RONIN Release: Tuesday 20th December 2022
Additional Object Storage Archiving Options! Improved filter when selecting machine types! Lots of helpful minor features, and some bug fixes to boot!
Major Features
Additional Object Storage Archiving Options
- Previously, when archiving was enabled on an object storage bucket in RONIN, the standard Glacier (Flexible Retrieval) object storage archive class was used by default. Now you can select from all the three available Glacier object storage archive classes available in AWS including Glacier Instant Retrieval, Glacier Flexible Retrieval or Glacier Deep Archive to better manage cost vs access.
Filter on CPU, Memory and Storage when selecting instance types
- When searching for an instance type for a machine or cluster, you can now filter on the number of CPUs, amount of RAM, or type of storage you desire, in addition to the existing general search option.
Minor Features
- Instance types will now be automatically filtered based on the selected OS architecture. For example, if you select an arm-based OS or package, only arm instance types will be available for select. Similarly for X86 architectures.
- Added permissions for users to list, access and delete object versions in their object storage buckets.
- Added instance and system check information for running single machines and clusters.
- Added an alert when project spend is greater than the project limit in the Admin budget management screen.
- Improved reporting of machine errors when restarting a machine e.g. catch capacity limitations.
- Added "created by" column to the machine table on the project dashboard and Admin machine table screen to easily identify which users created which machines.
- Added timestamps to machine creation and modification (start/stop).
- More accurate Windows Running status.
Bug Fixes
- Fixed bug where stop only smart schedules were not displaying correctly in the UI upon edit.
- Prevent object storage names exceeding the maximum number of characters.
- Fixed bug where users were unable to delete failed machine/cluster packages.
- Fixed bug where browser autocomplete was pre-filling search fields.
- Fixed bug where GPU specs for clusters were not being reported correctly for compute nodes.
- Fixed bug where RPIDs containing period characters would prevent successful cluster launch.
- Added catch where packages that do not support ENA cannot be added to the package catalogue.