Adobe, Inc.

Dominique Pfister

Michael A Marth

Thomas Müller Graf

Thomas Müller

Marcel Reutegger

Systems and methods herein provide for a clustered content management comprising at least two computing nodes. A first node comprises an instance of the content repository. The first computing node may perform content management operations on its instance of the content repository. Changes to the instance of the content repository of the first computing node are synchronized with the content repository by way of a second computing node. The second computing node is communicatively coupled to the first computing node through a network and is operable to synchronize the change with the content repository. The second computing node also determines that synchronization of the change is blocked due to an error. The second computing node identifies the error, determines that the error is correctable, and corrects the error to synchronize the change with the content repository.

Self healing cluster of a content management system

One embodiment involves communicatively connecting a master node and a plurality of slave nodes in a cluster without shared storage where each node has a separate repository. Each node in the cluster without shared storage can be configured to exclusively update its repository. The method may include receiving a lock request from a first slave node in the cluster without shared storage. The lock request may be associated with an update request received by the first slave node. The lock request can indicate that the first slave node requires exclusive write access to a data item in the repository of the first slave node. The method includes sending a response to the first slave node granting the lock request based at least in part on data in the repository of the master node. The response may authorize the first slave node to append data to the data item in the repository of the first slave node.

Growing community of inventors

Basel, Switzerland

Dominique Pfister