Releases: openvstorage/framework
2.11.2
Information
This release contains fixes from 2.9.10
Bugfixes
- #1920: Scrubber will not deploy the same proxy for all scrubbing workers and re-use wherever possible
- #1916: Scrubber will only remove the proxy when no other worker is using it. Removal failure can now be auto-resolved and and underlying issue with the SSHClient has been tackled
Enhancements
- #1933: Scrubbing logic has been updated to work concurrently. This significant change required a code rewrite
Features
- #1939: Celery now uses Arakoon as a result backend. The Memcached one has been replaced with an in-house Arakoon backend because keys would not persist
2.9.10
Bugfixes
- #1920: Scrubber will not deploy the same proxy for all scrubbing workers and re-use wherever possible
- #1916: Scrubber will only remove the proxy when no other worker is using it. Removal failure can now be auto-resolved and and underlying issue with the SSHClient has been tackled
Enhancements
- #1933: Scrubbing logic has been updated to work concurrently. This significant change required a code rewrite
Features
- #1939: Celery now uses Arakoon as a result backend. The Memcached one has been replaced with an in-house Arakoon backend because keys would not persist
2.11.1
Bugfixes
- #1867: Removing every VPool now removes the volumedriver Arakoon
- #1803: ExecStopPost should no longer kill other volumedrivers
- #1888: Garbage collection will now be applied to already running alba services
- #1751: VDisk 'unknown' state is properly shown in the GUI
- #1915: VDisk name validation is now a thing
- #1669: Different way of detecting the feature
Enhancements
- #1850: A VDisk now has an 'unknown' state when the volumedriver does not respond to calls
- #1876: The update logic has been improved. All information is now stored within the model and services will be restarted less
- #1836: Previous scrub api now requests vdisk_guids to be supplied. This new API no longer requires it and scrubs all disks of a VPool
- #1783: Less packaging queries due to caching layer
- #1735: The support-agent is able to run in every condition. Configuration can be found here: https://github.com/openvstorage/framework/blob/master/docs/support_agent.md
- #1613: SSHClient's exit_code can now be returned
- #1930: RabbitMQ is set up with auto-healing
- #1871: Unittests can now be listed easier
Features
- #1860: The token lifetime can be configured
- #1700: Users are able to filter vdisks in the GUI. This filtering can be done on API level too (https://github.com/openvstorage/ovs-documentation/tree/master/Administration/usingtheapi)
- #1646: New API which allows a number of vdisks to be moved towards a targetted StorageDriver
- #1573: StatsMonkey has been integrated in the Framework
2.11.0
2.9.9
Bugfixes
#1530 - Better errorhandling during MDS checkup
#1738 - Catch MDS removal errors while shrinking a vPool
#1755 - MDS master not always moving to local node
Enhancements
#1778 - Create required amount of new MDS services when overload is detected
#1841 - Initiate backend sync after create_clone and create_clone_from_template
Features
#1397 - Configurable MDS safety per vPool
2.9.8
2.10.3
2.9.7
2.10.2
2.10.1
Bugfixes
#1711 - NSM checkup no longer claims random externally managed NSM clusters
#1796 - Ignore disks reported by lsblk when their corresponding /dev/<id> does not exist (issue when installing OVS in a container)
#1798 - Installed packages are now displayed properly on support page
#1802 - Backwards incompatibility issue when linking Backends between 2 different OVS versions