Hello list,
Thank you for reading about my problem. I have been testing these new windows drivers and have a strange issue. I have built new Windows guest VM's of 2008 and 2012. Both Windows 2008 and 2012 guest VM's become unresponsive after a few days of essentially idling doing nothing. I have confirmed this a few weeks in a row checking each Monday after a reboot. Both become nonresponsive as for RDP and on VNC console.
My host server:
centos6.8 (fully updated)
xen_version: 4.6.1-11.el6
Xen is built via SolusVM slave script (http://files.soluslabs.com/install.sh)
Testing (2) Windows guests with the new drivers:
Windows 2008 R2 (fully updated) with pv 8.1 drivers
Windows 2012 R1 (fully updated) with pv 8.1 drivers
High-performance power option is set in Windows for each guest.
They do both respond to ping tests interestingly when they are in this problem state so not fully locked up but they will not give any console access. 2008 console is visible but does not respond to Ctrl-Alt-Del though the mouse moves. 2012 console goes black and does not respond at all.
The same production Xen server is running other Windows only guests on the previous drivers (ejbpv_Vista2008x64_1.0.1105.msi) without this issue. The Xen server has plenty of resources only running about 16% CPU and 32% memory usage. Only 15 VM's are currently running on this host.
The problem is repeatable having tested for few weeks confirming the problem.
I tried disabling all of the network interface offloading and rebooted as a test and it made no difference. This problem still happens.
I have installed all 5 driver packages of version 8.1 in each Windows guest.
Has anyone else had anything like this happen?
I'm lost as for what it could be. I have a large amount of Xen servers and Xen guests running the previous ejb drivers fine including on the same host server as these two test VM's.
I appreciate any suggestions if anyone has a suspicion what might be happening.
