Re: XenServer 6.5(SP1) - HVM 're0: watchdog timeout' errors...

2015-09-07 Thread Adam McDougall 
On 09/07/2015 11:00, Karl Pielorz wrote:
> 
> --On 07 September 2015 16:51 +0200 Roger Pau Monné
>  wrote:
> 
>> This sounds like a problem in the if_re driver on FreeBSD or a bug in
>> Qemu emulation. As a workaround, could you try to use the emulated e1000
>> instead of the realtek? AFAIK it should also provide better performance.
>>
>> Since you are using XenServer I'm not sure of the exact rune, with OSS
>> Xen you basically need to add "model=e1000" to the network configuration
>> line.
> 
> Hi Roger,
> 
> I'd love to try that - but as we're running full 'XenServer' (i.e.
> installed from XenServer .ISO) I have no idea where (or even if it's
> possible) to set that :(
> 
> Last time I tried looking it up I hit a bit of a dead end, I'll have
> another look [in fact, I think it may have actually been impossible to
> set in XenServer as I seem to remember someone saying that XenServer
> doesn't have that NIC model/emulation available, only Realtek or PV!]
> 
> But I'll have another look...
> 
> -Karl

Try this:
http://discussions.citrix.com/topic/329848-openbsd-with-message-watchdog-timeout/

I've used it before for an OpenBSD guest VM.  If that patch is in place
on any XenServer you migrate to, it should set e1000 for the NIC type if
you set NicEmulation in the Custom Fields as described in step 2.

A cleaner patch would just let you specify any driver you want, but this
one is semi-hardcoded for e1000.  You could also try 'virtio' with:
argv = [arg.replace('rtl8139','virtio') for arg in argv]
___
freebsd-xen@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Re: Boot FreeBSD 10 in HVM mode under Xen?

2015-07-31 Thread Adam McDougall 
On 07/31/2015 11:22, Karl Pielorz wrote:
> 
> --On 31 July 2015 17:09 +0200 Roger Pau Monné  wrote:
> 
>> Ouch, I guess this is the Xen HyperV support badly interacting with
>> FreeBSD again. XenServer enables the "viridian" extensions by default,
>> which really messes up with FreeBSD. You should be able to disable the
>> extensions by hand, as reported in:
>>
>> https://lists.freebsd.org/pipermail/freebsd-xen/2013-October/001756.html
> 
> Fantastic - that works!
> 
> Cheers,
> 
> -Karl

I'm guessing my kernel boots fine probably because I've disabled hyperv
support in my kernel config.  viridian probably explains why I see this
in dmesg though:
Hypervisor: Origin = "Microsoft Hv"

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

Re: Boot FreeBSD 10 in HVM mode under Xen?

2015-07-31 Thread Adam McDougall 
On 07/31/2015 09:47, Karl Pielorz wrote:
> 
> Hi,
> 
> Is there any way of booting FreeBSD 10.x as HVM under Xen?
> 
> I've tried removing,
> 
>   options XENHVM
>   device  xenpci
> 
>> From the kernel - but this just panics on boot now.
> 
>  seems to suggest you can no
> longer boot 10 under Xen as HVM?
> 
> If I can't boot HVM, I can't get HVM network support (i.e. Realtek) -
> and I can't get that, all the VM's I have to run as HVM due to a number
> of xen/freebsd/network oddities can't be 10.x machines :(
> 
> This is painting those roles/systems into non-agile HVM 9.x boxes that
> have no upgrade path now :(
> 
> Unless there's a workaround to ditch xn and replace it with re and / or
> boot HVM under FreeBSD 10?
> 
> -Karl

I use a custom kernel config containing:

include MYNORMALCONFIGFILE
ident   SPECIALCONFIG

nooptions   XENHVM
nodevicexenpci

The system running that is 10.1-STABLE #0 r283511 from May 24 2015.  It
previously ran an earlier build, probably from March or Jan.  re0 is
available with this config.

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

Re: XenServer 6.5 migrate FBSD 10/11 results in clock reset to 1970?

2015-01-14 Thread Adam McDougall 
On 01/14/2015 17:46, Karl Pielorz wrote:
> 
> --On 14 January 2015 15:48:47 -0500 Adam McDougall
>  wrote:
> 
>> I really thought this was fixed in Creedence alphas that I tested... I
>> still have those test systems up but I need to make a test environment
>> for 6.5.  Are you using AMD or Intel?  I don't know if it makes a
>> difference but I've only seen the problem on AMD so far.
> 
> We're looking at getting a Creedence test setup soon, so I can re-test
> then - at the moment our platform is XenServer 6.5 - with HP Proliant
> DL320e's (all Intel / Haswell based).
> 
> I'll check tomorrow to see how far off having a Creedence test setup
> will be.
> 
> -Karl

Are you sure you are running 6.5 and not 6.2?  The guest's kernel should
detect Xen 4.4 instead of 4.1.  6.5 is Creedence but it was called
Creedence during the long alpha stage before it became 6.5.
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Re: XenServer 6.5 migrate FBSD 10/11 results in clock reset to 1970?

2015-01-14 Thread Adam McDougall 
On 01/14/2015 15:42, Mark Felder wrote:
> 
> 
> On Wed, Jan 14, 2015, at 09:16, Karl Pielorz wrote:
>>
>> Hi,
>>
>> This has been seen before e.g. see
>>
>>  
>>
>> We're now seeing this now we've started using 10.x boxes under XenServer
>> 6.5
>>
>> Is there any work around for it?
>>
>> The 'hit' rate for us seems to be quite high (+70%?) i.e. the clock
>> resets 
>> most of the time we do even just a storage migration.
>>
>> The guests are running NTP - and that continues running after the event, 
>> but  is obviously unwilling to make such a big clock adjustment to drag
>> the 
>> guests time from 1970 to present day.
>>
>> Unfortunately - this also causes various other things on the box to break 
>> as well :(
>>
>> Is there no 'after migration' hook or script I can lodge some code to 
>> shutdown NTP, do an ntpdate - then restart NTP again?
>>
> 
> When I ran into this I manually stopped NTP, migrated, ran ntpdate,
> started NTP again. It was painful.
> 
> The problem as I recall is in the PVHVM code and is fixed upstream in
> Xen but wasn't pulled into XenServer. Roger will know more details, but
> if you have a Citrix support contract you should pressure them to open a
> bug / regression on this. Unfortunately I'm no longer in a position to
> do so...
> ___
> freebsd-xen@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-xen
> To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"
> 

I really thought this was fixed in Creedence alphas that I tested... I
still have those test systems up but I need to make a test environment
for 6.5.  Are you using AMD or Intel?  I don't know if it makes a
difference but I've only seen the problem on AMD so far.
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Re: [HEADS UP] FreeBSD guest integrated into Xen push gate tests

2013-12-07 Thread Adam McDougall 

On 12/07/2013 06:01, Roger Pau Monné wrote:

Hello,

Today we had the first successful run of the Xen push gate tester
(OSSTest) containing a FreeBSD PVHVM guest. This means that from now
onwards every commit on the Xen repository will be tested against a
FreeBSD PVHVM guest (i386 and amd64), to make sure new changes in Xen
doesn't break current support in FreeBSD. You can see the results of the
first run at:

http://www.chiark.greenend.org.uk/~xensrcts/logs/22323/

The jobs are test-amd64-i386-freebsd10-amd64 and
test-amd64-i386-freebsd10-i386, you can see what's being tested on the
results matrix. This guests are created using the official FreeBSD VM
images, found at:

ftp://ftp.freebsd.org/pub/FreeBSD/snapshots/VM-IMAGES/10.0-BETA3/

Right now it's using the BETA3 images, but I plan to update that once
10.0 is released.

A very big thanks goes to Ian Jackson for integrating my crappy FreeBSD
install script into OSSTest and for solving the various problems that
arose during the integration.

Roger.


Thanks for the work everyone put in!
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Re: FreeBSD 10-BETA4 r258899 PVHVM Xen clock -> 1970 on VM migration

2013-12-06 Thread Adam McDougall 
On Fri, Dec 06, 2013 at 06:36:17PM +0100, Roger Pau Monné wrote:

  On 06/12/13 18:14, Adam McDougall wrote:
  > Hello,
  > 
  > I noticed on recent (PVHVM) builds of FreeBSD 10 that I randomly lose the
  > proper date/time when I migrate my VM from one XenServer to another in the
  > pool.  I was intending to reproduce a much smaller time difference I've seen
  > on 9 but this was a much bigger jump so I started concentrating on the big
  > jumps.  When I migrate my VM back and forth between two servers, it will
  > randomly jump to a date in 1970 or 1969.  Continued migrations often bring
  > the proper date/time back, but there is not a strong pattern.  The 1970
  > comes and goes.  I could not reproduce this kind of jump on 9.  This happens
  > on XenServer 6.0 and 6.2.
  
  Hello,
  
  Thanks for testing, there were some bugs in Xen regarding the migration
  of PVHVM guests, the one that comes to mind, and that could cause this
  kind of time disruption is:
  
  http://lists.xenproject.org/archives/html/xen-devel/2013-06/msg01110.html
  
  The patch should be present in at least Xen 4.3, not sure about previous
  versions. Is there anyway you could try this on a recent Xen version (4.3)?

It seems fine in 4.3, I installed the 2013-12-05 ISO on 
http://www.xenserver.org/overview-xenserver-open-source-virtualization/download/2-uncategorised/115-development-snapshots.html
which identifies itself as 6.2.50.  I migrated a VM about 8 times.  It did
become off by a couple seconds but ntpd should fix that up.
  
  > I will take a stab at trying PVH time permitting,
  > but since 10.0 will ship with PVHVM it seems appropriate to try to fix it
  > if possible.  Let me know if there is more testing or information gathering
  > I should do.  Thanks.
  
  Xen 4.4 will be the first version to support PVH, which has not yet been
  released, and right now doesn't support migration. If you could try a
  recent version of Xen that contains the mentioned patch it will
  certainly help diagnose it (or at least make sure it's caused by this
  issue).
  
  Roger.
  
Sorry, I missed the part in your original email about needing Dom0 changes for
PVH.
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Re: FreeBSD 10-BETA4 r258899 PVHVM Xen clock -> 1970 on VM migration

2013-12-06 Thread Adam McDougall 
On Fri, Dec 06, 2013 at 12:14:07PM -0500, Adam McDougall wrote:

  Hello,
  
  I noticed on recent (PVHVM) builds of FreeBSD 10 that I randomly lose the
  proper date/time when I migrate my VM from one XenServer to another in the
  pool.  I was intending to reproduce a much smaller time difference I've seen
  on 9 but this was a much bigger jump so I started concentrating on the big
  jumps.  When I migrate my VM back and forth between two servers, it will
  randomly jump to a date in 1970 or 1969.  Continued migrations often bring
  the proper date/time back, but there is not a strong pattern.  The 1970
  comes and goes.  I could not reproduce this kind of jump on 9.  This happens
  on XenServer 6.0 and 6.2.  I will take a stab at trying PVH time permitting,
  but since 10.0 will ship with PVHVM it seems appropriate to try to fix it
  if possible.  Let me know if there is more testing or information gathering
  I should do.  Thanks.

The same happens with PVH it seems:
FreeBSD pvhvm 11.0-CURRENT FreeBSD 11.0-CURRENT #0 2740cfb(master)

Maybe as the VM is resuming after migration it can sync clock with the Dom0?
Some of my servers run daemons that are sensitive to abrupt time changes 
(dovecot)
which will purposely stop running if the time is altered by more than 60 
seconds.
Even on 9 I've seen the time jump by a few minutes or more after a migration.
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

FreeBSD 10-BETA4 r258899 PVHVM Xen clock -> 1970 on VM migration

2013-12-06 Thread Adam McDougall 
Hello,

I noticed on recent (PVHVM) builds of FreeBSD 10 that I randomly lose the
proper date/time when I migrate my VM from one XenServer to another in the
pool.  I was intending to reproduce a much smaller time difference I've seen
on 9 but this was a much bigger jump so I started concentrating on the big
jumps.  When I migrate my VM back and forth between two servers, it will
randomly jump to a date in 1970 or 1969.  Continued migrations often bring
the proper date/time back, but there is not a strong pattern.  The 1970
comes and goes.  I could not reproduce this kind of jump on 9.  This happens
on XenServer 6.0 and 6.2.  I will take a stab at trying PVH time permitting,
but since 10.0 will ship with PVHVM it seems appropriate to try to fix it
if possible.  Let me know if there is more testing or information gathering
I should do.  Thanks.
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Re: Current problem reports assigned to freebsd-xen@FreeBSD.org

2013-09-03 Thread Adam McDougall 
On 09/03/13 10:09, Justin T. Gibbs wrote:
> Has anyone made any progress analyzing these bugs?  I'm wondering if there are
> bugs here we can easily fix for 10 or that we've already fixed and can be 
> closed.
> 
>> S Tracker  Resp.  Description
>> 
>> o kern/176471  xen[xen] xn driver crash on detach
Can be closed, patch was committed and MFC'ed to 9, details in PR.

>> o kern/171118  xen[xen] FreeBSD XENHVM guest doesn't shutdown cleanly
Appears to still be relevant from a test just now on 9.2-BETA1
(AMD64-9) r253631.

>> o kern/154428  xen[xen] xn0 network interface and PF - Massive 
>> performan
I had similar issues, some testing I did last week makes me satisfied
with r252781 in 9-stable (MFC of r291296, r291297, r291393) but I can't
speak for everyone.

>> f kern/135178  xen[xen] Xen domU outgoing data transfer stall when 
>> TSO i
Looks like feedback timeout.
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Re: XenServer 6.2 tools for FreeBSD 9 'Guest'?

2013-08-28 Thread Adam McDougall 
On 08/28/13 11:31, Karl Pielorz wrote:
> 
> Hi,
> 
> We've got XenServer now running a couple of FreeBSD 9.2-RC2 VM's
> (running the XENHVM kernel).
> 
> Is there any up to date 'xen tools' setup guide or anything anywhere we
> can look at?
> 
> I've seen 'xen-tools' in the ports collection - but once installed it
> doesn't seem to do a lot. I'm looking to enable things like suspending
> VM's and migration etc.
> 
> Any pointers? - Google turns up a lot of stuff, but mostly for much
> older versions of Xen & FreeBSD.
> 
> -Karl

Try the xe-guest-utilities which uses xen-tools.  You may have to hack
/usr/local/sbin/xe-update-guest-attrs near the bottom to provide a
MinorVersion and/or MicroVersion higher than the default to get the
hypervisor to accept it as recent enough.  For example under a 6.0.2
environment you could try:

xenstore_write_cached "attr/PVAddons/MajorVersion" "6"
xenstore_write_cached "attr/PVAddons/MinorVersion" "0"
xenstore_write_cached "attr/PVAddons/MicroVersion" "99"

Once you enable and run /usr/local/etc/rc.d/xenguest you should be in
business.
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Re: XENHVM and XenServer 6.2

2013-08-11 Thread Adam McDougall 

On 08/11/13 05:30, Karl Pielorz wrote:


--On 10 August 2013 19:01:38 +0400 Alexey Degtyarev 
 wrote:



Try to remove DVD device from problem virtual machine:

# xe vm-list params=uuid name-label=

# xe vbd-list empty=true params=uuid vm-uuid=

# xe vbd-destroy uuid=


Presumably / obviously removing the DVD will stop you booting a .ISO 
image? - I just built a 9.1-Beta2 release .ISO image w/XENHVM kernel - 
and ran into the same issue booting that .ISO :(


Can you boot a GENERIC .ISO, use that to do the install, then switch 
to XENHVM kernel (i.e. once installed and you don't need to boot off 
of ISO)? [first time with XEN]



-Karl

For now, yes, but there was talk of enabling XENHVM in GENERIC which would
make this really hard without a boot loader switch to turn XENHVM off.
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Re: svn commit: r251297 - head/sys/dev/xen/netfront

2013-06-07 Thread Adam McDougall 
On Mon, Jun 03, 2013 at 02:06:57PM -0700, Colin Percival wrote:

  On 06/03/13 06:00, Andre Oppermann wrote:
  > Log:
  >   Specify a maximum TSO length limiting the segment chain to what the
  >   Xen host side can handle after defragmentation.
  >   
  >   This prevents the driver from throwing away too long TSO chains and
  >   improves the performance on Amazon AWS instances with 10GigE virtual
  >   interfaces to the normally expected throughput.
  
  Thanks!
  
  For the benefit of the list: This (and the r251296, which provides the
  network stack infrastructure which this commit uses) was the last of the
  "make FreeBSD behave well in Amazon EC2" patches I've been carrying around
  (and building into my EC2 images) for the past two years.
  
  Now the only step remaining to make FreeBSD/EC2 pure "straight off the
  RELEASE ISO" FreeBSD is to get Xen HVM functionality into GENERIC.
  
  -- 
  Colin Percival
  Security Officer Emeritus, FreeBSD | The power to serve
  Founder, Tarsnap | www.tarsnap.com | Online backups for the truly paranoid
  
Is it possible to turn it into a boot time tunable in case XENHVM puts
a system at a disadvantage?  I'm specifically thinking of some XENHVM 
environments
under XenServer and related products that aren't currently compat with having
a virtual cdrom due to a bug.  Removing the virtual cdrom drive is a workaround
but that makes it a lot harder to boot from a RELEASE ISO :)

Apologies if it already has a tunable, but I did some searching and could not
locate it.  It was desirable too back when a XENHVM kernel would panic on bare 
hw
but that was fixed a while ago.
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Re: kern/176471: [xen] xn driver crash on detach

2013-05-21 Thread Adam McDougall 
The following reply was made to PR kern/176471; it has been noted by GNATS.

From: Adam McDougall 
To: bug-follo...@freebsd.org, p...@helenius.fi
Cc:  
Subject: Re: kern/176471: [xen] xn driver crash on detach
Date: Tue, 21 May 2013 09:57:06 -0400

 This patch is working well for me as well, using it on 9.1-stable.  Can
 we get it committed?  Thanks.
___
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"