On Tue, Jan 26, 2016 at 01:31:25PM +0300, Maxim Nestratov wrote:
26.01.2016 12:29, Martin Kletzander пишет:
On Mon, Jan 25, 2016 at 12:16:12PM +0300, Maxim Nestratov wrote:
A pretty nasty deadlock occurs while trying to rename a VM in parallel
with virDomainObjListNumOfDomains.
The short
On Mon, 25 Jan 2016, John Ferlan wrote:
Since this has been sitting for a bit, I'll poke the bear, but also CC
Peter since he reviewed V1 and had some specific concerns...
Thanks for picking up this review.
The concern before was how the code would handle the user passing a
different,
I'm looking forward to be reviewed)
Sincerely yours,
Patch Series
On 13.01.2016 14:01, Nikolay Shirokovskiy wrote:
> Current libvirt + qemu pair lacks secure migrations in case of
> VMs with non-shared disks. The only option to migrate securely
> natively is to use
On 01/26/2016 04:24 PM, Peter Krempa wrote:
On Tue, Jan 26, 2016 at 10:25:05 +0800, Luyao Huang wrote:
After commit 57177f1, the cpu-stats command format change to:
CPU0:
cpu_time 14401.507878990 seconds
vcpu_time14378732785511
vcpu_time is not user friendly. After
In commit 7039bb3c we have removed code that saves uuid to vzDomObj.uuid
So this field is no longer needed.
---
src/vz/vz_sdk.c | 5 -
src/vz/vz_utils.h | 1 -
2 files changed, 6 deletions(-)
diff --git a/src/vz/vz_sdk.c b/src/vz/vz_sdk.c
index d610979..5e4c4ac 100644
--- a/src/vz/vz_sdk.c
The VIR_DOMAIN_STATS_VCPU flag to virDomainListGetStats
enables reporting of stats about vCPUs. Currently we
only report the cumulative CPU running time and the
execution state.
This adds reporting of the wait time - time the vCPU
wants to run, but the host schedular has something else
running
On 25.01.2016 10:16, Maxim Nestratov wrote:
> A pretty nasty deadlock occurs while trying to rename a VM in parallel
> with virDomainObjListNumOfDomains.
> The short description of the problem is as follows:
>
> Thread #1:
>
> qemuDomainRename:
> --> aquires domain lock by
On Mon, Jan 25, 2016 at 12:16:12PM +0300, Maxim Nestratov wrote:
A pretty nasty deadlock occurs while trying to rename a VM in parallel
with virDomainObjListNumOfDomains.
The short description of the problem is as follows:
Thread #1:
qemuDomainRename:
--> aquires domain lock by
On Mon, Jan 25, 2016 at 09:00:27PM -0500, Laine Stump wrote:
> >> +retries = MACVLAN_MAX_ID + 1;
> >> +while (!ifnameCreated && retries) {
> >> virMutexLock();
> >> -for (c = 0; c < 8192; c++) {
> >> -snprintf(ifname, sizeof(ifname), pattern, c);
> >> -
18.01.2016 13:08, Alexander Burluka пишет:
This functions setups per-domain cpu bandwidth parameters
Signed-off-by: Alexander Burluka
---
src/qemu/qemu_cgroup.c | 54 +
src/qemu/qemu_cgroup.h | 1 +
On 25.01.2016 18:28, Stefan Hajnoczi wrote:
> The QEMU wiki page for Google Summer of Code 2016 is now available here:
>
> http://qemu-project.org/Google_Summer_of_Code_2016
>
> QEMU will apply for Google Summer of Code 2016 (https://g.co/gsoc/).
> If QEMU is accepted there will be funding for
26.01.2016 12:29, Martin Kletzander пишет:
On Mon, Jan 25, 2016 at 12:16:12PM +0300, Maxim Nestratov wrote:
A pretty nasty deadlock occurs while trying to rename a VM in parallel
with virDomainObjListNumOfDomains.
The short description of the problem is as follows:
Thread #1:
On Tue, Jan 26, 2016 at 12:59:00 +0300, Alexander Burluka wrote:
> On 01/25/2016 04:16 PM, Daniel P. Berrange wrote:
> > On Mon, Jan 25, 2016 at 02:55:30PM +0300, Alexander Burluka wrote:
> >> For example, we want to store suspended state of VM.
> >> I'm aware that some other careless application
The migration would fail if the checks in virDomainDefPostParseMemory() fail
after ABI update. The ABI update normally done during the guest start and not
during define. If someone is using --persistent with the virsh
migrate it should be treated equivalent to domain define and not start. The
Please drop this patch. I sent it too early. I think the existing migration
behavior with --persistent is expected.
Thanks,
Shiva
On Tue, Jan 26, 2016 at 1:31 PM, Shivaprasad G Bhat <
shivaprasadb...@gmail.com> wrote:
> The migration would fail if the checks in virDomainDefPostParseMemory()
>
On Tue, Jan 26, 2016 at 10:25:05 +0800, Luyao Huang wrote:
> After commit 57177f1, the cpu-stats command format change to:
>
> CPU0:
> cpu_time 14401.507878990 seconds
> vcpu_time14378732785511
>
> vcpu_time is not user friendly. After this patch, it will
> change back:
>
On Tue, Jan 26, 2016 at 13:26:32 +0100, Michal Privoznik wrote:
> On 25.01.2016 10:16, Maxim Nestratov wrote:
...
> >
> > virCheckFlags(0, ret);
> >
> > -if (!(vm = qemuDomObjFromDomain(dom)))
> > +virObjectLock(driver->domains);
>
> This is rather ugly. While driver->domains is
On Fri, Jan 22, 2016 at 03:56:08PM +, Daniel P. Berrange wrote:
> We have had virtlockd available for a long time now but
> have always defaulted to the 'nop' lock driver which does
> no locking. This gives users an unsafe deployment by
> default unless they know to turn on lockd.
Does the
On Tue, Jan 26, 2016 at 02:14:23PM +0100, Jiri Denemark wrote:
On Tue, Jan 26, 2016 at 13:26:32 +0100, Michal Privoznik wrote:
On 25.01.2016 10:16, Maxim Nestratov wrote:
...
>
> virCheckFlags(0, ret);
>
> -if (!(vm = qemuDomObjFromDomain(dom)))
> +virObjectLock(driver->domains);
On Fri, Jan 22, 2016 at 05:21:03PM -0500, John Ferlan wrote:
> When commit id '82c1740a' made changes to the output format (changing from
> using a ',' separator to '#'), the examples in the lvs output from the
> comments weren't changed.
>
> Additionally, the two new fields added ('segtype' and
On Tue, Jan 26, 2016 at 02:28:56PM +0100, Ján Tomko wrote:
> On Fri, Jan 22, 2016 at 03:56:08PM +, Daniel P. Berrange wrote:
> > We have had virtlockd available for a long time now but
> > have always defaulted to the 'nop' lock driver which does
> > no locking. This gives users an unsafe
26.01.2016 16:33, Martin Kletzander пишет:
On Tue, Jan 26, 2016 at 02:14:23PM +0100, Jiri Denemark wrote:
On Tue, Jan 26, 2016 at 13:26:32 +0100, Michal Privoznik wrote:
On 25.01.2016 10:16, Maxim Nestratov wrote:
...
[snip]
What if, instead of introducing bunch of Locked() functions we
On Fri, Jan 22, 2016 at 03:56:08PM +, Daniel P. Berrange wrote:
> We have had virtlockd available for a long time now but
> have always defaulted to the 'nop' lock driver which does
> no locking. This gives users an unsafe deployment by
> default unless they know to turn on lockd. virtlockd
On Tue, Jan 26, 2016 at 11:20:59AM +, Daniel P. Berrange wrote:
> The VIR_DOMAIN_STATS_VCPU flag to virDomainListGetStats
> enables reporting of stats about vCPUs. Currently we
> only report the cumulative CPU running time and the
> execution state.
>
> This adds reporting of the wait time -
Hello,
It would seem this mailing list is pretty busy - however I have not seen any
responses to my email?
Is there anybody who can take a peek at this and give me some suggestions?
Thanks,
Maarten Jacobs
From: maarten...@hotmail.com
To: libvir-list@redhat.com
Date: Sun, 24 Jan 2016 12:57:51
On Tue, Jan 26, 2016 at 01:47:02PM +, Daniel P. Berrange wrote:
> On Tue, Jan 26, 2016 at 02:28:56PM +0100, Ján Tomko wrote:
> > On Fri, Jan 22, 2016 at 03:56:08PM +, Daniel P. Berrange wrote:
> > > We have had virtlockd available for a long time now but
> > > have always defaulted to the
On the host when we start a container, it will be
placed in a cgroup path of
/machine.slice/machine-lxc\x2ddemo.scope
under /sys/fs/cgroup/*
Inside the containers' namespace we need to setup
/sys/fs/cgroup mounts, and currently will bind
mount /machine.slice/machine-lxc\x2ddemo.scope on
the
On Sun, Jan 24, 2016 at 12:57:51PM -0500, Maarten Jacobs wrote:
> Hello,
> I have been trying to get libvirt compiled on Cygwin but I have run into a
> rather perplexing issue.
> I am running a 32-bit installation of Cygwin on Windows 10:
> CYGWIN_NT-10.0-WOW dell-desktop 2.4.0(0.293/5/3)
On Fri, Jan 22, 2016 at 05:21:04PM -0500, John Ferlan wrote:
> Create a helper routine in order to parse the 'device' string contained
> within the generated 'lvs' output string.
>
> A future patch would then be able to avoid the code more cleanly
>
> Signed-off-by: John Ferlan
On Fri, Jan 22, 2016 at 05:21:06PM -0500, John Ferlan wrote:
> From: Joe Harvell
>
> Since our 'devices' parsing logic now will use the 'nextents' (or
> lvs 'stripes' output) to decide whether or not to parse the field,
> use the regex of "(\\S*)" (e.g. zero or more)
ping
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
On 26.01.2016 15:37, Daniel P. Berrange wrote:
> On the host when we start a container, it will be
> placed in a cgroup path of
>
>/machine.slice/machine-lxc\x2ddemo.scope
>
> under /sys/fs/cgroup/*
>
> Inside the containers' namespace we need to setup
> /sys/fs/cgroup mounts, and currently
On 12.01.2016 13:06, Pavel Hrdina wrote:
> Ok, so consider this squashed in, I've forgot to regenerate the tests after
> rebase to current HEAD and there seems to be new tests.
>
> ---
>
Not only that. This needs to be squashed in too:
diff --git
On 12.01.2016 12:59, Pavel Hrdina wrote:
> Some of the patches from v1 series are already pushed, only those 4 left for
> review.
>
> Pavel Hrdina (4):
> tests: use virtTestDifferenceFull in tests where we have output file
> tests: add some missing tests to qemuxml2xmltest
> device: cleanup
On 12.01.2016 12:59, Pavel Hrdina wrote:
> The current code was a little bit odd. At first we've removed all
> possible implicit input devices from domain definition to add them later
> back if there was any graphics device defined while parsing XML
> description. That's not all, while formating
On Wed, Jan 20, 2016 at 16:47:56 +0200, Alberto Garcia wrote:
> Hi Peter,
Hi Alberto,
>
> I'm the current maintainer of Quorum in QEMU and I'd like to try to
> answer some of your comments.
>
> On Fri, Jan 08, 2016 at 06:20:04PM +0100, Peter Krempa wrote:
>
> > So I have a few
Hi,
I'm using libvirt (1.2.12) with qemu (2.2.0) in the context of OpenStack.
If I live-migrate a guest with virtio network interfaces, I see a ~1200msec
delay in processing the network packets, and several hundred of them get
dropped. I get the dropped packets, but I'm not sure why the
On 01/22/2016 02:11 PM, Cole Robinson wrote:
On 01/22/2016 02:09 PM, Cole Robinson wrote:
Nowadays we only support qemu 0.12.0+ which provides QEMU_CAPS_DEVICE,
so this is all dead code.
---
src/qemu/qemu_hotplug.c | 480 +++-
1 file changed, 144
In dc576025c360 we renamed virCgroupIsolateMount function to
virCgroupBindMount. However, we forgot about one occurrence in
section of the code which provides stubs for platforms without
support for CGroups like *BSD for instance.
Signed-off-by: Michal Privoznik
---
Pushed
On Tue, Jan 26, 2016 at 10:41:12AM -0600, Chris Friesen wrote:
> Hi,
>
> I'm using libvirt (1.2.12) with qemu (2.2.0) in the context of OpenStack.
>
> If I live-migrate a guest with virtio network interfaces, I see a ~1200msec
> delay in processing the network packets, and several hundred of
On 26/01/2016 17:41, Chris Friesen wrote:
> I'm using libvirt (1.2.12) with qemu (2.2.0) in the context of OpenStack.
>
> If I live-migrate a guest with virtio network interfaces, I see a
> ~1200msec delay in processing the network packets, and several hundred
> of them get dropped. I get the
On 01/26/2016 10:50 AM, Paolo Bonzini wrote:
On 26/01/2016 17:41, Chris Friesen wrote:
I'm using libvirt (1.2.12) with qemu (2.2.0) in the context of OpenStack.
If I live-migrate a guest with virtio network interfaces, I see a
~1200msec delay in processing the network packets, and several
On 26/01/2016 18:21, Chris Friesen wrote:
>>>
>>> My question is, why doesn't qemu continue processing virtio packets
>>> while the dirty page scanning and memory transfer over the network is
>>> proceeding?
>>
>> QEMU (or vhost) _are_ processing virtio traffic, because otherwise you'd
>> have
On 01/26/2016 11:31 AM, Paolo Bonzini wrote:
On 26/01/2016 18:21, Chris Friesen wrote:
My question is, why doesn't qemu continue processing virtio packets
while the dirty page scanning and memory transfer over the network is
proceeding?
QEMU (or vhost) _are_ processing virtio traffic,
On 26/01/2016 18:49, Chris Friesen wrote:
>>
>> That doesn't exclude a bug somewhere in net/ code. It doesn't pinpoint
>> it to QEMU or vhost-net.
>>
>> In any case, what I would do is to use tracing at all levels (guest
>> kernel, QEMU, host kernel) for packet rx and tx, and find out at which
libvirtd crashes on free()ing portData for an open vswitch port if that port
was deleted. To reproduce:
ovs-vsctl del-port vnet0
virsh migrate --live kvm1 qemu+ssh://dstHost/system
Error message:
libvirtd: *** Error in `/usr/sbin/libvirtd': free(): invalid pointer:
0x03ff90001e20 ***
The
On 01/26/2016 10:45 AM, Daniel P. Berrange wrote:
On Tue, Jan 26, 2016 at 10:41:12AM -0600, Chris Friesen wrote:
My question is, why doesn't qemu continue processing virtio packets while
the dirty page scanning and memory transfer over the network is proceeding?
The
Need something here...
Perhaps to the effect of instead of inlining most of
virHostdevOnlyReattachPCIDevice, call virHostdevOnlyReattachPCIDevice.
One extra "call" or check being made by doing this is the
virPCIDeviceGetStubDriver check and virPCIDeviceWaitForCleanup. Are
there any "timing"
w/r/t: your [0/7] from initial series...
As much as you don't want to keep living Groundhog Day - resolution of
bugs like this are job security :-)...
The subtleties of hostdev device Attach, Reattach, ReAttach, and Detach
are such a tangle morass of shinola (ref: Steve Martin in "The
On 01/25/2016 11:20 AM, Andrea Bolognani wrote:
> This ensures the behavior for reattach is consistent, no matter how it
> was triggered (eg. 'virsh nodedev-reattach', 'virsh detach-device' or
> shutdown of a domain that is configured to use hostdevs).
Similar to 2/9 - using
On 01/25/2016 11:20 AM, Andrea Bolognani wrote:
> This function mirrors virHostdevOnlyDetachPCIDevice().
>
> The handling of active and inactive devices is updated and made more
> explicit, which means virHostdevPreparePCIDevices() has to be
> updated as well.
> ---
> src/util/virhostdev.c |
On 01/25/2016 11:21 AM, Andrea Bolognani wrote:
> This ensures the behavior for detach is consistent, no matter how it
> was triggered (eg. 'virsh nodedev-detach', 'virsh attach-device' or
> startup of a domain that is configured to use hostdevs).
> ---
> src/util/virhostdev.c | 28
On 01/25/2016 11:21 AM, Andrea Bolognani wrote:
> The name is confusing, and the only use left is in a test case.
> ---
> src/libvirt_private.syms | 1 -
> src/util/virpci.c| 8
> src/util/virpci.h| 1 -
> tests/virpcitest.c | 5 -
> 4 files changed, 4
On 01/25/2016 11:21 AM, Andrea Bolognani wrote:
> Unbinding a PCI device from the stub driver can require several steps,
> and it can be useful for debugging to be able to trace which of these
> steps are performed and which are skipped for each device.
> ---
> src/util/virpci.c | 22
On 01/25/2016 11:21 AM, Andrea Bolognani wrote:
> Some of the comments are no longer accurate after the recent changes,
> others have been expanded and hopefully improved.
>
> The initial summary of the operations has been removed so that two
> separate edits are not needed when making changes.
On 01/25/2016 11:21 AM, Andrea Bolognani wrote:
> Mostly labels names and whitespace.
>
> No functional changes.
> ---
> src/util/virhostdev.c | 23 ++-
> 1 file changed, 14 insertions(+), 9 deletions(-)
>
Seems reasonable, ACK
John
--
libvir-list mailing list
If virTimeMillisNow() fails, the SIGINT action must be reset back to
its previous state.
Signed-off-by: Michael Chapman
---
tools/virsh-domain.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c
index
When waiting for a block job, the various statuses (COMPLETED, READY,
CANCELED, etc.) should all be treated consistently by having the loop be
exited with "break". Use "goto cleanup" for the error cases only, when
no block job status is available.
Signed-off-by: Michael Chapman
v2 discussion at:
http://www.redhat.com/archives/libvir-list/2016-January/msg01063.html
As requested I've split this into separate commits. All of the comment
nitpicks have been applied.
With regards to using "break" or "goto cleanup" to exit the loop, I decided
to use "goto" on all the error
There is no need to call virshPrintJobProgress() unless the block job's
cur or end cursors have changed since the last iteration.
Signed-off-by: Michael Chapman
---
tools/virsh-domain.c | 7 +--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git
After a block job hits 100%, we only need to apply a timeout waiting for
a block job event if exactly one of the BLOCK_JOB or BLOCK_JOB_2
callbacks were able to be registered.
If neither callback could be registered, there's clearly no need for a
timeout.
If both callbacks were registered, then
Commit 370608b added new functions:
- virNetDevMacVLanReleaseName,
- virNetDevMacVLanReserveName.
Add stubbed versions of them to fix build without macvtap.
---
src/util/virnetdevmacvlan.c | 17 +
1 file changed, 17 insertions(+)
diff --git a/src/util/virnetdevmacvlan.c
On 01/25/2016 04:16 PM, Daniel P. Berrange wrote:
On Mon, Jan 25, 2016 at 02:55:30PM +0300, Alexander Burluka wrote:
For example, we want to store suspended state of VM.
I'm aware that some other careless application dealing with libvirt
may erase metadata section and info about additional
63 matches
Mail list logo