> No, this boot= option is deprecated too. AFAIK boot=off does (and always
> did) nothing. boot=on tells qemu to boot from the disk using extboot option
> rom. This was needed to boot from virtio disks. Now SeaBIOS can boot from
> virtio disk natively, so extboot no longer needed. But due to the w
> > Another possibility to disable network boot would be to avoid loading
> > the pxe-XXX.rom network boot ROMs. Or is that a bad idea?
>
> Just add the romfile property to your nic with an empty string, i.e.
>
> qemu $otheropts -device e1000,romfile=,mac=...
Thanks, I can use that a sworkaround
> No, this boot= option is deprecated too. AFAIK boot=off does (and always
> did) nothing. boot=on tells qemu to boot from the disk using extboot option
> rom. This was needed to boot from virtio disks. Now SeaBIOS can boot from
> virtio disk natively, so extboot no longer needed. But due to the w
> Agree. noboot sounds optimal, but also requires more codding that other
> options (sigh, isn't it always this way?). But how do we have it for disks?
-drive [file=file][,if=type][,bus=n][,unit=m][,media=d][,index=i]
[,cyls=c,heads=h,secs=s[,trans=t]][,snapshot=on|off]
[,cache=write
> We can change BIOS to only boot from devices that has bootindex, but then you
> will have to always specify it for all/most devices, or we can add noboot
> device
> property, but that will require changes on qemu side too.
I guess 'noboot' is the way to go. We already have that for disks.
- Di
> > Also, I think the behavior was different with earlier versions.
> Yes, it was. The behaviour changed when bootindex was introduced. I think it
> should be easy to switch it back to what it was for -boot option, but -boot
> is/should be deprecated in favor of bootindex anyway.
> Implementing opt
> Then it is expected (as in "this is how code works currently"). Why would you
> want to disable network boot if other method failed?
If I start kvm with:
# kvm -boot order=cad
The it tries to boot: cdrom, net, disk, floppy (dnca)
So the boot order is completely wrong!
But the order is corre
> Then it is expected (as in "this is how code works currently"). Why would you
> want to disable network boot if other method failed?
Because I do not want to start/install a new VM only because I have some other
error. Also, I think the behavior was different with earlier versions.
For example
> Does it tries it only when c & d are not bootable or always?
Only if c &d are not bootable.
- Dietmar
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
If I use '-boot order=cd' the VM still tries to boot from network. Is that
expected? If so, how can I disable network boot?
- Dietmar
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel
gt; -Original Message-
> From: Nikola Ciprich [mailto:extmaill...@linuxbox.cz]
> Sent: Montag, 21. März 2011 21:48
> To: Dietmar Maurer
> Cc: kvm@vger.kernel.org; nikola.cipr...@linuxbox.cz
> Subject: Re: qemu-kvm-0.14.0 msix_mask_notifier failed.
>
> Hello Dietmar,
&g
Seems the bug is related to EVENTFD support.
int event_notifier_init(EventNotifier *e, int active)
That simply return -ENOSYS on older glibc (< 2.8)
I can define the syscall manually, but I wonder if there is a better way?
- Dietmar
> I get this error when I run new kvm 0.14.0 on kernel 2.6.
Hi all,
I get this error when I run new kvm 0.14.0 on kernel 2.6.35 (vhost=on):
# kvm -netdev type=tap,id=n0,vhost=on -device virtio-net-pci,netdev=n0 -cdrom
ubuntu-10.10-desktop-amd64.iso
kvm: /home/code/qemu-kvm/hw/msix.c:639: msix_unset_mask_notifier: Assertion
`dev->msix_mask_notifier' fai
> Right. So the kvm-kmod replacement is also fine.
>
> Dietmar, if you could confirm that your issue is fixed,
Yes, it is fixed.
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.or
> > The ubuntu code puts some barrier around the read.
> >
> > http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-
> hardy.git;a=commitdiff;h=198b348d96c9769153e72ca2461f8d841ddff1cc
> >
> > You simply override this with your own code - without barrier. Do you
> think this is correct?
>
> Unless I messe
Hi Jan,
The ubuntu code puts some barrier around the read.
http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=commitdiff;h=198b348d96c9769153e72ca2461f8d841ddff1cc
You simply override this with your own code - without barrier. Do you think
this is correct?
- Dietmar
--
To unsubscribe f
t: Montag, 16. November 2009 11:52
> To: Dietmar Maurer
> Cc: kvm
> Subject: Re: problem wit svm_get_msr on kvm-kmod-2.6.31.6
>
> On 11/16/2009 12:46 PM, Dietmar Maurer wrote:
> >> Nothing changed between these two versions to warrant this.
> >>
> > Oh,
> Nothing changed between these two versions to warrant this.
Oh, sorry - the one which works is kvm-kmod-2.6.30.1
> Can you post a disassembly of svm_get_msr() around the offending
> address?
Please can you tell me how to do that?
> Did you change qemu-kvm as well?
no, same qemu-kvm version (
Hi all,
We are testing kvm-kmod-2.6.31.6, and several people reported problems with AMD
cpus:
Nov 14 21:17:59 bigproxmox kernel: Pid: 3616, comm: kvm Not tainted
2.6.24-9-pve #1 ovz005
Nov 14 21:17:59 bigproxmox kernel: RIP: 0010:[]
[] :kvm_amd:svm_get_msr+0x146/0x300
...
see http://www.prox
> >> Locally, you could
> >> simply comment out the double definition for now.
> >
> > Yes, that will work. Many thanks for the fast help!
> >
>
> Is native_read_tsc the only conflict with hardy's kernel?
yes.
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a mes
> Do you know if there is anything to tell this ubuntu-patched kernel
> apart from a vanilla version? If so, we could adjust the check in
> x86/external-module-compat.h to avoid this breakage.
no, not really
> Locally, you could
> simply comment out the double definition for now.
Yes, that will
twoch, 28. Oktober 2009 12:55
> To: Dietmar Maurer
> Cc: kvm
> Subject: Re: [ANNOUNCE] kvm-kmod-2.6.31.5
>
> Dietmar Maurer wrote:
> >> This release has been tested on x86 down to host kernel 2.6.27 and
> >> builds down to 2.6.24. Building against older kernels is e
> This release has been tested on x86 down to host kernel 2.6.27 and
> builds down to 2.6.24. Building against older kernels is expected to be
> broken, but if anyone provides patches to fix it, I'm open to merge
> them.
I am still not able to build with our 2.6.24 kernel.
./sync -v kvm-kmod-2.6
> >> Do you support multiple guests accessing the same image?
> >
> > A VM image can be attached to any VMs but one VM at a time; multiple
> > running VMs cannot access to the same VM image.
I guess this is a problem when you want to do live migrations?
- Dietmar
> Also, on _loaded_ systems, I noticed creating/removing logical volumes
> can take really long (several minutes); where allocating a file of a
> given size would just take a fraction of that.
Allocating a file takes much longer, unless you use a 'sparse' file.
- Dietmar
--
To unsubscribe from
> Anyways, I do not know JGroups - maybe that 'reliable multicast' solves
> all network problems somehow - Is there any documentation about how
> they do it?
OK, found the papers on their web site - quite interesting too.
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the bo
Another suggestion: use LVM instead of btrfs (to get better performance)
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
> We use JGroups (Java library) for reliable multicast communication in
> our cluster manager daemon.
I doubt that there is something like 'reliable multicast' - you will run into
many problems when you try to handle errors.
> We don't worry about the performance much
> since the cluster manager
> -Original Message-
> From: kvm-ow...@vger.kernel.org [mailto:kvm-ow...@vger.kernel.org] On
> Behalf Of Wolfgang Mauerer
> Sent: Mittwoch, 21. Oktober 2009 16:21
> To: Dietmar Maurer
> Cc: kvm@vger.kernel.org; Kiszka, Jan
> Subject: Re: [PATCH 2/2] kvm-kmod: Docum
> + Before the kvm module can be built, the linux submodule must be
> initialised
> + and populated. The required sequence of commands is
> +
> + git submodule init
> + git submodule update
> + ./configure
> + make sync
> + make
> +
> + Notice that you can also specify an existing Linux tre
Quite interesting. But would it be possible to use corosync for the cluster
communication? The point is that we need corosync anyways for pacemaker, it is
written in C (high performance) and seem to implement the feature you need?
> -Original Message-
> From: kvm-ow...@vger.kernel.org [m
> > What is the reasoning behind such short downtimes? Are there any
> application that will fail with longer downtimes (let say 1s)?
> >
> > Note: on a 1Gbit/s net you can transfer only 10MB within 100ms
>
> which accounts for more than 2 thousand pages, which sounds like enough
> for a first pas
> > The default downtime is set to 30ms. This value triggers the
> convergence problem quite often. Maybe a longer default is more
> reasonable.
> What do you feel about 100 ms?
What is the reasoning behind such short downtimes? Are there any application
that will fail with longer downtimes (let
> > 'bandwidth' is something that changes dynamically (or by user
> settings), so why don't we simply abort after some amount of
> transferred memory (constant * memory size). This can be implemented by
> the management application without problems, although it's much easier
> inside kvm.
> >
> Eas
> We used to have a heuristic that said 'if an iteration transfers more
> pages than the previous iteration, we've stopped converging'. Why
> wouldn't that work?
I agree that this is the 'right' approach - but it is just too difficult to
detect that we are not 'converging', and it does not set a
> -Original Message-
> From: Avi Kivity [mailto:a...@redhat.com]
> Sent: Montag, 05. Oktober 2009 16:06
> To: Dietmar Maurer
> Cc: Glauber Costa; Anthony Liguori; kvm
> Subject: Re: migrate_set_downtime bug
>
> On 10/05/2009 04:01 PM, Dietmar Maurer wrote
> Heuristics like number of pages, maybe. But since we don't export
> iteration information, we can't expect management tools to stop the
> guest if migration doesn't converge.
>
> I suppose it could issue a 'stop' after some amount of time (constant *
> memory size / bandwidth).
'bandwidth' is
> On 10/05/2009 04:01 PM, Dietmar Maurer wrote:
> >> We used to have a heuristic that said 'if an iteration transfers
> more
> >> pages than the previous iteration, we've stopped converging'. Why
> >> wouldn't that work?
> >>
> We used to have a heuristic that said 'if an iteration transfers more
> pages than the previous iteration, we've stopped converging'. Why
> wouldn't that work?
This does not protect you from very long migration times.
- Dietmar
--
To unsubscribe from this list: send the line "unsubscribe kvm"
> > > > +if ((stage == 2) && (bytes_transferred >
> 2*ram_bytes_total())) {
> > > > +return 1;
> > > > +}
> > > why 2 * ?
> > > This means we'll have to transfer the whole contents of RAM at
> least
> > > twice to hit this condition, right?
> >
> > Yes, this is just an arbitrary lim
> On Wed, Sep 30, 2009 at 10:55:24AM +0200, Dietmar Maurer wrote:
> > Another problem occur when max_downtime is too short. This can
> results in never ending migration task.
> >
> > To reproduce just play a video inside a VM and set max_downtime to
> 30ns
> >
>
) && (bytes_transferred > 2*ram_bytes_total())) {
+return 1;
+}
Or do you think that is not reasonable?
- Dietmar
> -Original Message-
> From: Glauber Costa [mailto:glom...@redhat.com]
> Sent: Mittwoch, 30. September 2009 06:49
> To: Dietmar Maurer
>
> Since the problem you pinpointed do exist, I would suggest measuring
> the average load of the last,
> say, 10 iterations.
The "last 10 interation" does not define a fixed time. I guess it is much more
reasonable to measure the average of the last '10 seconds'.
But usually a migration only tak
> Also, if this is really the case (buffered), then the bandwidth capping
> part
> of migration is also wrong.
>
> Have you compared the reported bandwidth to your actual bandwith ? I
> suspect
> the source of the problem can be that we're currently ignoring the time
> we take
> to transfer the st
this patch solves the problem by calculation an average bandwidth.
- Dietmar
> -Original Message-
> From: kvm-ow...@vger.kernel.org [mailto:kvm-ow...@vger.kernel.org] On
> Behalf Of Dietmar Maurer
> Sent: Dienstag, 29. September 2009 16:37
> To: kvm
> Subject: RE: mi
kvm-ow...@vger.kernel.org [mailto:kvm-ow...@vger.kernel.org] On
> Behalf Of Dietmar Maurer
> Sent: Dienstag, 29. September 2009 15:01
> To: kvm
> Subject: migrate_set_downtime bug
>
> using 0.11.0, live migration works as expected, but max downtime does
> not seem to work, for example:
&g
using 0.11.0, live migration works as expected, but max downtime does not seem
to work, for example:
# migrate_set_downtime 1
After that tcp migration has much longer downtimes (up to 20 seconds).
Also, it seems that the 'monitor' is locked (take up to 10 seconds until I get
a monitor prompt).
VirtIO is still broken in qemu-kvm-0-11.0. Reverting the patch solves the
problem.
Should I report that bug somewhere else?
- Dietmar
> -Original Message-
> From: kvm-ow...@vger.kernel.org [mailto:kvm-ow...@vger.kernel.org] On
> Behalf Of Alexey Eromenko
> Sent: Donnerstag, 23. Juli 20
Hi all,
the kvm-XX packages always included the kernel module, but this is not the case
for stable qemu-kvm-0.11.0? So I am trying to compile the latest kernel module
for our 2.6.24 kernel.
# git clone git://git.kernel.org/pub/scm/virt/kvm/kvm-kmod.git
# cd kvm-kmod
# git submodule update --ini
Hi all,
what is the advantage of the --enable-io-thread options. It is disabled by
default - why?
- Dietmar
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
> I think this is a known problem in qemu mentioned here on qemu-devel
>
> http://article.gmane.org/gmane.comp.emulators.qemu/42321
Ok, it is fixed in the latest qemu git code.
- Dietmar
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger
Hi all,
the new vnc code seems to be broken. At least all java based viewers
have serious problems, also realvnc for X
For example:
# kvm -cdrom ubuntu-9.04-desktop-amd64.iso -vnc :0 & vncviewer :0
CConn: Throughput 2 kbit/s - changing to hextile encoding
CConn: Throughput 2000
Hi all,
is there a Czech keyboard mapping for kvm (vnc)?
- Dietmar
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
I just wonder if 0.XX is a good name for a stable release. May is suggest to
increase the major number (1.XX).
> -Original Message-
> From: kvm-ow...@vger.kernel.org [mailto:kvm-ow...@vger.kernel.org] On
> Behalf Of Anthony Liguori
> Sent: Sonntag, 08. März 2009 17:13
> To: Avi Kivity
>
use the vmmouse driver (instead of mouse) for X
> Do I have to configure something special on the quest?
> On my guests, "vnc mouse pointer" moves much faster than the real one -
> both when X is started, but also in console, with gpm started.
> Dietmar Maurer schrieb:
> > use the vmmouse driver (instead of mouse) for X
>
> Assuming it'll work, what about console? Well, I guess I could live
> with it.
>
> But what about Windows guests? I've heard unconfirmed rumours that
> Windows doesn'
Hi all,
we notice some problems when using the boot=on option:
-drive file=test.qcow2,if=ide,index=0,boot=on
Seems that some vm don't boot anymore, get:
Disk I/o error: Status = 0001
Disk I/o error: Status = 0001
Disk I/o error: Status = 0001
Windows 2000 could not start because th
I already tried that - it does not work (the tape is seen as st0 as expected,
but read/write operation does not work).
- Dietmar
> -Original Message-
> From: Laurent Vivier [mailto:[EMAIL PROTECTED]
> Sent: Donnerstag, 11. September 2008 13:01
> To: Dietmar Maur
Is there a way to use a scsi tape? I tried -drive file=/dev/sg0, but
that does not work (eject, rewind works, but read/write fails - no date
is written - no errors)
- Dietmar
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to [EMAIL PROTECTED]
More major
Hi all,
This patch add better vnc keymap support, i.e. you can now type a '|'
and '@' on german keybords.
It is based on a patch from ricardo ribalda I found on the qemu list
some months ago.
- Dietmar
keymap.diff
Description: keymap.diff
Well, I guess I need to update my kernel headers
> -Ursprüngliche Nachricht-
> Von: Zhang, Xiantao [mailto:[EMAIL PROTECTED]
> Gesendet: Donnerstag, 28. August 2008 13:01
> An: Dietmar Maurer; kvm@vger.kernel.org
> Betreff: RE: MADV_DONTFORK undeclared
>
>
Compiling latest version 74 give me (on debian 4.0):
/home/dietmar/pve/pve-kvm-2.6.24/kvm-74/qemu/qemu-kvm.c: In function
'kvm_setup_guest_memory':
/home/dietmar/pve/pve-kvm-2.6.24/kvm-74/qemu/qemu-kvm.c:821: error:
'MADV_DONTFORK' undeclared (first use in this function)
/home/dietmar/pve/pve-kvm
Maybe this would also simplify/fasten online migration without shared
storage, because you know what blocks changed.
> On Thu, Aug 7, 2008 at 10:29 AM, Dietmar Maurer
> <[EMAIL PROTECTED]> wrote:
> > Or maybe it possible to modify kvm to store all changes
> made to a
> like a 'savevm' without qcow2 requirements?
>
> would be great.
Yes, something like:
# savediff [filname] [maxsize]
After a backup you can revert all changes, so you basically have a
snapshot.
- Dietmar
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a messa
Or maybe it possible to modify kvm to store all changes made to a vm
(opposite off -snapshot option)?
> you could run LVM in the VM. be careful about block scanning
> tools on Dom0, could mistake the LVM structure inside a LV
> for the 'outer' one.
> (reiserfsck has this problem with image fi
> you could run LVM in the VM. be careful about block scanning
> tools on Dom0, could mistake the LVM structure inside a LV
> for the 'outer' one.
> (reiserfsck has this problem with image files)
No, thats no option, because I cant depend on the guest (i.e lvm2 is not
available for windows)
-
> > What is the suggested way to backup a running kvm instance
> which uses
> > several disk images? Currently I simply use a LVM2 snapshot if all
> > disk images resides on one lvm volume. But what if it uses
> several lvm
> > volumes?
>
> i'd try to suspend KVM, do all LVM snapshots, unsusp
What is the suggested way to backup a running kvm instance which uses
several disk images? Currently I simply use a LVM2 snapshot if all disk
images resides on one lvm volume. But what if it uses several lvm
volumes?
Or is it possible to make a consistent backup without lvm?
- Dietmar
--
To u
Kvm: kvm-70 (also tested with kvm-snapshot-20080713.tar.gz)
Kernel: 2.6.24.4
Trying to install WinXp SP2 with -smp 2 fails:
/usr/bin/kvm -usbdevice tablet -name xp -smp 2 -localtime -k de -drive
file=/var/lib/vz/template/iso/de_winxp_pro_with_sp2_vl.iso,if=ide,index=
2,media=cdrom -drive
file=/va
> Out of curiosity, why would you switch from grub to lilo?
There is still many software around using lilo, and maybe the
bug affects other software too. So I tried to make the bug
reproducable, and reported it to this list so that
kvm developers can fix it.
- Dietmar
--
To unsubscribe from th
Just tried to install debian etch (netinstall) onto a 'virtual' scsi
disk.
-drive file=/tmp/vm-101-default.qcow2,index=0,if=scsi,media=disk,boot=on
First, i needed to pass 'noapic' as kernel parameter, else Installation
hangs (works without problem when using IDE disk).
Installation then went fi
71 matches
Mail list logo