Skip to content

Latest commit

 

History

History
84 lines (73 loc) · 3.44 KB

upgrade.md

File metadata and controls

84 lines (73 loc) · 3.44 KB

Upgrades

Automated upgrades via the GUI

The easiest and most reliable way to upgrade Open vStorage is through the GUI. Go to Administration > Updates to verify if updates are pending.

  • Click the Update Framework button to update the Open vStorage Framework.
  • Click the Update Storage Driver button to update the Volume Driver and Open vStorage Backend.

Manually upgrade an environment

The below procedure describes how to manually update all components of an Open vStorage environment. Do follow these steps closely as an incorrect update might result in data loss or an unrecoverable environment.

The packages of the different versions can be found on the repo's: http://apt.openvstorage.org/ and http://yum.openvstorage.org/.

The manual procedure consists out of 5 steps:

Stop the services

To stop all services execute

  • For the Framework
service ovs-watcher-framework stop
service memcached stop
service arakoon-ovsdb stop
  • For the Volume Driver
service ovs-volumedriver_vpoolname stop
service ovs-dtl_vpoolname stop
  • For ALBA
service ovs-albaproxy_vpoolname stop
service alba-asd-* stop
service ovs-alba-maintenance* stop
service ovs-alba-rebalancer* stop
Upgrade ALBA
  • Make sure you are also about to update Arakoon if required
  • If you have less than 3 master nodes, make sure all volumes are stopped
  • On all nodes, install the new alba package through apt or yum.
  • On all nodes, restart each ASD
    • Find them with ls -al /etc/init/alba-asd-*
  • On all nodes, restart all proxies by sending a kill signal kill
    • Find them with ls -al /etc/init/ovs-albaproxy* and for each entry service ovs-albaproxy_<vpool name> status
  • On all nodes where applicable, restart the alba arakoon clusters, foreach backend
    • Find them with ls -al /etc/init/ovs-arakoon-<backend name>*
      • If < 3: Stop them all, and start them all up again
      • If >= 3: Restart them one by one, waiting a minute in between
  • On all nodes, restart the maintenance and rebalancer processes
    • Find them with ls -al /etc/init/ovs-alba-maintenance* and ls -al /etc/init/ovs-alba-rebalancer*
Upgrade the Volume Driver
  • On all nodes, install new Volume Driver packages through apt or yum.
  • On all nodes, for each running vPool execute restart ovs-volumedriver_vpoolname
    • Find them with ls -al /etc/ovs-volumedriver*
Upgrade the Framework

You simply can't upgrade this one without the GUI unless you manually touch /etc/ready_for_upgrade it will refuse to install.

  • Stop all relevant services if not done before (watcher-framework, memcache, arakoon-ovsdb)
  • Upgrade the packages through apt or yum.
  • Start arakoon-ovsdb, memcached on the master nodes
    • On 1 node:
from ovs.dal.helpers import Migration}}
Migration.migrate()

* On all nodes:

from ovs.extensions.migration.migrator import Migrator
Migrator.migrate()
  • Restart memcache on all master nodes
  • Start the watcher-framework again
service ovs-watcher-framework start

Do not forget to patch files you manually patched after the installation as they might have been overwritten by the package update.