Upgrading Federated Instances
This version of OpenNebula introduces some changes in the federation data model. You need to coordinate the upgrade across zones and upgrade them at the same time.
Step 1. Check Federation Status
Check that federation is in sync and all zones are at the same index (FED_INDEX):
$ onezone list
C ID NAME ENDPOINT FED_INDEX
101 S-US-CA http://192.168.150.3:2633/RPC2 715438
100 S-EU-GE http://192.168.150.2:2633/RPC2 715438
* 0 M-EU-FR http://192.168.150.1:2633/RPC2 715438
It is a good idea to prevent any API access to the master zone during this step (e.g., by filtering out access to API).
onezone disable <zone_id>.Step 2. Stop All Zones
Stop OpenNebula and any other related services you may have running: OneFlow, OneGate & FireEdge. It’s preferable to use the system tools, like systemctl or service as root in order to stop the services.
systemctl list-units | grep opennebula should be empty.Step 3. Upgrade Master Zone
You can now upgrade the master zone:
Step 4. Back up Federated Tables
Once the master zone has been updated, you need to export federated tables:
$ onedb backup -v --federated
--user and --password, and --host if the database is not on localhost. Please refer to the CLI Reference for further information.Step 5. Restore Federated Backup in Slave Zones
The backup that has been generated needs to be restored in all slave zones:
$ scp <backup_file> <slave_ip>:/tmp
$ onedb restore <backup_file> -v --federated
Step 6. Upgrade Slave Zones
You can now upgrade the slave zones:
You will restart OpenNebula in each zone as part of the upgrade. Once you finish upgrading your master, remove any access restriction to the API imposed in Step 1.
onezone enable <zone_id>.GIVE FEEDBACK
Was this resource helpful?
Glad to hear it
Sorry to hear that