System and method providing object messages in a compute environment
View Patent ↗The present invention provides a system, method and computer-readable media for creating object messages within a compute environment such as a cluster or grid. The method comprises upon an event affecting the availability or performance of an object within a cluster environment, retrieving messages related to the event from messages logs in different layers of a cluster environment. Next, the method comprises aggregating the retrieved messages into a single location, associating the aggregated messages with the object and, if a user submits a job and the performance of the job within the cluster environment is affected by the event, notifying the user of the event using the associated messages. Administrators may also create messages and associate the messages with objects such as nodes or reservations.
1. A method comprising:
retrieving a first message from a first message log from a node layer of a compute environment;
retrieving a second message from a second message log of a resource manager layer of the compute environment, wherein the resource manager layer maps workload to the node layer as instructed by a workload manager;
retrieving a third message from a third manager log of a workload manager layer of the compute environment comprising the workload manager;
aggregating the first message, the second message and the third message into a single location to yield aggregated messages; and
associating the aggregated messages with one of a user and a group of users.
2. The method of claim 1 , further comprising:
upon receiving an inquiry from a user regarding the reservation, presenting the user with the messages associated with the reservation and the aggregated messages.
3. A system comprising:
means for retrieving a first message from a first message log from a node layer of a compute environment;
means for retrieving a second message from a second message log of a resource manager layer of the compute environment, wherein the resource manager layer maps workload to the node layer as instructed by a workload manager;
means for retrieving a third message from a third manager log of a workload manager layer of the compute environment comprising the workload manager;
means for aggregating the first message, the second message and the third message into a single location to yield aggregated messages; and
means for associating the aggregated messages with an object, wherein the object is one of a user and a group of users.
4. A non-transitory computer-readable medium comprising instructions for controlling a computing device to create an object messages in a compute environment, the instructions causing the computing device to perform operations comprising:
retrieving a first message from a first message log from a node layer of a compute environment;
retrieving a second message from a second message log of a resource manager layer of the compute environment, wherein the resource manager layer maps workload to the node layer as instructed by a workload manager;
retrieving a third message from a third manager log of a workload manager layer of the compute environment comprising the workload manager;
aggregating the first message, the second message and the third message into a single location to yield aggregated messages; and
associating the aggregated messages with one of a user and a group of users.
5. A system for managing messages in a compute environment, the system comprising:
a processor; and
a computer-readable storage medium storing instructions, which, when executed by the processor, cause the processor to perform operations comprising:
retrieving a first message from a first message log from a node layer of a compute environment;
retrieving a second message from a second message log of a resource manager layer of the compute environment, wherein the resource manager layer maps workload to the node layer as instructed by a workload manager;
retrieving a third message from a third manager log of a workload manager layer of the compute environment comprising the workload manager;
aggregating the first message, the second message and the third message into a single location to yield aggregated messages; and
associating the aggregated messages with one of a user and a group of users.