DevOps (Artifactory, Distribution)
Update | Details |
|---|---|
New REST APIs for Pruning Unreferenced Data (PUD) | Three new REST APIs have been added for pruning unreferenced data (PUD): Start PUD Process, Get Status of the PUD Process, and Stop PUD Process. These APIs are used to delete randomly existing binaries in the filestore that are not referenced in Artifactory and cannot be deleted by Artifactory Garbage Collection. These APIs offer greater control over the PUD process from what was available prior to this release through the PUD user interface. The Start API can be used to start the PUD process from a specific directory or even from a specific binary in a directory, and there is an option to refrain from deleting unreferenced binaries that were created after a certain date. The Get Status API returns the current status of the PUD process (i.e., running, stopped, finished, or error), along with relevant data on the process such as total binaries deleted, total size of disk space cleaned, information on the last handled directory, and other information. And the Stop API allows you to stop the PUD process before it reaches the end of the directories list. For more information, see Pruning Unreferenced Data (PUD) APIs |
Hugging Face Naming Limitations and Model Names Enhancement | Artifactory now supports the naming limitations enforced by Hugging Face when deploying models to local repositories. For more information, see Hugging Face Naming limitations. Artifactory now supports hugging face models with names that do not follow the "Organization/Name" structure. |
Get Release Bundle Versions API Enhancement | A new query parameter option in the Get Release Bundle Versions API makes it possible to return the permission settings for promoting, distributing, and deleting the specified Release Bundle v2 versions. |
Federation Recovery and Auto-Healing | Artifactory now features an auto-healing mechanism that can recover automatically from error exhaustion and reduce the need to perform full sync operations on Federated repositories. Auto-healing checks Federated repositories at regular intervals for exhausted queues (queues that have exceeded the maximum number of attempts to send events to other Federation members). This mechanism resets the failed events automatically and tries again to sync with the target mirror. For more information, see Federation Recovery and Auto-Healing. There is also a new Federation Recovery API that enables administrators to perform recovery manually as an alternative to performing a full sync. |
Hide SSO Providers | You can now hide the SSO providers that you configured in the JPD from the login screen. If you hide the providers, the users need to type in the SSO providers to log in to the JFrog Platform UI. For more information, see Hide SSO Providers. |