1. Download nimbula director documents here: nimbula_documents_ppt.rar
2. And of course you can get more on nimbula director site http://nimbula.com/resources/documentation/public/ungated/

To check how much cpu one vm is consuming, we can use xentop for this analyzing:

[test@test ~]# xentop -b -i 2 -d 1
NAME STATE CPU(sec) CPU(%) MEM(k) MEM(%) MAXMEM(k) MAXMEM(%) VCPUS NETS NETTX(k) NETRX(k) VBDS VBD_OO VBD_RD VBD_WR SSID
11572_test_0106_us_oracle_com –b— 8412 0.0 34603008 34.4 34603008 34.4 6 2 196796 1111779 2 90 37651 3174172 0
16026_test_0093_us_oracle_com –b— 4255 0.0 1048576 1.0 1048576 1.0 2 2 2092803 2914101 3 851 49446 1918010 0
16051_test_0094_us_oracle_com —–r 3636909 0.0 56623104 56.3 56623104 56.3 24 2 1553871 970055 2 417 101921 10195220 0
Domain-0 —–r 36197 0.0 2621440 2.6 no limit n/a 24 0 0 0 0 0 0 0 0
NAME STATE CPU(sec) CPU(%) MEM(k) MEM(%) MAXMEM(k) MAXMEM(%) VCPUS NETS NETTX(k) NETRX(k) VBDS VBD_OO VBD_RD VBD_WR SSID
11572_test_0106_us_oracle_com –b— 8412 0.1 34603008 34.4 34603008 34.4 6 2 196796 1111780 2 90 37651 3174172 0
16026_test_0093_us_oracle_com –b— 4255 0.1 1048576 1.0 1048576 1.0 2 2 2092803 2914102 3 851 49446 1918015 0
16051_test_0094_us_oracle_com —–r 3636933 2396.8 56623104 56.3 56623104 56.3 24 2 1553895 970090 2 417 101921 10195220 0
Domain-0 —–r 36197 2.7 2621440 2.6 no limit n/a 24 0 0 0 0 0 0 0 0

So we can see that for vm ’16051_test_0094_us_oracle_com’, it has 24 vcpus, but the CPU(%) has reached 2396.8. We can calculate from 2396.8/24, that’s almost 100% usage of all the vcpus. So we can see that this vm is quite busy.

• PRECHECK

1. standalone OVS(source server) and OVS managed by Oracle VM Manager(destination server) must be the same type of machine, we can use command dmidecode |grep ‘Product Name’ to confirm.
2. make sure xend relocation server has been configured and is running, run the following commands to confirm:

grep xend-relocation /etc/xen/xend-config.sxp |grep -v ‘#’
(xend-relocation-server yes)
(xend-relocation-ssl-server yes)
(xend-relocation-port 8002)
(xend-relocation-server-ssl-key-file /etc/ovs-agent/cert/key.pem)
(xend-relocation-server-ssl-cert-file /etc/ovs-agent/cert/certificate.pem)
(xend-relocation-address ”)

lsof -i :8002
COMMAND PID USER FD TYPE DEVICE SIZE NODE NAME
xend 8372 root 5u IPv4 17979 TCP *:teradataordbms (LISTEN)

3.make sure ports are open between source and destination servers, run telnet <server_name> 8002 to confirm

4.Make sure source & destination servers are in the same subnet

• CREATE STORAGE REPO

To live migrate xen vm, the source & destination servers should have one NFS mounted. In Oracle VM, we can fulfill this by creating another storage repo for the current server pool.
The steps for creating storgage repo:
First, make sure the NFS share are writable to OVSes managed by Oracle VM Manager;
Second, run “/opt/ovs-agent-2.3/utils/repos.py -n <NFS share>” on master OVS;
Third, run “/opt/ovs-agent-2.3/utils/repos.py -i” on master OVS to make the storage repo seen by all OVSes managed by Oracle VM Manager;

• CREATE SYMBOLIC LINK

For live migration, the mount directories of the NFS share must be the same on source & destination OVS. But as the mount directory is automatically created by Oracle VM when creating the storage repo, so we must create symbolic link on destination OVS.
Assuming we have xen VM configuration on source OVS like the following:

disk = ['file:/repo_standalone/testvm/System.img,xvda,w']

Then we’ll link storage repo dir to /repo_standalone:

cd /
ln -s /var/ovs/mount/<uuid> /repo_standalone

• LIVE MIGRATE

Now on source OVS, let’s do the migration to destination OVS which has enough free memory

time xm migrate -l <vm> <destination OVS>

• IMPORT IMAGE

After the VM live migrated to destination OVS, we’ll need import the migrated VM to Oracle VM Manager. We’ll create another soft link under running_pool so that Oracle VM Manager can see the image:

cd /var/ovs/mount/<uuid>
ln -s /var/ovs/mount/<uuid>/<vm> .

After this, open GUI of Oracle VM Manager and then import & approve the system image.

PS:
You don’t need change VM configuration file(vm.cfg) manually, as after image imported to Oracle VM Manager the configuration file will be changed automatically by Oracle VM.

• To check current settings of o2cb, check files under /sys/kernel/config/cluster/ocfs2/
• To set new value for o2cb:

service o2cb unload
service o2cb configure

heartbeat dead threshold 151 #Iterations before a node is considered dead
network idle timeout 120000 #Time in ms before a network connection is considered dead
network keepalive delay 5000 #Max time in ms before a keepalive packet is sent
network reconnect delay 5000 #Min time in ms between connection attempts

service o2cb load

service o2cb status #will show new configuration if OVS in server pool; or it will show offline

PS: