Method and system for managing switch workloads in a cluster
AssigneeArista Networks, Inc.
InventorBenoit Sigoure
Life of this patent
Estimating the term…
Abstract
In general, embodiments of the invention relate to a switch that includes a processor, ports, and memory that includes instructions, which when executed by the processor perform a method. The method includes obtaining, via a port, a global task-server mapping from a first server, detecting that one of the ports is congested. The method further includes, based on the detecting, identifying a task associated with the one of the ports using the global-task server mapping, generating a task listing based on the identifying, generating an alert including the task listing, and transmitting the alert to an administrator.
Claims20
A switch, comprising:
- a processor;
- a plurality of ports;
- memory comprising instructions, which when executed by the processor perform a method, the method comprising:
- obtaining, via a first port of the plurality of ports, a global task-server mapping from a first server;
- detecting that one of the plurality of ports is congested;
- based on the detecting:
- identifying a task associated with the one of the plurality of ports using the global-task server mapping;
- generating a task listing based on the identifying;
- generating an alert comprising the task listing;
- transmitting the alert to an administrator.
A switch, comprising:
- a processor;
- a plurality of ports;
- memory comprising instructions, which when executed by the processor perform a method, the method comprising:
- obtaining local switch load information, wherein the local switch load information specifies bandwidth utilization of the switch;
- and
- transmitting the local switch load information to a server.
A switch, comprising:
- a processor;
- a plurality of ports;
- memory comprising instructions, which when executed by the processor perform a method, the method comprising:
- detecting that a link to a first server is down, wherein the first server is directly connected to a port of the plurality of ports on the switch;
- generating, based on the detecting, a link down message specifying that the link to the first server is down;
- and
- transmitting, via one of the plurality of ports, the link down message to the second server, wherein the link down message is sent to the second server prior to a time that the second server expected to receive a heartbeat message from the first server.
A switch, comprising:
- a processor;
- a plurality of ports;
- memory comprising instructions, which when executed by the processor perform a method, the method comprising:
- obtaining a first set of task executing in a cluster;
- obtaining a first list of servers in the cluster;
- determining which servers on the first list of servers is the cluster are directly connected to the switch to obtain a second list of servers;
- querying each server specified in the second list of servers to which of tasks are executing on the second list of servers to obtain a second set of tasks;
- generating a local task-server mapping using the second list of servers, the first set of tasks and the second set of tasks, wherein only tasks that are present in both the first set of tasks and the second set of tasks are used to generate the local task-server mappings;
- and
- annotating an alert message using the local task-server mappings.
Dependencies are inferred from the claim text.