[Qemu-devel] Boot Problems Windows XP guest

[Qemu-devel] Boot Problems Windows XP guest

[Erik Rull]

Hi all,

is it possible to get back to the "legacy" BIOS instead of the (u)efi based
BIOS? I have problems booting the guest on a SSD HDD, there it reboots
infinitely, when running the guest on a rotating HDD, it works. On qemu 1.2.0
both disks work properly. I didn't find a way to select the BIOS type in QEMU.

Best regards,

Erik

Re: [Qemu-devel] Boot Problems Windows XP guest

[Stefan Hajnoczi]

On Mon, Aug 26, 2013 at 11:52:26AM +0200, Erik Rull wrote:
> is it possible to get back to the "legacy" BIOS instead of the (u)efi based
> BIOS? I have problems booting the guest on a SSD HDD, there it reboots
> infinitely, when running the guest on a rotating HDD, it works. On qemu 1.2.0
> both disks work properly. I didn't find a way to select the BIOS type in QEMU.

You can specify the BIOS blob to use with the -bios <filename> option.

Not sure about the issue you are describing.  SeaBIOS is a BIOS, not
UEFI.

Stefan

Re: [Qemu-devel] Boot Problems Windows XP guest

[Erik Rull]

Stefan Hajnoczi wrote:
> On Mon, Aug 26, 2013 at 11:52:26AM +0200, Erik Rull wrote:
>> is it possible to get back to the "legacy" BIOS instead of the (u)efi based
>> BIOS? I have problems booting the guest on a SSD HDD, there it reboots
>> infinitely, when running the guest on a rotating HDD, it works. On qemu 1.2.0
>> both disks work properly. I didn't find a way to select the BIOS type in QEMU.
>
> You can specify the BIOS blob to use with the -bios <filename> option.
>
> Not sure about the issue you are describing.  SeaBIOS is a BIOS, not
> UEFI.
>
> Stefan
>

Hi Stefan,

which BIOS is selected by default? It's more a guess, there must be a 
change between 1.2.0 and 1.6.0 that prevents a simple Windows XP from 
booting completely, if the guest HDD image is placed on a SSD. On a 
rotating HDD (with the same commandline except the path to the image) it 
boots successfully. The only difference is the speed of the disk access.
I will try an alternative BIOS, maybe this fixes it, if not I will try to 
do some regression tests, first trying 1.4.0.

Best regards,

Erik

Re: [Qemu-devel] Boot Problems Windows XP guest

[Paolo Bonzini]

Il 27/08/2013 22:26, Erik Rull ha scritto:
> Hi Stefan,
> 
> which BIOS is selected by default?

QEMU only ships with SeaBIOS.

> It's more a guess, there must be a
> change between 1.2.0 and 1.6.0 that prevents a simple Windows XP from
> booting completely, if the guest HDD image is placed on a SSD. On a
> rotating HDD (with the same commandline except the path to the image) it
> boots successfully. The only difference is the speed of the disk access.

It could be a real difference, actually.  An unexpectedly fast disk
might screw a sloppy driver.  IIRC you're not the first person reporting
it.  Stefan, do you think using block throttling could fix it (with some
trial and error)?

> I will try an alternative BIOS, maybe this fixes it, if not I will try
> to do some regression tests, first trying 1.4.0.

This could help.

Paolo

Re: [Qemu-devel] Boot Problems Windows XP guest

[Anthony Liguori]

[-- Attachment #1: Type: text/plain, Size: 1268 bytes --]

On Aug 27, 2013 4:32 PM, "Paolo Bonzini" <pbonzini@redhat.com> wrote:
>
> Il 27/08/2013 22:26, Erik Rull ha scritto:
> > Hi Stefan,
> >
> > which BIOS is selected by default?
>
> QEMU only ships with SeaBIOS.
>
> > It's more a guess, there must be a
> > change between 1.2.0 and 1.6.0 that prevents a simple Windows XP from
> > booting completely, if the guest HDD image is placed on a SSD. On a
> > rotating HDD (with the same commandline except the path to the image) it
> > boots successfully. The only difference is the speed of the disk access.
>
> It could be a real difference, actually.  An unexpectedly fast disk
> might screw a sloppy driver.  IIRC you're not the first person reporting
> it.  Stefan, do you think using block throttling could fix it (with some
> trial and error)?


Add cache=writethrough and I bet it'll work even on an SSD.


We changed the default in that timeframe.  The Windows IDE drivers can have
an issue if the IRQ comes too quickly to indicate the request has
completed.  This is what -win2k-hack is for.  That may also work here too.


Regards,


Anthony Liguori

>
> > I will try an alternative BIOS, maybe this fixes it, if not I will try
> > to do some regression tests, first trying 1.4.0.
>
> This could help.
>
> Paolo
>
>

[-- Attachment #2: Type: text/html, Size: 1708 bytes --]

Re: [Qemu-devel] Boot Problems Windows XP guest

[Erik Rull]

[-- Attachment #1: Type: text/plain, Size: 1958 bytes --]

> On August 27, 2013 at 11:37 PM Anthony Liguori <anthony@codemonkey.ws> wrote:
>
>
> On Aug 27, 2013 4:32 PM, "Paolo Bonzini" <pbonzini@redhat.com> wrote:
> >
> > Il 27/08/2013 22:26, Erik Rull ha scritto:
> > > Hi Stefan,
> > >
> > > which BIOS is selected by default?
> >
> > QEMU only ships with SeaBIOS.
> >
> > > It's more a guess, there must be a
> > > change between 1.2.0 and 1.6.0 that prevents a simple Windows XP from
> > > booting completely, if the guest HDD image is placed on a SSD. On a
> > > rotating HDD (with the same commandline except the path to the image) it
> > > boots successfully. The only difference is the speed of the disk access.
> >
> > It could be a real difference, actually.  An unexpectedly fast disk
> > might screw a sloppy driver.  IIRC you're not the first person reporting
> > it.  Stefan, do you think using block throttling could fix it (with some
> > trial and error)?
>
>
> Add cache=writethrough and I bet it'll work even on an SSD.
>
>
> We changed the default in that timeframe.  The Windows IDE drivers can have
> an issue if the IRQ comes too quickly to indicate the request has
> completed.  This is what -win2k-hack is for.  That may also work here too.
>
>
> Regards,
>
>
> Anthony Liguori
>
> >
> > > I will try an alternative BIOS, maybe this fixes it, if not I will try
> > > to do some regression tests, first trying 1.4.0.
> >
> > This could help.
> >
> > Paolo
> >
> >

Hi all,

thanks for the input. I tested both cache=writethrough and -win2k-hack but it
didn't help.
But I was able to capture a screenshot from the Windows XP bluescreen, it shows
only a kind of disk access problem, but this could still be related to the
"speed" of the disk access (see cropped screen attached). Even with disabled kvm
it does not work, it just takes longer until it reboots.

Any further ideas? Are we now too quick for XP?

Best regards,

Erik

[-- Attachment #2: qemu-xp-bluescreen.png --]
[-- Type: image/png, Size: 10618 bytes --]

Re: [Qemu-devel] Boot Problems Windows XP guest

[Paolo Bonzini]

Il 27/08/2013 23:37, Anthony Liguori ha scritto:
> On Aug 27, 2013 4:32 PM, "Paolo Bonzini" <pbonzini@redhat.com
> <mailto:pbonzini@redhat.com>> wrote:
>>
>> Il 27/08/2013 22:26, Erik Rull ha scritto:
>> > Hi Stefan,
>> >
>> > which BIOS is selected by default?
>>
>> QEMU only ships with SeaBIOS.
>>
>> > It's more a guess, there must be a
>> > change between 1.2.0 and 1.6.0 that prevents a simple Windows XP from
>> > booting completely, if the guest HDD image is placed on a SSD. On a
>> > rotating HDD (with the same commandline except the path to the image) it
>> > boots successfully. The only difference is the speed of the disk access.
>>
>> It could be a real difference, actually.  An unexpectedly fast disk
>> might screw a sloppy driver.  IIRC you're not the first person reporting
>> it.  Stefan, do you think using block throttling could fix it (with some
>> trial and error)?
> 
> Add cache=writethrough and I bet it'll work even on an SSD.

You'll also kill the SSD rather quickly though. :)

Paolo

> We changed the default in that timeframe.  The Windows IDE drivers can
> have an issue if the IRQ comes too quickly to indicate the request has
> completed.  This is what -win2k-hack is for.  That may also work here too.

Re: [Qemu-devel] Boot Problems Windows XP guest

[Stefan Hajnoczi]

On Tue, Aug 27, 2013 at 11:31:28PM +0200, Paolo Bonzini wrote:
> Il 27/08/2013 22:26, Erik Rull ha scritto:
> > It's more a guess, there must be a
> > change between 1.2.0 and 1.6.0 that prevents a simple Windows XP from
> > booting completely, if the guest HDD image is placed on a SSD. On a
> > rotating HDD (with the same commandline except the path to the image) it
> > boots successfully. The only difference is the speed of the disk access.
> 
> It could be a real difference, actually.  An unexpectedly fast disk
> might screw a sloppy driver.  IIRC you're not the first person reporting
> it.  Stefan, do you think using block throttling could fix it (with some
> trial and error)?

That might work.  You could start with something like -drive ...,iops=20
and then disable the limit from the QEMU monitor once the guest OS is
booting (block_set_io_throttle virtio0 0 0 0 0 0 0).

It would be easier to try -drive ...,cache=writethrough and -win2k-hack
first as Anthony suggests.

Stefan

Re: [Qemu-devel] Boot Problems Windows XP guest

[Erik Rull]

> On August 28, 2013 at 9:50 AM Stefan Hajnoczi <stefanha@gmail.com> wrote:
>
>
> On Tue, Aug 27, 2013 at 11:31:28PM +0200, Paolo Bonzini wrote:
> > Il 27/08/2013 22:26, Erik Rull ha scritto:
> > > It's more a guess, there must be a
> > > change between 1.2.0 and 1.6.0 that prevents a simple Windows XP from
> > > booting completely, if the guest HDD image is placed on a SSD. On a
> > > rotating HDD (with the same commandline except the path to the image) it
> > > boots successfully. The only difference is the speed of the disk access.
> >
> > It could be a real difference, actually.  An unexpectedly fast disk
> > might screw a sloppy driver.  IIRC you're not the first person reporting
> > it.  Stefan, do you think using block throttling could fix it (with some
> > trial and error)?
>
> That might work.  You could start with something like -drive ...,iops=20
> and then disable the limit from the QEMU monitor once the guest OS is
> booting (block_set_io_throttle virtio0 0 0 0 0 0 0).
>
> It would be easier to try -drive ...,cache=writethrough and -win2k-hack
> first as Anthony suggests.
>
> Stefan
>

Thanks.
I tried that, but when should I reset the throttle? When I reset it some seconds
after the BIOS screen disappeared same result as without throttling. When I keep
it, Windows still reboots, the cycle just takes longer (half an hour), but the
progress seems to be the same as without throttle.
Btw.: my disks are not virtio0, but ide-hd0 and ide-hd1 (yes, I set the throttle
on both!)
To exclude a damage of the image meanwhile, I went back to 1.2.0 and it boots
perfectly.

Best regards,

Erik

Re: [Qemu-devel] Boot Problems Windows XP guest

[Paolo Bonzini]

Il 28/08/2013 11:07, Erik Rull ha scritto:
>>> It could be a real difference, actually.  An unexpectedly fast disk
>>> might screw a sloppy driver.  IIRC you're not the first person reporting
>>> it.  Stefan, do you think using block throttling could fix it (with some
>>> trial and error)?
>>
>> That might work.  You could start with something like -drive ...,iops=20
>> and then disable the limit from the QEMU monitor once the guest OS is
>> booting (block_set_io_throttle virtio0 0 0 0 0 0 0).
>>
>> It would be easier to try -drive ...,cache=writethrough and -win2k-hack
>> first as Anthony suggests.
>>
>> Stefan
> 
> Thanks.
> I tried that, but when should I reset the throttle?

Never.  The bug will be there through the whole execution of the guest.

> When I reset it some seconds
> after the BIOS screen disappeared same result as without throttling. When I keep
> it, Windows still reboots, the cycle just takes longer (half an hour), but the
> progress seems to be the same as without throttle.

On second thought that is expected.  Until throttling kicks in, I/O will
complete just as fast as without throttling.  Maybe limiting the number
of bytes per second instead of I/O ops would be better.  Can you try
-drive ...,bps=1048576 (possibly higher or lower numbers too)?

And maybe Benoit's new algorithm could help too.  Benoit, do you have a
tree for Erik to try?

Paolo

Re: [Qemu-devel] Boot Problems Windows XP guest

[Benoît Canet]

Le Wednesday 28 Aug 2013 à 15:14:41 (+0200), Paolo Bonzini a écrit :
> Il 28/08/2013 11:07, Erik Rull ha scritto:
> >>> It could be a real difference, actually.  An unexpectedly fast disk
> >>> might screw a sloppy driver.  IIRC you're not the first person reporting
> >>> it.  Stefan, do you think using block throttling could fix it (with some
> >>> trial and error)?
> >>
> >> That might work.  You could start with something like -drive ...,iops=20
> >> and then disable the limit from the QEMU monitor once the guest OS is
> >> booting (block_set_io_throttle virtio0 0 0 0 0 0 0).
> >>
> >> It would be easier to try -drive ...,cache=writethrough and -win2k-hack
> >> first as Anthony suggests.
> >>
> >> Stefan
> > 
> > Thanks.
> > I tried that, but when should I reset the throttle?
> 
> Never.  The bug will be there through the whole execution of the guest.
> 
> > When I reset it some seconds
> > after the BIOS screen disappeared same result as without throttling. When I keep
> > it, Windows still reboots, the cycle just takes longer (half an hour), but the
> > progress seems to be the same as without throttle.
> 
> On second thought that is expected.  Until throttling kicks in, I/O will
> complete just as fast as without throttling.  Maybe limiting the number
> of bytes per second instead of I/O ops would be better.  Can you try
> -drive ...,bps=1048576 (possibly higher or lower numbers too)?
> 
> And maybe Benoit's new algorithm could help too.  Benoit, do you have a
> tree for Erik to try?

Hi Eric,

You can find my latest throttling work on the leaky_continuous_final2 branch
of the following git tree.

git clone git@gitorious.org:benoit-canet-qemu/benoit-canet-qemu.git
git checkout leaky_continuous_final2

Hope this help

Best regards

Benoît

Re: [Qemu-devel] Boot Problems Windows XP guest

[Paolo Bonzini]

Il 28/08/2013 15:34, Benoît Canet ha scritto:
>> On second thought that is expected.  Until throttling kicks in, I/O will
>> complete just as fast as without throttling.  Maybe limiting the number
>> of bytes per second instead of I/O ops would be better.  Can you try
>> -drive ...,bps=1048576 (possibly higher or lower numbers too)?
>>
>> And maybe Benoit's new algorithm could help too.  Benoit, do you have a
>> tree for Erik to try?
> 
> Hi Eric,
> 
> You can find my latest throttling work on the leaky_continuous_final2 branch
> of the following git tree.
> 
> git clone git@gitorious.org:benoit-canet-qemu/benoit-canet-qemu.git
> git checkout leaky_continuous_final2

What is the command line to use for avg bps = 1MB/s, max bps = 0?

Paolo

Re: [Qemu-devel] Boot Problems Windows XP guest

[Benoît Canet]

> What is the command line to use for avg bps = 1MB/s, max bps = 0?

On the monitor it would be:

block_set_io_throttle 1048576 0 0 0 0 0

When configuring throttle on the command line it would be:

-drive file=foo.raw,if=virtio,cache=none,bps=1048576

Also remember that the code is setting bps_max = avg/10 to please guest cfq
schedulers so a little burst will be done at startup.

Best regards

Benoît

Re: [Qemu-devel] Boot Problems Windows XP guest

[Paolo Bonzini]

Il 28/08/2013 15:51, Benoît Canet ha scritto:
>> What is the command line to use for avg bps = 1MB/s, max bps = 0?
> On the monitor it would be:
> 
> block_set_io_throttle 1048576 0 0 0 0 0
> 
> When configuring throttle on the command line it would be:
> 
> -drive file=foo.raw,if=virtio,cache=none,bps=1048576
> 
> Also remember that the code is setting bps_max = avg/10 to please guest cfq
> schedulers so a little burst will be done at startup.

Yes, I'm afraid that the burst is problematic here since the current
throttling algorithm doesn't help (and IIUC it behaves as if it had
max=avg/10 due to the 100ms slice time).  This is Windows so no CFQ
here.  Do you need to set bps_max=1 or will bps_max=0 work too?

Paolo

Re: [Qemu-devel] Boot Problems Windows XP guest

[Benoît Canet]

> Yes, I'm afraid that the burst is problematic here since the current
> throttling algorithm doesn't help (and IIUC it behaves as if it had
> max=avg/10 due to the 100ms slice time).  This is Windows so no CFQ
> here.  Do you need to set bps_max=1 or will bps_max=0 work too?

In it's current state the code will overwrite the burst if it's set smaller than
avg/10.
The result will be a small initial spike of size avg/10 then it will cap to avg.

Best regards

Benoît

Re: [Qemu-devel] Boot Problems Windows XP guest

[Paolo Bonzini]

Il 28/08/2013 16:05, Benoît Canet ha scritto:
>> Yes, I'm afraid that the burst is problematic here since the current
>> > throttling algorithm doesn't help (and IIUC it behaves as if it had
>> > max=avg/10 due to the 100ms slice time).  This is Windows so no CFQ
>> > here.  Do you need to set bps_max=1 or will bps_max=0 work too?
> In it's current state the code will overwrite the burst if it's set smaller than
> avg/10.
> The result will be a small initial spike of size avg/10 then it will cap to avg.

Can you whip up a patch to avoid that?  Or to honor bps_max if it is
specified (making avg/10 simply the default)?

Paolo

Re: [Qemu-devel] Boot Problems Windows XP guest

[Benoît Canet]

> Can you whip up a patch to avoid that?  Or to honor bps_max if it is
> specified (making avg/10 simply the default)?

Pushed in the same branch (it honor bps_max if specified).

Best regards

Benoît

Re: [Qemu-devel] Boot Problems Windows XP guest

[Erik Rull]

Benoît Canet wrote:
>> Can you whip up a patch to avoid that?  Or to honor bps_max if it is
>> specified (making avg/10 simply the default)?
>
> Pushed in the same branch (it honor bps_max if specified).
>
> Best regards
>
> Benoît
>

Hi all,

thanks for your help. I cloned the git and compiled it - but I'm not 
completely sure how to enable the throttling finally - there were several 
mails regarding averages and max values... And the "unit" of the values 
would be interesting.

Merci beaucoup.

Best regards,

Erik

Re: [Qemu-devel] Boot Problems Windows XP guest

[Benoît Canet]

> thanks for your help. I cloned the git and compiled it - but I'm not
> completely sure how to enable the throttling finally - there were
> several mails regarding averages and max values... And the "unit" of
> the values would be interesting.

Hi Erik,

The main settings are bps, bps_rd and bps_wr for total, read and write bandwith
throttling (unit is bytes) and iops, iops_rd, iops_wr for IO per second
throttling (unit is IO operation).

You should specify your settings on the -drive command line like in:
-drive file=foo.raw,if=virtio,cache=none,bps=1048576 for a 1 MB total bandwith.

In addition to that there is another set of parameters to configure the burst
ability of the throttling.
These setting are: bps_max, bps_rd_max, bps_wr_max, iops_mx, iops_rd_max, and
iops_wr_max.

Some bursting is enabled by default.

>From what Paolo said you should set bps_max = 1 to specify that the default
bursting is not set.

So:
-drive file=foo.raw,if=virtio,cache=none,bps=1048576,bps_max=1 for a 1 MB total
bandwith with almost no burst.

Best regards

Benoît

ps: take care of updating your repository and recompiling it since I made some
changes for you to use. (use git fetch origin)

> 
> Merci beaucoup.
> 
> Best regards,
> 
> Erik
> 
> 
> 

Re: [Qemu-devel] Boot Problems Windows XP guest

[Erik Rull]

Benoît Canet wrote:
>> thanks for your help. I cloned the git and compiled it - but I'm not
>> completely sure how to enable the throttling finally - there were
>> several mails regarding averages and max values... And the "unit" of
>> the values would be interesting.
>
> Hi Erik,
>
> The main settings are bps, bps_rd and bps_wr for total, read and write bandwith
> throttling (unit is bytes) and iops, iops_rd, iops_wr for IO per second
> throttling (unit is IO operation).
>
> You should specify your settings on the -drive command line like in:
> -drive file=foo.raw,if=virtio,cache=none,bps=1048576 for a 1 MB total bandwith.
>
> In addition to that there is another set of parameters to configure the burst
> ability of the throttling.
> These setting are: bps_max, bps_rd_max, bps_wr_max, iops_mx, iops_rd_max, and
> iops_wr_max.
>
> Some bursting is enabled by default.
>
>  From what Paolo said you should set bps_max = 1 to specify that the default
> bursting is not set.
>
> So:
> -drive file=foo.raw,if=virtio,cache=none,bps=1048576,bps_max=1 for a 1 MB total
> bandwith with almost no burst.
>
> Best regards
>
> Benoît
>
> ps: take care of updating your repository and recompiling it since I made some
> changes for you to use. (use git fetch origin)
>

Great thanks! I cloned half an hour before your email, and updated right 
now, but there was only an "already up to date"... (git fetch origin didn't 
say anything git pull says Already up-to-date.)

I will test it tomorrow and let you know my results.

Best regards,

Erik

Re: [Qemu-devel] Boot Problems Windows XP guest

[Erik Rull]

> On August 28, 2013 at 9:22 PM Erik Rull <erik.rull@rdsoftware.de> wrote:
>
>
> Benoît Canet wrote:
> >> thanks for your help. I cloned the git and compiled it - but I'm not
> >> completely sure how to enable the throttling finally - there were
> >> several mails regarding averages and max values... And the "unit" of
> >> the values would be interesting.
> >
> > Hi Erik,
> >
> > The main settings are bps, bps_rd and bps_wr for total, read and write
> > bandwith
> > throttling (unit is bytes) and iops, iops_rd, iops_wr for IO per second
> > throttling (unit is IO operation).
> >
> > You should specify your settings on the -drive command line like in:
> > -drive file=foo.raw,if=virtio,cache=none,bps=1048576 for a 1 MB total
> > bandwith.
> >
> > In addition to that there is another set of parameters to configure the
> > burst
> > ability of the throttling.
> > These setting are: bps_max, bps_rd_max, bps_wr_max, iops_mx, iops_rd_max,
> > and
> > iops_wr_max.
> >
> > Some bursting is enabled by default.
> >
> >  From what Paolo said you should set bps_max = 1 to specify that the default
> > bursting is not set.
> >
> > So:
> > -drive file=foo.raw,if=virtio,cache=none,bps=1048576,bps_max=1 for a 1 MB
> > total
> > bandwith with almost no burst.
> >
> > Best regards
> >
> > Benoît
> >
> > ps: take care of updating your repository and recompiling it since I made
> > some
> > changes for you to use. (use git fetch origin)
> >
>
> Great thanks! I cloned half an hour before your email, and updated right
> now, but there was only an "already up to date"... (git fetch origin didn't
> say anything git pull says Already up-to-date.)
>
> I will test it tomorrow and let you know my results.
>
> Best regards,
>
> Erik

Hi all,

it still fails :-(

My commandline section is (I played with bps between 0.5 and 2.0 MB/sec and iops
with 1000 and 500):
-drive file=/dev/sda2,cache=none,bps=548576,bps_max=1,iops_max=1000
Within qemu it looks like that:
QEMU 1.6.50 monitor - type 'help' for more information
(qemu) info block
ide0-hd0: /dev/sda2 (raw)
    I/O throttling:   bps=548576 bps_rd=0 bps_wr=0 bps_max=1 bps_rd_max=0
bps_wr_max=0 iops=0 iops_rd=0 iops_wr=0 iops_max=1000 iops_rd_max=0
iops_wr_max=0 iops_size=0
(qemu)

Any further ideas?

Thanks.

Best regards,

Erik

Re: [Qemu-devel] Boot Problems Windows XP guest

[Benoît Canet]

> My commandline section is (I played with bps between 0.5 and 2.0 MB/sec and iops
> with 1000 and 500):
> -drive file=/dev/sda2,cache=none,bps=548576,bps_max=1,iops_max=1000
> Within qemu it looks like that:
> QEMU 1.6.50 monitor - type 'help' for more information
> (qemu) info block
> ide0-hd0: /dev/sda2 (raw)
>     I/O throttling:   bps=548576 bps_rd=0 bps_wr=0 bps_max=1 bps_rd_max=0
> bps_wr_max=0 iops=0 iops_rd=0 iops_wr=0 iops_max=1000 iops_rd_max=0
> iops_wr_max=0 iops_size=0
> (qemu)

Why did you set iops_max but not iops ?

Best regards

Benoît

Re: [Qemu-devel] Boot Problems Windows XP guest

[Erik Rull]

> On August 29, 2013 at 11:25 AM Benoît Canet <benoit.canet@irqsave.net> wrote:
>
>
> > My commandline section is (I played with bps between 0.5 and 2.0 MB/sec and
> > iops
> > with 1000 and 500):
> > -drive file=/dev/sda2,cache=none,bps=548576,bps_max=1,iops_max=1000
> > Within qemu it looks like that:
> > QEMU 1.6.50 monitor - type 'help' for more information
> > (qemu) info block
> > ide0-hd0: /dev/sda2 (raw)
> >     I/O throttling:   bps=548576 bps_rd=0 bps_wr=0 bps_max=1 bps_rd_max=0
> > bps_wr_max=0 iops=0 iops_rd=0 iops_wr=0 iops_max=1000 iops_rd_max=0
> > iops_wr_max=0 iops_size=0
> > (qemu)
>
> Why did you set iops_max but not iops ?
>
> Best regards
>
> Benoît
>

Good point, I added that, but it still keeps rebooting:
(qemu) info block
ide0-hd0: /dev/sda2 (raw)
    I/O throttling:   bps=548576 bps_rd=0 bps_wr=0 bps_max=1 bps_rd_max=0
bps_wr_max=0 iops=500 iops_rd=0 iops_wr=0 iops_max=500 iops_rd_max=0
iops_wr_max=0 iops_size=0

Best regards,

Erik

Re: [Qemu-devel] Boot Problems Windows XP guest

[Paolo Bonzini]

Il 29/08/2013 16:51, Erik Rull ha scritto:
> 
> 
>> On August 29, 2013 at 11:25 AM Benoît Canet <benoit.canet@irqsave.net> wrote:
>>
>>
>>> My commandline section is (I played with bps between 0.5 and 2.0 MB/sec and
>>> iops
>>> with 1000 and 500):
>>> -drive file=/dev/sda2,cache=none,bps=548576,bps_max=1,iops_max=1000
>>> Within qemu it looks like that:
>>> QEMU 1.6.50 monitor - type 'help' for more information
>>> (qemu) info block
>>> ide0-hd0: /dev/sda2 (raw)
>>>     I/O throttling:   bps=548576 bps_rd=0 bps_wr=0 bps_max=1 bps_rd_max=0
>>> bps_wr_max=0 iops=0 iops_rd=0 iops_wr=0 iops_max=1000 iops_rd_max=0
>>> iops_wr_max=0 iops_size=0
>>> (qemu)
>>
>> Why did you set iops_max but not iops ?
>>
>> Best regards
>>
>> Benoît
>>
> 
> Good point, I added that, but it still keeps rebooting:
> (qemu) info block
> ide0-hd0: /dev/sda2 (raw)
>     I/O throttling:   bps=548576 bps_rd=0 bps_wr=0 bps_max=1 bps_rd_max=0
> bps_wr_max=0 iops=500 iops_rd=0 iops_wr=0 iops_max=500 iops_rd_max=0
> iops_wr_max=0 iops_size=0

Too bad.

But iops_max w/o iops makes sense: it means no more than 1000 iops will
be in flight at the same time.  Note that "s" is a plural in "*s_max",
not "per seconds". :)

You could try iops_max=1 and make it higher if it works.

Paolo

Re: [Qemu-devel] Boot Problems Windows XP guest

[Erik Rull]

> On August 29, 2013 at 4:55 PM Paolo Bonzini <pbonzini@redhat.com> wrote:
>
>
> Il 29/08/2013 16:51, Erik Rull ha scritto:
> >
> >
> >> On August 29, 2013 at 11:25 AM Benoît Canet <benoit.canet@irqsave.net>
> >> wrote:
> >>
> >>
> >>> My commandline section is (I played with bps between 0.5 and 2.0 MB/sec
> >>> and
> >>> iops
> >>> with 1000 and 500):
> >>> -drive file=/dev/sda2,cache=none,bps=548576,bps_max=1,iops_max=1000
> >>> Within qemu it looks like that:
> >>> QEMU 1.6.50 monitor - type 'help' for more information
> >>> (qemu) info block
> >>> ide0-hd0: /dev/sda2 (raw)
> >>>     I/O throttling:   bps=548576 bps_rd=0 bps_wr=0 bps_max=1 bps_rd_max=0
> >>> bps_wr_max=0 iops=0 iops_rd=0 iops_wr=0 iops_max=1000 iops_rd_max=0
> >>> iops_wr_max=0 iops_size=0
> >>> (qemu)
> >>
> >> Why did you set iops_max but not iops ?
> >>
> >> Best regards
> >>
> >> Benoît
> >>
> >
> > Good point, I added that, but it still keeps rebooting:
> > (qemu) info block
> > ide0-hd0: /dev/sda2 (raw)
> >     I/O throttling:   bps=548576 bps_rd=0 bps_wr=0 bps_max=1 bps_rd_max=0
> > bps_wr_max=0 iops=500 iops_rd=0 iops_wr=0 iops_max=500 iops_rd_max=0
> > iops_wr_max=0 iops_size=0
>
> Too bad.
>
> But iops_max w/o iops makes sense: it means no more than 1000 iops will
> be in flight at the same time.  Note that "s" is a plural in "*s_max",
> not "per seconds". :)
>
> You could try iops_max=1 and make it higher if it works.
>
> Paolo
>

My SSD died - I took another one, cloned again from the spinning disk - works
*douh* - without any throttling!

If I run again in this situation, the real hardware might be the culprit.
Sorry for the time this issue took.
But I will try to make some deeper investigation on this issue - because it
worked with 1.2.0 and 1.6.50 failed...

Best regards,

Erik

Re: [Qemu-devel] Boot Problems Windows XP guest

[Stefan Hajnoczi]

On Wed, Aug 28, 2013 at 11:07:55AM +0200, Erik Rull wrote:
> To exclude a damage of the image meanwhile, I went back to 1.2.0 and it boots
> perfectly.

If you have any doubts about image corruption or differences, try:

  $ qemu-img compare /spinning-disk/vm.img /ssd/vm.img

Their contents should be identical if you just copied it over.

Stefan