Release 3.6.1

Release Date

v3.6.1 was released to the public on January 3, 2024.

New in This Release

This release introduces the following changes since version 3.5.x:

1. api-service: Description fields in create/update API calls are now limited to 256B at most.
2. api-service: You can now specify either a UUID or a server name when enabling or disabling a server. When both a name and a UUID are specified, the UUID is used and the name is ignored.
3. azure: Add a new option to deploy a support virtual machine connected to the Lightbits cluster. The support VM is used for managing the Lightbits managed application.
4. azure: Add support for deploying a single Lightbits cluster across multiple Azure Availability Zones.
5. azure: Add to the UI definition a checkbox explicitly alloliwing Lightbits Support to connect to the Lightbits cluster. Keep this unchecked to disable Lightbits from connecting to the Lightbits cluster.
6. azure: AVS (Azure VMware Solution) options are now hidden when deploying to an existing VNET.
7. azure: Added support for all regions that have the selected instance type available.
8. cluster-manager: Fixed a rare bug due to which volumes could have stopped responding to system updates, thus potentially leaving them in incorrect protection states or under extreme circumstances - even making them unavailable to the user.
9. cluster-manager: In some rare cases a "zombie" cluster key was left behind in etcd after node deletion. You should make sure to remove this.
10. k8s CSI plugin: The Lightbits Kubernetes CSI driver is now open source and available via https://github.com/LightBitsLabs/los-csi.
11. discovery-service: Fixed an issue that caused a failure to return a recovery log to the client and consequently failed nvme discover or nvme connect-all commands. The issue affected specifically nvme-cli 2.4, used in Rocky 9.3 and likely also RHEL 9.3 and Alma Linux 9.3.
12. docs: Clarified the requirements for the VPC endpoints in the AWS documentation.
13. installation: Ansible could sometimes try to reconnect to a server undergoing reboot before the server had a chance to reboot. This was fixed by setting a small delay after reboot and before starting to reconnect.
14. lbavs: Added an error message in case mount/unmount operations failed when invoked on non-existing datastores.
15. lbcli: Added a new command line option to evict data from servers when they are disabled, to prepare them for maintenance or replacement. This new command line option is currently disabled and will be enabled in future Lightbits releases.
16. lbcli: Added a new command line option --local-timezone. When passed, lbcli will print human readable timestamps for events and upgrades in the local timezone rather than UTC.
17. lbcli: Fixed issue where fetching logs deleted an existing older fetched log file when the command was reissued and failed on a file that already exists.
18. lbcli: The default timeout for the fetch log command was increased from 5 minutes to 20 minutes.
19. lb_top: Fixed display of nvme metrics while I/O is flowing through the system.
20. exporter: Updated the documentation with new metrics and clarified metric descriptions.
21. node-manager: In rare cases when trying to load a non-existing key from etcd, we could have been stuck in an endless loop. This was fixed by stopping on permanent (as opposed to transient) errors.
22. node-manager: In case of a software failure, the system will now attempt to recover by restarting the Lightbits services rather than performing a server reboot. This decreases the system downtime and recovery time.
23. node-manager: In rare cases the node-manager could panic due to encountering a node without an attached server, due to a temporary disconnection from etcd. This is handled gracefully.
24. openstack: Added a new option to the Cinder driver to use Lightbits IPACLs. When configured to use IPACLs, Cinder will create IPACL entries for each volume, limiting which hosts can access it to the specific compute hosts that the volume is attached to.
25. lb_irq_balance: improved the way lb_irq_balance handles scenarios where there are not enough idle (or less busy) CPU cores to handle network interrupts.
26. upgrade-manager: From this version forward, upgrades will no longer require a reboot to complete.
27. upgrade-manager: In rare cases, the upgrade-manager could fail to set the current upgrade task as complete and give up, leading to an inability to perform further upgrades. This is fixed by retrying until the task state is updated successfully.
28. userlbe: Made reading stats atomic so that the stats output never changes between opening a stats file and reading its contents.
29. userlbe: Time-based statistics were broken and are now fixed and based on the TSC register.

Installation and Upgradeability

This release is upgradeable from prior releases listed below. An (x) indicates upgradeability from the listed release. An empty bracket "( )" indicates that this release is not directly upgradeable from that particular release and upgrading requires an upgrade to an intermediary release (e.g., 3.3.2 to 3.4.2 to 3.5.1 to 3.6.1) or a new installation.

( ) Lightbits 2.2.x

( ) Lightbits 2.3.x

( ) Lightbits 3.0.x

( ) Lightbits 3.1.1

( ) Lightbits 3.1.2

( ) Lightbits 3.2.1

( ) Lightbits 3.3.1

( ) Lightbits 3.3.2

( ) Lightbits 3.3.3

( ) Lightbits 3.4.1

( ) Lightbits 3.4.2

( ) Lightbits 3.4.3

(x) Lightbits 3.5.1