Il 05/03/2014 09:01, Zhanghailiang ha scritto:
Hi all:
Currently, we use cgroup(memory) to support memory QoS on KVM
platform, and use mlock on qemu to support memory reserved.
The mlock seems to be not appropriate.
Now qemu mlock memory in the main thread, which would lock
Il 06/03/2014 09:06, Zhanghailiang ha scritto:
Il 05/03/2014 09:01, Zhanghailiang ha scritto:
Hi all:
Currently, we use cgroup(memory) to support memory QoS on KVM
platform, and use mlock on qemu to support memory reserved.
The mlock seems to be not appropriate.
Now qemu mlock memory in
On 03/04/2014 03:58 AM, Michael Chapman wrote:
- As of commit 2ff4c137, all virGet*() functions in datatypes.c always
return pointers to new objects. Objects are not cached in a
per-connection hashtable.
- As of commit 46ec5f85, the conn.lock mutex does not need to be held
when
On Wed, Mar 05, 2014 at 05:53:48PM -0700, Eric Blake wrote:
On 03/05/2014 10:53 AM, Daniel P. Berrange wrote:
Extracting capabilities from QEMU takes a notable amount of time
when all QEMU binaries are installed. Each system emulator
needs about 200-300ms multiplied by 26 binaries == ~5-8
On Wed, Mar 05, 2014 at 02:52:49PM -0500, Tomoki Sekiyama wrote:
Hello,
This is patchset v2 to add FSFreeze/FSThaw API for custom disk snapshotting.
Patch 1 in your series hasn't arrived in my inbox, nor is it visible
in the mail archves...
Regards,
Daniel
--
|: http://berrange.com -o-
On Wed, Mar 05, 2014 at 09:40:34PM +0100, Guido Günther wrote:
Hi Daniel,
On Wed, Mar 05, 2014 at 05:53:53PM +, Daniel P. Berrange wrote:
[..snip..]
+/* Discard if cache is older that QEMU binary */
+/* XXX must also compare to libvirtd timestamp */
+if (sb.st_mtime
This patch is used to add --compress and [--compression-format] string to
virsh dump --memory-only. And virsh dump --memory-only is going be
implemented by new virDomainCoreDumpWithFormat API.
Signed-off-by: Qiao Nuohan qiaonuo...@cn.fujitsu.com
---
tools/virsh-domain.c | 45
dumping guest's memroy is introduced without compression supported, and this is
a freature regression of 'virsh dump --memory-only'. This patchset is used to
add support in libvirt side to make qemu dump guest's memory in kdump-compressed
format and please refer the following address to see
--memory-only option is introduced without compression supported. Therefore,
this is a freature regression of virsh dump. Now qemu has support dumping memory
in kdump-compressed format. This patch is adding new virDomainCoreDumpWithFormat
API, so that the format in which qemu dump domain's memory
This patch makes qemu driver supprot virDomainCoreDumpWithFormat API.
---
src/qemu/qemu_driver.c | 49
src/qemu/qemu_monitor.c | 7 ---
src/qemu/qemu_monitor.h | 3 ++-
src/qemu/qemu_monitor_json.c | 4 +++-
Amos Kong ak...@redhat.com writes:
vm_config_groups[] only contains part of the options which have
argument, and all options which have no argument aren't added
to vm_config_groups[]. Current query-command-line-options only
checks options from vm_config_groups[], so some options will
be
Libvirt tarball contains po/stamp-po file which prevents any po/*.gmo
file to be regenerated even if a corresponding po/*.po file is newer. By
removing the stamp-po file, all *.gmo files are properly updated if
required. This allows downstreams to provide patches that update
translations.
On 03/06/2014 04:06 AM, Jiri Denemark wrote:
Libvirt tarball contains po/stamp-po file which prevents any po/*.gmo
file to be regenerated even if a corresponding po/*.po file is newer. By
removing the stamp-po file, all *.gmo files are properly updated if
required. This allows downstreams to
On 03/06/2014 12:02 AM, Jason Helfman wrote:
On Sun, Mar 2, 2014 at 7:26 AM, Daniel Veillard veill...@redhat.com wrote:
As planned I have tagged libvirt-1.2.2 in git and it seem the git tree
finally updated despite the horrible network I get in china ATM, the
tarballs may finally arrive at
On Thu, Mar 06, 2014 at 06:20:37 -0700, Eric Blake wrote:
On 03/06/2014 04:06 AM, Jiri Denemark wrote:
Libvirt tarball contains po/stamp-po file which prevents any po/*.gmo
file to be regenerated even if a corresponding po/*.po file is newer. By
removing the stamp-po file, all *.gmo files
On 03/06/2014 06:37 AM, Jiri Denemark wrote:
On Thu, Mar 06, 2014 at 06:20:37 -0700, Eric Blake wrote:
On 03/06/2014 04:06 AM, Jiri Denemark wrote:
Libvirt tarball contains po/stamp-po file which prevents any po/*.gmo
file to be regenerated even if a corresponding po/*.po file is newer. By
Hi Team,
We are calling virDomainShutdown() API for shutting the VM. Sometime we see,
this API is failing because of 'unknown' reason. It took almost 15 secs to
bringdown the VM and at last failed. And the domain goes into shut off state.
You can find the time stamp in below log. Could you
On Wed, Mar 05, 2014 at 06:12:41AM -0700, Eric Blake wrote:
On 03/05/2014 03:37 AM, Martin Kletzander wrote:
On Tue, Mar 04, 2014 at 02:04:05PM -0700, Eric Blake wrote:
On 03/04/2014 07:15 AM, Martin Kletzander wrote:
When domain is started with setting that cannot be done, i.e. those
[To be squashed in the previous commit]
I've found two more places where shareSpecified should be checked
and added a test.
---
src/parallels/parallels_driver.c | 1 +
src/vmx/vmx.c | 2 +-
.../qemuxml2argv-cputune-zero-shares.args
Many of the domain xml format functions (including all of the device
format functions) had hard-coded spaces, which made for incorrect
indentation when those functions were called in a different context
(for example, commit 2122cf39 added interface XML into the document
provided to a network hook
On Thu, Mar 06, 2014 at 03:46:42PM +0100, Ján Tomko wrote:
[To be squashed in the previous commit]
I've found two more places where shareSpecified should be checked
and added a test.
And you can find another one in qemuBuildCommandLine(), see commit
45ad1adb4a5ae5ce46287c570e87abab6ffe62d6
These will freeze and thaw filesystems within guest. The APIs take flags
arguments which are currently not used, for future extensions.
Signed-off-by: Tomoki Sekiyama tomoki.sekiy...@hds.com
---
include/libvirt/libvirt.h.in |6
src/driver.h | 10 ++
src/libvirt.c
This was very simple, since the only place that had hardcoded
indentation was a few items in the network status xml.
---
src/conf/network_conf.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/conf/network_conf.c b/src/conf/network_conf.c
index bac0465..1d5781e
All leading spaces in domain snapshot xml format functions have been
replaced with appropriate calls to virBufferAdjustIndent(). This will
make it easier to call other similarly fixed format functions
(e.g. domain device format functions).
---
src/conf/snapshot_conf.c | 48
This file was using multiple virBuffers, inserting the contents of
buf3 into buf2, then inserting the contents of buf2 into buf1, rather
than the more conventional method of just passing around a single
virBufferPtr and streaming everything into that single buffer. This
was unnecessary, and also
Again completely mechanical, but a lot of lines.
---
src/conf/storage_conf.c| 174 +
src/conf/storage_encryption_conf.c | 6 +-
2 files changed, 105 insertions(+), 75 deletions(-)
diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c
These format functions needed the ability to be indented by an
arbitrary amount, but were written before the introduction of
virBufferAdjustIndent(). They instead used the much more clunky method
of adding a level arg to every format function, and padding with
spaces using the %*s printf format
These last files had such a small change count I just put them into a
single patch.
---
src/conf/cpu_conf.c | 11 ++-
src/conf/netdev_bandwidth_conf.c | 6 --
src/conf/netdev_vlan_conf.c | 6 --
src/conf/netdev_vport_profile_conf.c | 6 --
Completely mechanical changes, but there were a lot of lines so I made
it a separate patch.
---
src/conf/node_device_conf.c | 207
1 file changed, 112 insertions(+), 95 deletions(-)
diff --git a/src/conf/node_device_conf.c
There were a lot of changes here, but all very mechanical. For some
reason, the virBufferPtr had been named xml instead of buf in this
file, so since the indentation changing touched almost every line
using the buffer, I took this chance to change its name for buf for
consistency with every other
These will freeze and thaw filesystems within guest. The APIs take flags
arguments which are currently not used, for future extensions.
Signed-off-by: Tomoki Sekiyama tomoki.sekiy...@hds.com
---
include/libvirt/libvirt.h.in |6
src/driver.h | 10 ++
src/libvirt.c
On 3/6/14 4:32 , Daniel P. Berrange berra...@redhat.com wrote:
On Wed, Mar 05, 2014 at 02:52:49PM -0500, Tomoki Sekiyama wrote:
Hello,
This is patchset v2 to add FSFreeze/FSThaw API for custom disk
snapshotting.
Patch 1 in your series hasn't arrived in my inbox, nor is it visible
in the mail
Starting with libvirt 0.9.0+, hook scripts can be called from several new
locations. These locations must also be reflected in the expected logs of
the hook tests.
The final test in 052-domain-hook.t intentionally produces a failed start,
which should show the first stage in the startup hook
This test completes successfully, but results in a return code of 141 due to
a broken pipe when restarting libvirtd. This patch just masks the SIGPIPE
and undefines $tck to avoid the 141 return code. If there is a way to
reestablish the tck connection after the restart, that would be a better
On 03/06/2014 08:24 AM, Laine Stump wrote:
Many of the domain xml format functions (including all of the device
format functions) had hard-coded spaces, which made for incorrect
indentation when those functions were called in a different context
(for example, commit 2122cf39 added interface
On 03/06/2014 08:24 AM, Laine Stump wrote:
All leading spaces in domain snapshot xml format functions have been
replaced with appropriate calls to virBufferAdjustIndent(). This will
make it easier to call other similarly fixed format functions
(e.g. domain device format functions).
---
When matching capabilities of a guest, if multiple domain types exist (for
example, 'qemu' and 'kvm') the order in which they are returned can change.
To avoid unpredictable test results, this patch prefers kvm if that domain
type exists. If not, the behavior matches what existed before, and the
On 03/06/2014 08:24 AM, Laine Stump wrote:
This was very simple, since the only place that had hardcoded
indentation was a few items in the network status xml.
---
src/conf/network_conf.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
ACK.
--
Eric Blake eblake redhat com
On 03/06/2014 08:24 AM, Laine Stump wrote:
These format functions needed the ability to be indented by an
arbitrary amount, but were written before the introduction of
virBufferAdjustIndent(). They instead used the much more clunky method
of adding a level arg to every format function, and
On 03/06/2014 08:24 AM, Laine Stump wrote:
This file was using multiple virBuffers, inserting the contents of
buf3 into buf2, then inserting the contents of buf2 into buf1, rather
than the more conventional method of just passing around a single
virBufferPtr and streaming everything into that
On 03/06/2014 08:24 AM, Laine Stump wrote:
There were a lot of changes here, but all very mechanical. For some
reason, the virBufferPtr had been named xml instead of buf in this
file, so since the indentation changing touched almost every line
using the buffer, I took this chance to change its
On 03/06/2014 08:24 AM, Laine Stump wrote:
Completely mechanical changes, but there were a lot of lines so I made
it a separate patch.
---
src/conf/node_device_conf.c | 207
1 file changed, 112 insertions(+), 95 deletions(-)
ACK
--
Eric
On 03/06/2014 08:24 AM, Laine Stump wrote:
Again completely mechanical, but a lot of lines.
---
src/conf/storage_conf.c| 174
+
src/conf/storage_encryption_conf.c | 6 +-
2 files changed, 105 insertions(+), 75 deletions(-)
ACK
--
Eric
On 03/06/2014 08:24 AM, Laine Stump wrote:
These last files had such a small change count I just put them into a
single patch.
---
src/conf/cpu_conf.c | 11 ++-
src/conf/netdev_bandwidth_conf.c | 6 --
src/conf/netdev_vlan_conf.c | 6 --
On 03/06/2014 08:24 AM, Laine Stump wrote:
[0/9 in the subject line]
Many of the domain xml format functions (including all of the device
format functions) had hard-coded spaces, which made for incorrect
indentation when those functions were called in a different context
(for example, commit
On Thu, Mar 06, 2014 at 06:57:14 -0700, Eric Blake wrote:
On 03/06/2014 06:37 AM, Jiri Denemark wrote:
On Thu, Mar 06, 2014 at 06:20:37 -0700, Eric Blake wrote:
On 03/06/2014 04:06 AM, Jiri Denemark wrote:
Libvirt tarball contains po/stamp-po file which prevents any po/*.gmo
file to be
On 03/05/2014 07:36 PM, Amos Kong wrote:
vm_config_groups[] only contains part of the options which have
argument, and all options which have no argument aren't added
to vm_config_groups[]. Current query-command-line-options only
checks options from vm_config_groups[], so some options will
be
Reload libvirtd after hook testing has completed. Otherwise, libvirtd
is still expecting hook test scripts to exist.
---
scripts/hooks/052-domain-hook.t | 4
1 file changed, 4 insertions(+)
diff --git a/scripts/hooks/052-domain-hook.t b/scripts/hooks/052-domain-hook.t
index
On Thu, Mar 6, 2014 at 5:22 AM, Eric Blake ebl...@redhat.com wrote:
On 03/06/2014 12:02 AM, Jason Helfman wrote:
On Sun, Mar 2, 2014 at 7:26 AM, Daniel Veillard veill...@redhat.com
wrote:
As planned I have tagged libvirt-1.2.2 in git and it seem the git tree
finally updated despite the
While investigating https://bugzilla.redhat.com/show_bug.cgi?id=1061827
I noticed that we pass user input unscathed for block-pull, but
always pass a canonical absolute name through for block-commit.
[Note that we probably _ought_ to validate that the user's request
for block-pull actually matches
-Original Message-
From: kvm-ow...@vger.kernel.org [mailto:kvm-ow...@vger.kernel.org]
On Behalf Of Stefan Hajnoczi
Sent: Friday, February 14, 2014 7:58 AM
To: Cedric Bosdonnat
Cc: Jan Kiszka; qemu-devel; kvm; libvir-list@redhat.com; Kevin Wolf; Peter
Maydell; Peter Crosthwaite; Max
bug address: https://bugzilla.redhat.com/show_bug.cgi?id=1066801
I can reproduce it at master/v1.2.1-maint/v1.1.4-maint/v1.1.3.4 release
branches everytime,
but cann't reproduce it at v1.2.1,so I believe this bug is imported by the
maintain patches after v1.2.1 release.
Steps to Reproduce:
I installed libvirt 1.2 tarball and then I use pip install libvirt-python
inside virtual env.. it worked..
thanks
-Sijo
On Thu, Mar 6, 2014 at 12:02 PM, Roman Bogorodskiy bogorods...@gmail.comwrote:
Sijo Jose wrote:
@Michel
using virtualenv we can create an isolated environment in
vm_config_groups[] only contains part of the options which have
parameters, and all options which have no parameter aren't added
to vm_config_groups[]. Current query-command-line-options only
checks options from vm_config_groups[], so some options will
be lost.
We have macro in qemu-options.hx to
This patchset fixed some issues of query-command-line-options:
* some new options that haven't argument can't be queried. (eg: -enable-fips)
* some legacy options that have argument can't be queried. (eg: -vnc display)
More discussion:
http://marc.info/?l=qemu-develm=139081830416684w=2
V2:
On Thu, Mar 06, 2014 at 02:23:15PM -0700, Eric Blake wrote:
On 03/05/2014 07:36 PM, Amos Kong wrote:
vm_config_groups[] only contains part of the options which have
argument, and all options which have no argument aren't added
to vm_config_groups[]. Current query-command-line-options only
Signed-off-by: Amos Kong ak...@redhat.com
Reviewed-by: Eric Blake ebl...@redhat.com
---
util/qemu-config.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/util/qemu-config.c b/util/qemu-config.c
index f610101..d2facfd 100644
--- a/util/qemu-config.c
+++
57 matches
Mail list logo