Re: Timekeeping in stable/9

2012-01-22 Thread Petro Rossini 
On Sun, Jan 22, 2012 at 4:42 AM, Kevin Oberman  wrote:
> I suspect that most VB users on FreeBSD use it to get access to a small
> number of took on Windows...the ones in Office. It's still better than
> either LibreOffice or OpenOffice.org for either documents or presentations,
> especially presentations that will be displayed on a Windows system (as
> most are).

We are running a Zimbra Mail Server on FreebSD + VirtualBox.

I have seen timekeeping issues with various virtualization solutions
over the years, including VMWare ESXi and ESX. I tried NTP inside the
VMs as well as some tricks to enhance the syncing with the host - all
with mixed success.

One of the reasons I love jails;-)

Sorry, it does not help to solve the specific problem but I thought I
add my usage example - it is not all Windows..

Regards
Peter
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-21 Thread Kevin Oberman 
On Sat, Jan 21, 2012 at 7:46 AM, Jeremy Chadwick
wrote:

> On Sat, Jan 21, 2012 at 02:11:51PM +0100, Martin Sugioarto wrote:
> > Am Sat, 21 Jan 2012 13:20:51 +0100
> > schrieb "Ronald Klop" :
>  > > BTW: My experience with VBox is that it is nice for hobby stuff, but
> > > not for heavy load server stuff. VMWare does a better job there.
>
> Agreed.
>
> The number of problems with VirtualBox continues to grow (you should see
> my Mail/sent folder sometime).  I don't know or why anyone on earth
> would use VirtualBox with problems like what's described in this thread,
> combined with problems like what's described below (which TMK still
> hasn't been addressed):
>
> http://lists.freebsd.org/pipermail/freebsd-stable/2011-February/061642.html
> http://lists.freebsd.org/pipermail/freebsd-stable/2011-July/063172.html
> http://lists.freebsd.org/pipermail/freebsd-stable/2011-July/063221.html
> http://lists.freebsd.org/pipermail/freebsd-stable/2011-July/063322.html
>
> I would recommend you stick with bare metal, or use something that's at
> least well-established like VMware products (Workstation, for example,
> is actually affordable, and ESXi is free -- though there have been
> problems posted to the list about FreeBSD on ESXi as well).  Xen gets
> praise here on the lists, but I haven't tried it myself.  I stick with
> bare metal for everything, sans "tinkering around".
>

Hmm. The OP has XP as a guest on FreeBSD9. I am unaware that VMware has any
software for this purpose. Specifically, the only virtualization system
that I am aware can run on FreeBSD as the host is VirtualBox.  I won't ague
about its many issues and limitations, but for any use of FreeBSD as a host
system, it's the only game in town.

I suspect that most VB users on FreeBSD use it to get access to a small
number of took on Windows...the ones in Office. It's still better than
either LibreOffice or OpenOffice.org for either documents or presentations,
especially presentations that will be displayed on a Windows system (as
most are).
-- 
R. Kevin Oberman, Network Engineer
E-mail: kob6...@gmail.com
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-21 Thread Joe Holden 

Ronald Klop wrote:
On Sat, 21 Jan 2012 14:11:51 +0100, Martin Sugioarto 
 wrote:



Am Sat, 21 Jan 2012 13:20:51 +0100
schrieb "Ronald Klop" :


Hi,

As I understand it.
Host: FreeBSD 9
Guest: WinXP

Which one has troubles with its clock? The host or the guest or both?


Hi,

only inside VirtualBox, I think it's only an application problem and
my emails would be probably better addressed to ports@. ONLY the guest
is affected when host is loaded.

I noticed additionally:

You get better results with a desync'ed clock in the guest system, when
you start "openssl speed -multi 20" or similar. Within a few seconds the
clock gets a 20 seconds difference.


How many CPU's did you assign to the guest?
Did you install virtualbox guest additions to the guest?


Here a few details (guest additions are installed):

Memory size: 1600MB
Page Fusion: off
VRAM size:   256MB
HPET:on/off (tried both settings)
Chipset: piix3
Firmware:BIOS
Number of CPUs:  1
Synthetic Cpu:   off
CPUID overrides: None
[...]
ACPI:on
IOAPIC:  off
PAE: on
Time offset: 0 ms
RTC: local time
Hardw. virt.ext: on
Hardw. virt.ext exclusive: on
Nested Paging:   on
Large Pages: on
VT-x VPID:   on
[...]
3D Acceleration: off
2D Video Acceleration: on


Do you run NTP on the guest XP also? If yes, turn it off.


Windows XP default installation (synch'ed to time.windows.com).
Switching this off, does not have any influence. I think MS-Windows
does not do continuous synchronization, only at system start, I guess.


VBox guest additions can sync the guest clock with the host.


I'll try to deinstall them. But I somehow like my shared folder.


BTW: My experience with VBox is that it is nice for hobby stuff, but
not for heavy load server stuff. VMWare does a better job there.


Yes. I know. Still VirtualBox ist nice and cheap solution.

--
Martin


BTW: I used VBox on Linux at work. Same problems. Different problems 
come and go with different versions of Linux in combination with 
different versions of VirtualBox. Using VmWare ESXI solved it. If you 
search a lot on the vmware website you will find a free version.


Ronald.
In the extreme case I have here, the host isn't taxed at all, cpu, disk 
i/o and such are almost idle but the time is skewed dramatically regardless.


For reference the settings I have are:

4 VCPUS (4 physical cores)
1GB ram
ICH9, SAS controller

If I toggle the sysctl in my previous post the problem goes way, and 
doesn't return even if the sysctl is changed back... until a reboot of 
course.  None of the pre-9 guests (there are quite a few spread across a 
couple of identical machines) exhibit the behaviour, nor does this 
particular one when reverted to a pre-upgrade snapshot, so in this case 
it is certainly not the hardware but whatever is used to keep track of 
the "ticks" (terminology probably incorrect)


Thanks,
J
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-21 Thread Jeremy Chadwick 
On Sat, Jan 21, 2012 at 02:11:51PM +0100, Martin Sugioarto wrote:
> Am Sat, 21 Jan 2012 13:20:51 +0100
> schrieb "Ronald Klop" :
> > Do you run NTP on the guest XP also? If yes, turn it off.
> 
> Windows XP default installation (synch'ed to time.windows.com).
> Switching this off, does not have any influence. I think MS-Windows
> does not do continuous synchronization, only at system start, I guess.

Incorrect; Windows XP syncs the clock using the W32Time service, which
polls NTP at the interval specified in DWORD registry key
HKLM\SYSTEM\CurrentControlSet\Services\W32Time\TimeProviders\NtpClient\SpecialPollInterval.

You can read more about it in detail here (see section "Windows XP
Professional and all versions of Windows Server 2003"):

http://support.microsoft.com/kb/884776

> > BTW: My experience with VBox is that it is nice for hobby stuff, but
> > not for heavy load server stuff. VMWare does a better job there.

Agreed.

The number of problems with VirtualBox continues to grow (you should see
my Mail/sent folder sometime).  I don't know or why anyone on earth
would use VirtualBox with problems like what's described in this thread,
combined with problems like what's described below (which TMK still
hasn't been addressed):

http://lists.freebsd.org/pipermail/freebsd-stable/2011-February/061642.html
http://lists.freebsd.org/pipermail/freebsd-stable/2011-July/063172.html
http://lists.freebsd.org/pipermail/freebsd-stable/2011-July/063221.html
http://lists.freebsd.org/pipermail/freebsd-stable/2011-July/063322.html

I would recommend you stick with bare metal, or use something that's at
least well-established like VMware products (Workstation, for example,
is actually affordable, and ESXi is free -- though there have been
problems posted to the list about FreeBSD on ESXi as well).  Xen gets
praise here on the lists, but I haven't tried it myself.  I stick with
bare metal for everything, sans "tinkering around".

-- 
| Jeremy Chadwick j...@parodius.com |
| Parodius Networking http://www.parodius.com/ |
| UNIX Systems Administrator Mountain View, CA, US |
| Making life hard for others since 1977. PGP 4BD6C0CB |

___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-21 Thread Ronald Klop 
On Sat, 21 Jan 2012 14:11:51 +0100, Martin Sugioarto  
 wrote:



Am Sat, 21 Jan 2012 13:20:51 +0100
schrieb "Ronald Klop" :


Hi,

As I understand it.
Host: FreeBSD 9
Guest: WinXP

Which one has troubles with its clock? The host or the guest or both?


Hi,

only inside VirtualBox, I think it's only an application problem and
my emails would be probably better addressed to ports@. ONLY the guest
is affected when host is loaded.

I noticed additionally:

You get better results with a desync'ed clock in the guest system, when
you start "openssl speed -multi 20" or similar. Within a few seconds the
clock gets a 20 seconds difference.


How many CPU's did you assign to the guest?
Did you install virtualbox guest additions to the guest?


Here a few details (guest additions are installed):

Memory size: 1600MB
Page Fusion: off
VRAM size:   256MB
HPET:on/off (tried both settings)
Chipset: piix3
Firmware:BIOS
Number of CPUs:  1
Synthetic Cpu:   off
CPUID overrides: None
[...]
ACPI:on
IOAPIC:  off
PAE: on
Time offset: 0 ms
RTC: local time
Hardw. virt.ext: on
Hardw. virt.ext exclusive: on
Nested Paging:   on
Large Pages: on
VT-x VPID:   on
[...]
3D Acceleration: off
2D Video Acceleration: on


Do you run NTP on the guest XP also? If yes, turn it off.


Windows XP default installation (synch'ed to time.windows.com).
Switching this off, does not have any influence. I think MS-Windows
does not do continuous synchronization, only at system start, I guess.


VBox guest additions can sync the guest clock with the host.


I'll try to deinstall them. But I somehow like my shared folder.


BTW: My experience with VBox is that it is nice for hobby stuff, but
not for heavy load server stuff. VMWare does a better job there.


Yes. I know. Still VirtualBox ist nice and cheap solution.

--
Martin


BTW: I used VBox on Linux at work. Same problems. Different problems come  
and go with different versions of Linux in combination with different  
versions of VirtualBox. Using VmWare ESXI solved it. If you search a lot  
on the vmware website you will find a free version.


Ronald.
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-21 Thread Alexander Motin 

On 01/21/12 15:20, Martin Sugioarto wrote:

Am Sat, 21 Jan 2012 14:30:53 +0200
schrieb Alexander Motin:

I am not using VirtualBox right now, so I'll need to setup it to test
this. Meanwhile you could try to experiment with switching to
different timecounters and eventtimers. May be some change in 9.0
changed default timecounter for you, causing the problem.


I think we have a misunderstanding here. The host (FreeBSD 9.0R) works
fine. The time is being updated under heavy load without problems.

I already said that this seems to be an application problem and this
email(s) should be rather seen by the VBox maintainer. The problem is
that VBox seems to stop working properly when you put heavy CPU load on
the host. It even does not keep the clock up-to-date.

I can desync the guest clock to -1 minute in a few seconds, just by
running "openssl speed -multi 20".


Ah. I'm sorry. I was sure we are debugging FreeBSD inside VirtualBox. If 
we are speaking about FreeBSD outside, then neither timecounter nor 
eventtimer choice should not affect guest if host is working fine. It is 
more question to VirtualBox and may be host system scheduler.


--
Alexander Motin
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-21 Thread Martin Sugioarto 
Am Sat, 21 Jan 2012 14:30:53 +0200
schrieb Alexander Motin :

Hi Alexander,

> I am not using VirtualBox right now, so I'll need to setup it to test 
> this. Meanwhile you could try to experiment with switching to
> different timecounters and eventtimers. May be some change in 9.0
> changed default timecounter for you, causing the problem.

I think we have a misunderstanding here. The host (FreeBSD 9.0R) works
fine. The time is being updated under heavy load without problems.

I already said that this seems to be an application problem and this
email(s) should be rather seen by the VBox maintainer. The problem is
that VBox seems to stop working properly when you put heavy CPU load on
the host. It even does not keep the clock up-to-date.

I can desync the guest clock to -1 minute in a few seconds, just by
running "openssl speed -multi 20".

> timecounter wrap should be the main cause of time drift (if
> timecounter hardware is emulated correctly at all). Different
> timecounters have different wrap periods that can be calculated by
> dividing kern.timecounter.tc.X.mask on
> kern.timecounter.tc.X.frequency. In my case there are: 300s for HPET,
> 5s for ACPI-fast, 2s for TSC and 55ms for i8254. If system won't get
> timer interrupts within half of that time -- time will drift. Start
> from looking what you are using and how good it is in your case.

This is my current timecounter setting:
kern.timecounter.choice: TSC-low(1000) HPET(950) i8254(0)
ACPI-fast(900) dummy(-100)
kern.timecounter.hardware: TSC-low
kern.timecounter.tc.ACPI-fast.mask: 16777215
kern.timecounter.tc.ACPI-fast.counter: 10400371
kern.timecounter.tc.ACPI-fast.frequency: 3579545
kern.timecounter.tc.ACPI-fast.quality: 900
kern.timecounter.tc.i8254.mask: 65535
kern.timecounter.tc.i8254.counter: 41849
kern.timecounter.tc.i8254.frequency: 1193182
kern.timecounter.tc.i8254.quality: 0
kern.timecounter.tc.HPET.mask: 4294967295
kern.timecounter.tc.HPET.counter: 3255982446
kern.timecounter.tc.HPET.frequency: 14318180
kern.timecounter.tc.HPET.quality: 950
kern.timecounter.tc.TSC-low.mask: 4294967295
kern.timecounter.tc.TSC-low.counter: 1587561917
kern.timecounter.tc.TSC-low.frequency: 8593928
kern.timecounter.tc.TSC-low.quality: 1000
kern.timecounter.smp_tsc: 1
kern.timecounter.invariant_tsc: 1

I can try other timer counter settings, if you think it will improve
responsivity inside VBox guest.

--
Martin


signature.asc
Description: PGP signature

Re: Timekeeping in stable/9

2012-01-21 Thread Martin Sugioarto 
Am Sat, 21 Jan 2012 13:20:51 +0100
schrieb "Ronald Klop" :

> Hi,
> 
> As I understand it.
> Host: FreeBSD 9
> Guest: WinXP
> 
> Which one has troubles with its clock? The host or the guest or both?

Hi,

only inside VirtualBox, I think it's only an application problem and
my emails would be probably better addressed to ports@. ONLY the guest
is affected when host is loaded.

I noticed additionally:

You get better results with a desync'ed clock in the guest system, when
you start "openssl speed -multi 20" or similar. Within a few seconds the
clock gets a 20 seconds difference.

> How many CPU's did you assign to the guest?
> Did you install virtualbox guest additions to the guest?

Here a few details (guest additions are installed):

Memory size: 1600MB
Page Fusion: off
VRAM size:   256MB
HPET:on/off (tried both settings)
Chipset: piix3
Firmware:BIOS
Number of CPUs:  1
Synthetic Cpu:   off
CPUID overrides: None
[...]
ACPI:on
IOAPIC:  off
PAE: on
Time offset: 0 ms
RTC: local time
Hardw. virt.ext: on
Hardw. virt.ext exclusive: on
Nested Paging:   on
Large Pages: on
VT-x VPID:   on
[...]
3D Acceleration: off
2D Video Acceleration: on

> Do you run NTP on the guest XP also? If yes, turn it off.

Windows XP default installation (synch'ed to time.windows.com).
Switching this off, does not have any influence. I think MS-Windows
does not do continuous synchronization, only at system start, I guess.

> VBox guest additions can sync the guest clock with the host.

I'll try to deinstall them. But I somehow like my shared folder.

> BTW: My experience with VBox is that it is nice for hobby stuff, but
> not for heavy load server stuff. VMWare does a better job there.

Yes. I know. Still VirtualBox ist nice and cheap solution.

--
Martin


signature.asc
Description: PGP signature

Re: Timekeeping in stable/9

2012-01-21 Thread Ronald Klop 
On Sat, 21 Jan 2012 10:18:42 +0100, Martin Sugioarto  
 wrote:



Am Wed, 18 Jan 2012 07:50:49 +0100
schrieb Martin Sugioarto :


I can confirm this on VirtualBox. I've been running WinXP inside
VirtualBox and measured network I/O during downloads. It showed me
very high download rates (around 800kB/s) while it's physically
possible to download 200kB/s through DSL here (Germany sucks with
DSL, even in largest cities, btw!).

I correlated this behavior with high disk I/O on the host. That means
that the timer issues on the virtual host appear when I start a
larger cp job on the host. I also immediately thought that this has
something to do with timers.


Hi everybody,

I just want to add some information on this. I tested a few things with
VirtualBox yesterday.

I switched off ntpd on the host and tested if there are differences,
but the clock is working correctly on the host. I tested it a few times,
it is stable, as I expect it to be.

It seems to be rather a software problem with VirtualBox. I can see that
when the host is under heavy load (CPU!) the guest does not get enough
runtime to adjust the clock correctly. After a few minutes there has
been a difference of 50 seconds between the host and guest clock. And
furthermore, I don't quite understand how the real time clock works in
VirtualBox but it seems to slide in the different directions causing
weird results with progress bars on MS-Windows XP.

I just want to explain why I thought that I/O influences this. I have
got my hard disk encrypted, so it puts some load on the CPU, too.

If you want to test VirtualBox behavior, you can simple dd
from /dev/random and look at the weird results in VirtualBox.

--
I hope it helps further,
Martin



Hi,

As I understand it.
Host: FreeBSD 9
Guest: WinXP

Which one has troubles with its clock? The host or the guest or both?
How many CPU's did you assign to the guest?
Did you install virtualbox guest additions to the guest?
Do you run NTP on the guest XP also? If yes, turn it off. VBox guest  
additions can sync the guest clock with the host.


BTW: My experience with VBox is that it is nice for hobby stuff, but not  
for heavy load server stuff. VMWare does a better job there.


Ronald.
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-21 Thread Alexander Motin 

Hi.

On 01/21/12 11:18, Martin Sugioarto wrote:

Am Wed, 18 Jan 2012 07:50:49 +0100
schrieb Martin Sugioarto:


I can confirm this on VirtualBox. I've been running WinXP inside
VirtualBox and measured network I/O during downloads. It showed me
very high download rates (around 800kB/s) while it's physically
possible to download 200kB/s through DSL here (Germany sucks with
DSL, even in largest cities, btw!).

I correlated this behavior with high disk I/O on the host. That means
that the timer issues on the virtual host appear when I start a
larger cp job on the host. I also immediately thought that this has
something to do with timers.


I just want to add some information on this. I tested a few things with
VirtualBox yesterday.

I switched off ntpd on the host and tested if there are differences,
but the clock is working correctly on the host. I tested it a few times,
it is stable, as I expect it to be.

It seems to be rather a software problem with VirtualBox. I can see that
when the host is under heavy load (CPU!) the guest does not get enough
runtime to adjust the clock correctly. After a few minutes there has
been a difference of 50 seconds between the host and guest clock. And
furthermore, I don't quite understand how the real time clock works in
VirtualBox but it seems to slide in the different directions causing
weird results with progress bars on MS-Windows XP.

I just want to explain why I thought that I/O influences this. I have
got my hard disk encrypted, so it puts some load on the CPU, too.

If you want to test VirtualBox behavior, you can simple dd
from /dev/random and look at the weird results in VirtualBox.


I am not using VirtualBox right now, so I'll need to setup it to test 
this. Meanwhile you could try to experiment with switching to different 
timecounters and eventtimers. May be some change in 9.0 changed default 
timecounter for you, causing the problem.


timecounter wrap should be the main cause of time drift (if timecounter 
hardware is emulated correctly at all). Different timecounters have 
different wrap periods that can be calculated by dividing 
kern.timecounter.tc.X.mask on kern.timecounter.tc.X.frequency. In my 
case there are: 300s for HPET, 5s for ACPI-fast, 2s for TSC and 55ms for 
i8254. If system won't get timer interrupts within half of that time -- 
time will drift. Start from looking what you are using and how good it 
is in your case.


--
Alexander Motin
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-21 Thread Martin Sugioarto 
Am Wed, 18 Jan 2012 07:50:49 +0100
schrieb Martin Sugioarto :

> I can confirm this on VirtualBox. I've been running WinXP inside
> VirtualBox and measured network I/O during downloads. It showed me
> very high download rates (around 800kB/s) while it's physically
> possible to download 200kB/s through DSL here (Germany sucks with
> DSL, even in largest cities, btw!).
> 
> I correlated this behavior with high disk I/O on the host. That means
> that the timer issues on the virtual host appear when I start a
> larger cp job on the host. I also immediately thought that this has
> something to do with timers.

Hi everybody,

I just want to add some information on this. I tested a few things with
VirtualBox yesterday.

I switched off ntpd on the host and tested if there are differences,
but the clock is working correctly on the host. I tested it a few times,
it is stable, as I expect it to be.

It seems to be rather a software problem with VirtualBox. I can see that
when the host is under heavy load (CPU!) the guest does not get enough
runtime to adjust the clock correctly. After a few minutes there has
been a difference of 50 seconds between the host and guest clock. And
furthermore, I don't quite understand how the real time clock works in
VirtualBox but it seems to slide in the different directions causing
weird results with progress bars on MS-Windows XP.

I just want to explain why I thought that I/O influences this. I have
got my hard disk encrypted, so it puts some load on the CPU, too.

If you want to test VirtualBox behavior, you can simple dd
from /dev/random and look at the weird results in VirtualBox.

--
I hope it helps further,
Martin


signature.asc
Description: PGP signature

Re: Timekeeping in stable/9

2012-01-20 Thread John Baldwin 
On Friday, January 20, 2012 2:29:21 am Adrian Chadd wrote:
> .. if you have a _reproducable_ case for this, mav@ needs to see it ASAP.
> 
> The trouble is that mav@ isn't handed reproducable cases and thus can't
> debug it.
> 
> If you can supply some kind of box that provides this as a reproducable
> issue, we can get it fixed ASAP. Otherwise it's a case of "can't reproduce
> in our environments, sorry."

However, if we haven't mentioned eventtimers at all in the release notes, that 
is a bit of a fail on our parts.  That is a substantial change which needs to 
be adequately documented.

-- 
John Baldwin
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-19 Thread Adrian Chadd 
.. if you have a _reproducable_ case for this, mav@ needs to see it ASAP.

The trouble is that mav@ isn't handed reproducable cases and thus can't
debug it.

If you can supply some kind of box that provides this as a reproducable
issue, we can get it fixed ASAP. Otherwise it's a case of "can't reproduce
in our environments, sorry."


Adrian

On 19 January 2012 13:09, Joe Holden  wrote:

> Joe Holden wrote:
>
>> Chuck Swiger wrote:
>>
>>> On Jan 19, 2012, at 12:18 PM, Joe Holden wrote:
>>>
 Sounds like you were looking for commercial support, since unpaid
> volunteers don't have an obligation to promptly leap out and provide
> solutions within your ETA.
>
 Not really, just an acknowledgement would be fine.  It is what it is,
 everyday I try to argue in favour of the project, I still use it for myself
 everywhere but increasingly things happen that just don't on other
 volunteer projects... it's rather difficult to argue the case when they can
 install Ubuntu or whatever nonsense distro is the current favourite and it
 just works.  Just a bit more accurate info would solve it, if it doesn't do
 X reliably, or Y has changed, note it.

>>>
>>> You asked a question and got two or three responses back in a day.  You
>>> mentioned trying different timekeeping choices, but I don't recall seeing
>>> what your kern.timecounter sysctl values looked like; without that, folks
>>> are missing info that is likely to be relevant.
>>>
>>> Ah, well
>>>
>>> Regards,
>>>
>> Yeah my gripe isn't with having no responses, the handful of people that
>> have responded have been helpful but ultimately no responses from anyone
>> involved.  Just a one liner saying "we changed the timecounter stuff in 9,
>> look at sysctl tree X" would have been more than sufficient, this sort of
>> thing should really be mentioned in the relnotes though...
>>
>> For the record though, setting kern.eventtimer.periodic to 1 fixes the
>> problem on all affected machines (returns my virtualbox guest to normality,
>> reduces the drift on physical machines to 8.2 figures).
>>
>> FWIW, I can't even see any notes relating to this in UPDATING either.
>>
> I should probably clarify here that some responses were received from the
> maintainers (eg: Qing for mpath) for a couple of bits of code but the wider
> issues weren't discussed further.  I'm not trying to say that no effort is
> made, but as a whole for the project to be comparable to the alternatives
> this sort of thing shouldn't happen.
>
>
>
> __**_
> freebsd-stable@freebsd.org mailing list
> http://lists.freebsd.org/**mailman/listinfo/freebsd-**stable
> To unsubscribe, send any mail to 
> "freebsd-stable-unsubscribe@**freebsd.org
> "
>
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-19 Thread Joe Holden 

Joe Holden wrote:

Chuck Swiger wrote:

On Jan 19, 2012, at 12:18 PM, Joe Holden wrote:
Sounds like you were looking for commercial support, since unpaid 
volunteers don't have an obligation to promptly leap out and provide 
solutions within your ETA.
Not really, just an acknowledgement would be fine.  It is what it is, 
everyday I try to argue in favour of the project, I still use it for 
myself everywhere but increasingly things happen that just don't on 
other volunteer projects... it's rather difficult to argue the case 
when they can install Ubuntu or whatever nonsense distro is the 
current favourite and it just works.  Just a bit more accurate info 
would solve it, if it doesn't do X reliably, or Y has changed, note it.


You asked a question and got two or three responses back in a day.  
You mentioned trying different timekeeping choices, but I don't recall 
seeing what your kern.timecounter sysctl values looked like; without 
that, folks are missing info that is likely to be relevant.


Ah, well

Regards,
Yeah my gripe isn't with having no responses, the handful of people that 
have responded have been helpful but ultimately no responses from anyone 
involved.  Just a one liner saying "we changed the timecounter stuff in 
9, look at sysctl tree X" would have been more than sufficient, this 
sort of thing should really be mentioned in the relnotes though...


For the record though, setting kern.eventtimer.periodic to 1 fixes the 
problem on all affected machines (returns my virtualbox guest to 
normality, reduces the drift on physical machines to 8.2 figures).


FWIW, I can't even see any notes relating to this in UPDATING either.
I should probably clarify here that some responses were received from 
the maintainers (eg: Qing for mpath) for a couple of bits of code but 
the wider issues weren't discussed further.  I'm not trying to say that 
no effort is made, but as a whole for the project to be comparable to 
the alternatives this sort of thing shouldn't happen.



___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-19 Thread Joe Holden 

Chuck Swiger wrote:

On Jan 19, 2012, at 12:18 PM, Joe Holden wrote:

Sounds like you were looking for commercial support, since unpaid volunteers 
don't have an obligation to promptly leap out and provide solutions within your 
ETA.

Not really, just an acknowledgement would be fine.  It is what it is, everyday 
I try to argue in favour of the project, I still use it for myself everywhere 
but increasingly things happen that just don't on other volunteer projects... 
it's rather difficult to argue the case when they can install Ubuntu or 
whatever nonsense distro is the current favourite and it just works.  Just a 
bit more accurate info would solve it, if it doesn't do X reliably, or Y has 
changed, note it.


You asked a question and got two or three responses back in a day.  You 
mentioned trying different timekeeping choices, but I don't recall seeing what 
your kern.timecounter sysctl values looked like; without that, folks are 
missing info that is likely to be relevant.

Ah, well

Regards,
Yeah my gripe isn't with having no responses, the handful of people that 
have responded have been helpful but ultimately no responses from anyone 
involved.  Just a one liner saying "we changed the timecounter stuff in 
9, look at sysctl tree X" would have been more than sufficient, this 
sort of thing should really be mentioned in the relnotes though...


For the record though, setting kern.eventtimer.periodic to 1 fixes the 
problem on all affected machines (returns my virtualbox guest to 
normality, reduces the drift on physical machines to 8.2 figures).


FWIW, I can't even see any notes relating to this in UPDATING either.
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-19 Thread Chuck Swiger 
On Jan 19, 2012, at 12:18 PM, Joe Holden wrote:
>> Sounds like you were looking for commercial support, since unpaid volunteers 
>> don't have an obligation to promptly leap out and provide solutions within 
>> your ETA.
> 
> Not really, just an acknowledgement would be fine.  It is what it is, 
> everyday I try to argue in favour of the project, I still use it for myself 
> everywhere but increasingly things happen that just don't on other volunteer 
> projects... it's rather difficult to argue the case when they can install 
> Ubuntu or whatever nonsense distro is the current favourite and it just 
> works.  Just a bit more accurate info would solve it, if it doesn't do X 
> reliably, or Y has changed, note it.

You asked a question and got two or three responses back in a day.  You 
mentioned trying different timekeeping choices, but I don't recall seeing what 
your kern.timecounter sysctl values looked like; without that, folks are 
missing info that is likely to be relevant.

Ah, well

Regards,
-- 
-Chuck

___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-19 Thread Joe Holden 

Chuck Swiger wrote:

On Jan 19, 2012, at 12:04 PM, Joe Holden wrote:

Looks like this is down to the dynamic/tickless changes in 9 (that aren't even 
noted in the release notes), the machines have now been switched to linux as 
the lack of responses/care given to my recent postings has been noted and it 
was deemed that using linux would be less hassle in the long run.


Sounds like you were looking for commercial support, since unpaid volunteers 
don't have an obligation to promptly leap out and provide solutions within your 
ETA.

Regards,
Not really, just an acknowledgement would be fine.  It is what it is, 
everyday I try to argue in favour of the project, I still use it for 
myself everywhere but increasingly things happen that just don't on 
other volunteer projects... it's rather difficult to argue the case when 
they can install Ubuntu or whatever nonsense distro is the current 
favourite and it just works.  Just a bit more accurate info would solve 
it, if it doesn't do X reliably, or Y has changed, note it.

___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-19 Thread Chuck Swiger 
On Jan 19, 2012, at 12:04 PM, Joe Holden wrote:
> Looks like this is down to the dynamic/tickless changes in 9 (that aren't 
> even noted in the release notes), the machines have now been switched to 
> linux as the lack of responses/care given to my recent postings has been 
> noted and it was deemed that using linux would be less hassle in the long run.

Sounds like you were looking for commercial support, since unpaid volunteers 
don't have an obligation to promptly leap out and provide solutions within your 
ETA.

Regards,
-- 
-Chuck

___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-19 Thread Joe Holden 
Looks like this is down to the dynamic/tickless changes in 9 (that 
aren't even noted in the release notes), the machines have now been 
switched to linux as the lack of responses/care given to my recent 
postings has been noted and it was deemed that using linux would be less 
hassle in the long run.


Unfortunate decision but I am inclined to agree.

Thanks,
J

Ian Lepore wrote:

On Tue, 2012-01-17 at 20:12 +, Joe Holden wrote:

Hi guys,

Has anyone else noticed the tendency for 9.0-R to be unable to 
accurately keep time?  I've got a couple of machines that have been 
upgraded from 8.2 that are struggling, in particular a Virtual box guest 
that was fine on 8.2, but now that's its been upgraded to 9.0 counts at 
anything from 2 to 20 seconds per 5 second sample, the result is similar 
with HPET, ACPI-fast and TSC.


I also have physical boxes which new seem to drift quite substantially, 
ntpd cannot keep up and as these boxes need to be able to report the 
time relatively accurately, it is causing problems with log times and 
such...


Any suggestions most welcome!

Thanks,
J


I finally got a 9.0 generic build done today and I've been watching the
timekeeping on 3 systems and they're all doing just fine.  Two of the
systems are performing pretty much identically to how they did on 8.2;
the clock frequency correction calculated by ntpd differs by less than
1ppm.  On the other system the kernel timekeeping routines are now
choosing to use a different clock so I don't get a direct comparison of
the old vs new drift rate, but the drift is still reasonable  (100ppm
now, used to be around 88, on an old 300mhz MediaGx-based system).

I haven't had time yet to learn about the new eventtimer stuff in 9.0,
but I know you can get some info on the choices it made via sysctl
kern.eventtimer.  Before 9.0 I'd check sysctl kern.clockrate and vmstat
-i and make sure the chosen clock is interrupting at the right rate, but
now with the eventtimer stuff there's not an obvious correlation between
hz and profhz and stathz and any particular device's interrupt rate, at
least for some clock choices (on the old MediaGx system without ACPI or
HPET it seems to work more like it used to).

-- Ian




___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-18 Thread Ian Lepore 
On Tue, 2012-01-17 at 20:12 +, Joe Holden wrote:
> Hi guys,
> 
> Has anyone else noticed the tendency for 9.0-R to be unable to 
> accurately keep time?  I've got a couple of machines that have been 
> upgraded from 8.2 that are struggling, in particular a Virtual box guest 
> that was fine on 8.2, but now that's its been upgraded to 9.0 counts at 
> anything from 2 to 20 seconds per 5 second sample, the result is similar 
> with HPET, ACPI-fast and TSC.
> 
> I also have physical boxes which new seem to drift quite substantially, 
> ntpd cannot keep up and as these boxes need to be able to report the 
> time relatively accurately, it is causing problems with log times and 
> such...
> 
> Any suggestions most welcome!
> 
> Thanks,
> J

I finally got a 9.0 generic build done today and I've been watching the
timekeeping on 3 systems and they're all doing just fine.  Two of the
systems are performing pretty much identically to how they did on 8.2;
the clock frequency correction calculated by ntpd differs by less than
1ppm.  On the other system the kernel timekeeping routines are now
choosing to use a different clock so I don't get a direct comparison of
the old vs new drift rate, but the drift is still reasonable  (100ppm
now, used to be around 88, on an old 300mhz MediaGx-based system).

I haven't had time yet to learn about the new eventtimer stuff in 9.0,
but I know you can get some info on the choices it made via sysctl
kern.eventtimer.  Before 9.0 I'd check sysctl kern.clockrate and vmstat
-i and make sure the chosen clock is interrupting at the right rate, but
now with the eventtimer stuff there's not an obvious correlation between
hz and profhz and stathz and any particular device's interrupt rate, at
least for some clock choices (on the old MediaGx system without ACPI or
HPET it seems to work more like it used to).

-- Ian


___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Timekeeping in stable/9

2012-01-17 Thread Martin Sugioarto 
Am Tue, 17 Jan 2012 20:12:51 +
schrieb Joe Holden :

> Hi guys,
> 
> Has anyone else noticed the tendency for 9.0-R to be unable to 
> accurately keep time?  I've got a couple of machines that have been 
> upgraded from 8.2 that are struggling, in particular a Virtual box
> guest that was fine on 8.2, but now that's its been upgraded to 9.0
> counts at anything from 2 to 20 seconds per 5 second sample, the
> result is similar with HPET, ACPI-fast and TSC.

Hi Joe,

I can confirm this on VirtualBox. I've been running WinXP inside
VirtualBox and measured network I/O during downloads. It showed me very
high download rates (around 800kB/s) while it's physically possible to
download 200kB/s through DSL here (Germany sucks with DSL, even in
largest cities, btw!).

I correlated this behavior with high disk I/O on the host. That means
that the timer issues on the virtual host appear when I start a
larger cp job on the host. I also immediately thought that this has
something to do with timers.

You can perhaps try yourself and confirm, if it's the disk I/O that
influences timers. I somehow don't like the hard disk behavior. It
makes desktop unusable in some situations (mouse pointer skipping,
applications lock for several seconds).

> I also have physical boxes which new seem to drift quite
> substantially, ntpd cannot keep up and as these boxes need to be able
> to report the time relatively accurately, it is causing problems with
> log times and such...

Not sure about physical boxes. I have not taken a look at this, yet.

--
Martin


signature.asc
Description: PGP signature

Timekeeping in stable/9

2012-01-17 Thread Joe Holden 

Hi guys,

Has anyone else noticed the tendency for 9.0-R to be unable to 
accurately keep time?  I've got a couple of machines that have been 
upgraded from 8.2 that are struggling, in particular a Virtual box guest 
that was fine on 8.2, but now that's its been upgraded to 9.0 counts at 
anything from 2 to 20 seconds per 5 second sample, the result is similar 
with HPET, ACPI-fast and TSC.


I also have physical boxes which new seem to drift quite substantially, 
ntpd cannot keep up and as these boxes need to be able to report the 
time relatively accurately, it is causing problems with log times and 
such...


Any suggestions most welcome!

Thanks,
J
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"