Open Stack

Hi Wenmao,

> Hi all:
> 
> I upgrade Openstack to Icehouse and find that there is a new horizon feature in which VM can be live migrated. So I tried to move a VM from node2 to node1, but got a "Error: Failed to live
> migrate instance to host "node1.expr-zone"." error.
> 
> There seems no error logs in nova compute services of node1 or node2, the nova api and nova schduler on controller does not give any hints either. The only error is found  in apache error
> log:
> 
> [Tue May 06 08:18:32 2014] [error] REQ: curl -i 'http://controller:8774/v2/53707d290204404dbff625378969c25c/os-hypervisors/detail' -X GET -H "X-Auth-Project-Id:
> 53707d290204404dbff625378969c25c" -H "User-Agent: python-novaclient" -H "Accept: application/json" -H "X-Auth-Token: 6e3dcd16941b44a6aa19b07c64a4374f"
> [Tue May 06 08:18:32 2014] [error]
> [Tue May 06 08:18:32 2014] [error] RESP: [200] CaseInsensitiveDict({'date': 'Tue, 06 May 2014 08:18:32 GMT', 'content-length': '1561', 'content-type': 'application/json',
> 'x-compute-request-id': 'req-75b0e3d3-4f66-459d-a54f-b35163c6a730'})
> [Tue May 06 08:18:32 2014] [error] RESP BODY: {"hypervisors": [{"service": {"host": "node1", "id": 6}, "vcpus_used": 12, "hypervisor_type": "QEMU", "local_gb_used": 240, "host_ip":
> "20.0.0.11", "hypervisor_hostname": "node1.expr-zone", "memory_mb_used": 25088, "memory_mb": 128898, "current_workload": 0, "vcpus": 24, "cpu_info": "{\\"vendor\\": \\"Intel\\",
> \\"model\\": \\"SandyBridge\\", \\"arch\\": \\"x86_64\\", \\"features\\": [\\"pdpe1gb\\", \\"osxsave\\", \\"dca\\", \\"pcid\\", \\"pdcm\\", \\"xtpr\\", \\"tm2\\", \\"est\\", \\"smx\\",
> \\"vmx\\", \\"ds_cpl\\", \\"monitor\\", \\"dtes64\\", \\"pbe\\", \\"tm\\", \\"ht\\", \\"ss\\", \\"acpi\\", \\"ds\\", \\"vme\\"], \\"topology\\": {\\"cores\\": 6, \\"threads\\": 2,
> \\"sockets\\": 1}}", "running_vms": 3, "free_disk_gb": 1728, "hypervisor_version": 1005000, "disk_available_least": 1396, "local_gb": 1968, "free_ram_mb": 103810, "id": 1}, {"service":
> {"host": "node2", "id": 8}, "vcpus_used": 13, "hypervisor_type": "QEMU", "local_gb_used": 222, "host_ip": "20.0.0.12", "hypervisor_hostname": "node2.expr-zone", "memory_mb_used": 24064,
> "memory_mb": 128909, "current_workload": 0, "vcpus": 24, "cpu_info": "{\\"vendor\\": \\"Intel\\", \\"model\\": \\"SandyBridge\\", \\"arch\\": \\"x86_64\\", \\"features\\": [\\"pdpe1gb\\",
> \\"osxsave\\", \\"dca\\", \\"pcid\\", \\"pdcm\\", \\"xtpr\\", \\"tm2\\", \\"est\\", \\"smx\\", \\"vmx\\", \\"ds_cpl\\", \\"monitor\\", \\"dtes64\\", \\"pbe\\", \\"tm\\", \\"ht\\", \\"ss\\",
> \\"acpi\\", \\"ds\\", \\"vme\\"], \\"topology\\": {\\"cores\\": 6, \\"threads\\": 2, \\"sockets\\": 1}}", "running_vms": 7, "free_disk_gb": 1746, "hypervisor_version": 1005000,
> "disk_available_least": 1398, "local_gb": 1968, "free_ram_mb": 104845, "id": 2}]}
> [Tue May 06 08:18:32 2014] [error]
> [Tue May 06 08:18:32 2014] [error]
> [Tue May 06 08:18:32 2014] [error] REQ: curl -i 'http://controller:8774/v2/53707d290204404dbff625378969c25c/servers/f1df31ff-6e42-44ed-8697-1bd0dd708762/action' -X POST -H
> "X-Auth-Project-Id: 53707d290204404dbff625378969c25c" -H "User-Agent: python-novaclient" -H "Content-Type: application/json" -H "Accept: application/json" -H "X-Auth-Token:
> 6e3dcd16941b44a6aa19b07c64a4374f" -d '{"os-migrateLive": {"disk_over_commit": false, "block_migration": false, "host": "node1.expr-zone"}}'
> [Tue May 06 08:18:32 2014] [error]
> [Tue May 06 08:18:33 2014] [error] RESP: [400] CaseInsensitiveDict({'date': 'Tue, 06 May 2014 08:18:33 GMT', 'content-length': '107', 'content-type': 'application/json; charset=UTF-8',
> 'x-compute-request-id': 'req-ede459ef-cea4-42d3-a212-abe63021c05d'})
> [Tue May 06 08:18:33 2014] [error] RESP BODY: {"badRequest": {"message": "Compute service of node1.expr-zone is unavailable at this time.", "code": 400}}
> [Tue May 06 08:18:33 2014] [error]
> [Tue May 06 08:18:33 2014] [error] Recoverable error: Compute service of node1.expr-zone is unavailable at this time. (HTTP 400) (Request-ID: req-ede459ef-cea4-42d3-a212-abe63021c05d)
> 
> 
> the services seems to be up :
> root at controller:/usr/src/horizon# nova-manage service list
> Binary           Host                                 Zone             Status     State Updated_At
> nova-scheduler   controller                           internal         enabled    :-)   2014-05-06 08:28:57
> nova-cert        controller                           internal         enabled    :-)   2014-05-06 08:28:57
> nova-consoleauth controller                           internal         enabled    :-)   2014-05-06 08:28:57
> nova-conductor   controller                           internal         enabled    :-)   2014-05-06 08:28:57
> nova-compute     node1                                nova             enabled    :-)   2014-05-06 08:28:52
> nova-conductor   node1                                internal         enabled    :-)   2014-05-06 08:28:37
> nova-compute     node2                                nova             enabled    :-)   2014-05-06 08:28:58
> nova-conductor   node2                                internal         enabled    :-)   2014-05-06 08:29:14
> 
> So how can I solve this problem?
> 
> Thanks
> 
> Wenmao Liu
> NSFOCUS Inc
> 
>

I haven't tried Icehouse, but I have block live migration working on 
Havana, but to set this up, you have to edit some files to allow libvirt 
to do the work, this post seems to sum up what is needed and it works 
reliable for me.

http://kimizhang.wordpress.com/2013/08/26/openstack-vm-live-migration/

However, the migation is all done from the command line, not through 
Horizon.

Regards,
Andrew.




-- 
Windows, for life without walls.
Linux, for life without Windows.