Re: [Openstack] [OpenStack][Nova] Live-Migration ---Libvirt related problem

Razique Mahroua Mon, 28 May 2012 01:08:33 -0700

Hey Eric,
does from the two nodes a simple :
$ virsh list --all works ?

also, have you enabled started the TCP socket by enabling it into /etc/default/libvirt-bin

Before :
libvirtd_opts="-d"

After :
libvirtd_opts="-d -l"

Restart libvirtd and check the logs, by the way, you can safely ignore the errors about power management or install the following package :
dbus

You can also validate a manual migration by issuing the following command from the node 1 :
$ virsh migrate --live $domain (get by virsh list) qemu+tcp://user@node2/system

And check the output

Best regards,
Razique

Eric Luo

28 mai 2012 08:24
Hello ,all
first ,let me introduce my develop environment:
I have a computer with following lshw information

*-cpu
description: CPU
product: AMD FX(tm)-6100 Six-Core Processor
vendor: Hynix Semiconductor (Hyundai Electronics)
physical id: 4
bus info: cpu@0
version: AMD FX(tm)-6100 Six-Core Processor
slot: Socket M2
size: 1400MHz
capacity: 1400MHz
width: 64 bits
clock: 200MHz
*-memory
description: System Memory
physical id: 29
slot: System board or motherboard
size: 16GiB
Then ,I installed ubuntu1204 LTS .
Then,I started 3 VMs with virt-manager as my OpenStack testing cluster.I named the 3VMs as os0 ,os1 and os2.
I installed cloud controller on os0 ,and nova-compute on both os1 and os2.

I set up all the live-migration preparations instructed
here:
http://docs.openstack.org/trunk/openstack-compute/admin/content/configuring-live-migrations.html

and here:
https://help.ubuntu.com/community/SettingUpNFSHowTo

BTW:I found some error in the document ,

Update the libvirt configurations. Modify /etc/libvirt/libvirt.conf:

before : #listen_tls = 0
after : listen_tls = 0

before : #listen_tcp = 1
after : listen_tcp = 1

add: auth_tcp = "none"

here ,we should update the configuration file /etc/libvirt/libvirtd.conf ,not /etc/libvirt/libvirt.conf

After all are set ,I just use the command "nova live-migration vm1 os2",here the vm1 is an instance running currently on os1 and I wanted it to be migrated to os2

the command returned without any error or infor.I guess it indicated the migration is done correctly .But I check the vm1 with "nova show vm1",it is still running on os1!
I checked the nova-compte.log file on both os1 and os2,and found the following live_migration related entries :

2012-05-28 10:26:39 DEBUG nova.rpc.amqp [-] received {u'_context_roles': [u'admin'], u'_msg_id': u'89dfda33bb7b4dd29d93867aaf1460ec', u'_context_read_deleted':

u'no', u'_context_request_id': u'req-f1415f26-f19e-4499-aa24-22e47b586378', u'args': {u'instance_id': 15, u'block_migration': False, u'disk': None}, u'_context_

auth_token': '<SANITIZED>', u'_context_is_admin': True, u'_context_project_id': u'7fa3accbd2b543a2a59c5249b8d07281', u'_context_timestamp': u'2012-05-28T02:26:3

9.026426', u'_context_user_id': u'297ad61b9a67448186c08729ab0f1971', u'method': u'pre_live_migration', u'_context_remote_address': u'192.168.0.221'} from (pid=7

290) _safe_log /usr/lib/python2.7/dist-packages/nova/rpc/common.py:160

2012-05-28 11:22:09 DEBUG nova.service [-] live_migration_retry_count : 30 from (pid=16393) wait /usr/lib/python2.7/dist-packages/nova/service.py:411

2012-05-28 11:22:09 DEBUG nova.service [-] live_migration_uri : qemu+tcp://%s/system from (pid=16393) wait /usr/lib/python2.7/dist-packages/nova/service.py:411

2012-05-28 11:22:09 DEBUG nova.service [-] live_migration_bandwidth : 0 from (pid=16393) wait /usr/lib/python2.7/dist-packages/nova/service.py:411

I didn't see any error in these log files.

Then I checked the /var/log/libvirt/libvirtd.log on both os1 and os2.and found the following.

2012-05-28 02:56:14.310+0000: 9481: error : virExecWithHook:328 : Cannot find 'pm-is-supported' in path: No such file or directory
2012-05-28 02:56:14.310+0000: 9481: warning : qemuCapsInit:856 : Failed to get host power management capabilities

2012-05-28 02:56:14.514+0000: 9480: error : virExecWithHook:328 : Cannot find 'pm-is-supported' in path: No such file or directory
2012-05-28 02:56:14.515+0000: 9480: warning : qemuCapsInit:856 : Failed to get host power management capabilities

2012-05-28 02:56:14.719+0000: 9482: error : virExecWithHook:328 : Cannot find 'pm-is-supported' in path: No such file or directory
2012-05-28 02:56:14.719+0000: 9482: warning : qemuCapsInit:856 : Failed to get host power management capabilities

2012-05-28 02:56:14.943+0000: 9484: error : virExecWithHook:328 : Cannot find 'pm-is-supported' in path: No such file or directory
2012-05-28 02:56:14.944+0000: 9484: warning : qemuCapsInit:856 : Failed to get host power management capabilities

2012-05-28 02:56:15.155+0000: 9481: error : virExecWithHook:328 : Cannot find 'pm-is-supported' in path: No such file or directory
2012-05-28 02:56:15.155+0000: 9481: warning : qemuCapsInit:856 : Failed to get host power management capabilities

2012-05-28 02:58:55.330+0000: 13514: info : libvirt version: 0.9.8
2012-05-28 02:58:55.330+0000: 13514: error : qemuMonitorIORead:513 : Unable to read from monitor: Connection reset by peer

2012-05-28 03:02:15.618+0000: 13518: error : qemuMonitorOpenUnix:295 : failed to connect to monitor socket: No such process
2012-05-28 03:02:15.618+0000: 13518: error : qemuProcessWaitForMonitor:1301 : internal error process exited while connecting to monitor: char device re

to /dev/pts/3
inet_listen_opts: bind(ipv4,192.168.0.205,5900): Cannot assign requested address

inet_listen_opts: FAILED

Please note the above ,the last two lines I marked as red ,I googled the error ,and do found something maybe related here :

https://bugzilla.redhat.com/show_bug.cgi?format=multiple&id=703851

I know the inet_listen_opts: bind error maybe related to my VNC settings in /etc/nova/nova.conf,So I just commented all the vnc related lines in /etc/nova/nova.conf

Then I run the
"nova live-migration vm1 os2" again ,and "tail -f /var/log/nova/nova-compute.log" to both os1 and os2.I found the the folloing:

2012-05-28 03:21:02.700+0000: 8872: error : virNetSocketReadWire:996 : End of file while reading data: Input/output error

2012-05-28 03:21:02.700+0000: 8872: error : virNetSocketReadWire:996 : End of file while reading data: Input/output error
2012-05-28 03:23:06.551+0000: 8879: error : virNetDevGetIndex:656 : Unable to get index for interface vnet0: No such device

2012-05-28 03:24:36.992+0000: 8875: error : virNetClientProgramDispatchError:174 : Unable to read from monitor: Connection reset by peer

note that the previous inet_listen_opts:bind error is not showing this time ,but I got something new as above.

Well ,these are all problems I met.And here is my question :

1.Are these libvirt related problems showing only because I am testing them on a "virtual "environment? I mean ,maybe I should buy some more physical computer?
2.If not ,what should I do with this ? Did anyone else get the same problem as I did?.

Please help me ,thank you very much in advance!

Eric Luo
2012-05-28

_______________________________________________
Mailing list: https://launchpad.net/~openstack
Post to : [email protected]
Unsubscribe : https://launchpad.net/~openstack
More help : https://help.launchpad.net/ListHelp

--
Nuage & Co - Razique Mahroua
[email protected]

_______________________________________________
Mailing list: https://launchpad.net/~openstack
Post to     : [email protected]
Unsubscribe : https://launchpad.net/~openstack
More help   : https://help.launchpad.net/ListHelp

Re: [Openstack] [OpenStack][Nova] Live-Migration ---Libvirt related problem

Reply via email to