<div dir="ltr"><div class="gmail_default" style="font-family:'courier new',monospace">I have a Grizzly system and I'm trying to figure out why VM's aren't able to be migrated (using "nova live-migration $UUID $NODE") from one node to another. The error message on the node that it is being migrated from is:</div><div class="gmail_default" style="font-family:'courier new',monospace"><br></div><div class="gmail_default" style="font-family:'courier new',monospace"><div class="gmail_default"> ERROR nova.virt.libvirt.driver [-] [instance: 0522b23c-5c2d-4c45-a66b-24c4c3f4ba9c] Live Migration failure: internal error process exited while connecting to monitor: W: kvm binary is deprecated, please use qemu-system-x86_64 instead</div><div><br></div></div><div class="gmail_default" style="font-family:'courier new',monospace">There is no message on the node that it is supposed to be migrating to.</div><div class="gmail_default" style="font-family:'courier new',monospace"><br></div><div class="gmail_default" style="font-family:'courier new',monospace">It was working fine for a while and then it fails. While investigating, I see AUDIT messages in nova-compute.log:</div><div class="gmail_default" style><div class="gmail_default" style><font face="courier new, monospace"><br></font></div><div class="gmail_default" style><font face="courier new, monospace">2014-09-23 18:00:45.274 4608 AUDIT nova.compute.resource_tracker [-] Free ram (MB): 80797</font></div><div class="gmail_default" style><font face="courier new, monospace">2014-09-23 18:00:45.274 4608 AUDIT nova.compute.resource_tracker [-] Free disk (GB): 31369</font></div><div class="gmail_default" style><font face="courier new, monospace">2014-09-23 18:00:45.274 4608 AUDIT nova.compute.resource_tracker [-] Free VCPUS: -30</font></div><div style="font-family:'courier new',monospace"><br></div><div style="font-family:'courier new',monospace">The system has 16 cores and has "cpu_allocation_ratio=8.0" in nova.conf so it should have a capacity of 128 VCPU's (right?). Checking with nova-manage:</div><div style="font-family:'courier new',monospace"><br></div><div style="font-family:'courier new',monospace">root@compute3:~# nova-manage service describe_resources test3</div><div style="font-family:'courier new',monospace"><div>HOST PROJECT cpu mem(mb) hdd</div><div>test3 (total) 16 128925 32330</div><div>test3 (used_now) 46 48128 960</div><div>test3 (used_max) 46 47104 960</div><div>.</div><div>.</div><div>.</div><div><br></div><div>It looks like it is calculating Free VCPUS by subtracting "used_now" from "total": 16 - 46 = -30. Is it somehow using this to decide that the node should not take more VM's? If so, I don't know why it allowed it to get as low as -30. </div><div><br></div><div>Can anyone explain what is going on? Is there other information I can look at to diagnose why the live-migration is failing?</div><div><br></div><div>Thanks a lot,</div><div><br></div><div>Steve</div></div></div><div><br></div>-- <br><span style="font-family:'courier new',monospace">________________________________________________________________</span><br style="font-family:'courier new',monospace"><span style="font-family:'courier new',monospace"> Steve Cousins Supercomputer Engineer/Administrator</span><br style="font-family:'courier new',monospace"><span style="font-family:'courier new',monospace"> Advanced Computing Group University of Maine System</span><br style="font-family:'courier new',monospace"><span style="font-family:'courier new',monospace"> 244 Neville Hall (UMS Data Center) (207) 561-3574</span><br style="font-family:'courier new',monospace"><span style="font-family:'courier new',monospace"> Orono ME 04469 steve.cousins at <a href="http://maine.edu" target="_blank">maine.edu</a></span><br><br>
</div>