exists
in multiple drivers, which will soon use these common functions.
v2:
Adjust libvirt_private.syms
isDispatching bool fixes
NONNULL tagging
v3:
Add requireTimer parameter to virDomainEventStateNew
This structure will be used to unify lots of duplicated event handling code
across the state drivers.
v2:
Check for state == NULL in StateFree
Add NONNULL tagging
Use bool for isDispatching
Signed-off-by: Cole Robinson <crobinso@redhat.com>
