Date: Fri, 6 Jan 2017 11:18:46 +0000 From: 'Roger Pau =?iso-8859-1?Q?Monn=E9'?= <roger.pau@citrix.com> To: Jay West <jwest@ezwind.net> Cc: "'Hoyer-Reuther, Christian'" <Christian.Hoyer-Reuther@cac-chem.de>, <freebsd-xen@freebsd.org>, <owner-freebsd-xen@freebsd.org> Subject: Re: 11-RELEASE and live migration Message-ID: <20170106111846.qnvrq6vwvdi3245y@dhcp-3-221.uk.xensource.com> In-Reply-To: <001701d2679d$f99eb560$ecdc2020$@ezwind.net> References: <001f01d250a6$05ca4160$115ec420$@ezwind.net> <a8486e84c86c73ae1a7d68a06d969671@ultra-secure.de> <003501d250af$032dd580$09898080$@ezwind.net> <41E487BC91654544B2B8F31096F2D9D4D1B060A925@ex1> <20161212103148.5prkggi3ix4zoxyv@dhcp-3-221.uk.xensource.com> <41E487BC91654544B2B8F31096F2D9D4D1B060AA3E@ex1> <002101d25484$c0a068c0$41e13a40$@ezwind.net> <20161214140224.rbg5h3v2opa3q3sj@dhcp-3-221.uk.xensource.com> <41E487BC91654544B2B8F31096F2D9D4D1B060ABBF@ex1> <001701d2679d$f99eb560$ecdc2020$@ezwind.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Jan 05, 2017 at 03:52:13PM -0600, Jay West wrote: > > Roger wrote... > > I'm able to see these glitches, as said the VM doesn't really hang, it > > just gets stuck for a long time (and I think that depends on the > > uptime differences between the source and destination hosts). In the > > meantime, could you please test if changing the timecounter from > > XENTIMER to any other (like HPET or > > ACPI-fast) solves the issue? > > > > # sysctl -w kern.timecounter.hardware=ACPI-fast > > Christian replied... > I changed the timecounter from XENTIMER to ACPI-fast and then I did about 10 > migrations between our 3 hosts. I always started the next migration a few > seconds after the previous migration finished. The VM didn't stuck now and I > didn't see the VGABios screen on the console. > > I can also add to this... changing timecounter to ACPI-fast definitely fixes > the problem. I can migrate a handful of freebsd11 vm's back and forth and > all stay running correctly. > > So... bug in XENTIMER? Not exactly, let's say it's a shortcoming of the current resume procedure. IF you want to know more about the gory details: https://lists.freebsd.org/pipermail/freebsd-arch/2016-December/018041.html I'm currently quite busy, so I don't really have an ETA of when this is going to be fixed. Roger.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20170106111846.qnvrq6vwvdi3245y>