Host-level persistent volume management for container clouds
Methods and systems for managing persistent volumes include receiving a request from a container on a processing node to access a local mount point. A distributed filesystem located outside the processing node is mounted to a local mount point. Access to the local mount point is provided to the container.
1. A computer-implemented method for managing persistent volumes, comprising:
receiving a request from a container on a processing node to access a distributed filesystem that is located outside the processing node;
mounting the distributed filesystem to a local mount point; and
giving access to the local mount point to the container.
2. The computer-implemented method of claim 1 , further comprising communicating synchronization information regarding the distributed filesystem to other processing nodes.
3. The computer-implemented method of claim 2 , wherein the distributed filesystem is non-exclusively mounted, such that other containers can also access the distributed filesystem.
4. The computer-implemented method of claim 1 , further comprising enforcing quota limits on the local mount point.
5. The computer-implemented method of claim 1 , wherein the local mount point is a directory in a local filesystem of the processing node.
6. The computer-implemented method of claim 1 , wherein receiving the request from the container invokes a hostpath driver in a container orchestration engine.
7. The computer-implemented method of claim 6 , wherein the hostpath driver communicates with a storage provision engine that performs the mounting and that communicates directory information back to the hostpath driver to give access to the local mount point.
8. The computer-implemented method of claim 6 , wherein the distributed filesystem is mounted as a persistent volume without invoking any native persistent volume driver within the container orchestration engine.
9. The computer-implemented method of claim 1 , wherein the distributed filesystem is also mounted to a container on at least one other processing node.
10. A non-transitory computer readable storage medium comprising a computer readable program for managing persistent volumes, wherein the computer readable program when executed on a computer causes the computer to perform the steps of:
receiving a request from a container on a processing node to access a distributed filesystem that is located outside the processing node;
mounting the distributed filesystem to a local mount point; and
giving access to the local mount point to the container.
11. A system for managing persistent volumes, comprising:
a container orchestration engine, executed in a hardware processing node, configured to receive a request from a container on a processing node to access a distributed filesystem that is located outside the processing node; and
a storage provision engine configured to mount the distributed filesystem to a local mount point and to give access to the local mount point to the container.
12. The system of claim 11 , further comprising a sync engine configured to communicate synchronization information regarding the distributed filesystem to other processing nodes.
13. The system of claim 12 , wherein the distributed filesystem is non-exclusively mounted, such that other containers can also access the distributed filesystem.
14. The system of claim 11 , further comprising a quota manager configured to enforce quota limits on the local mount point.
15. The system of claim 11 , wherein the local mount point is a directory in a local filesystem of the processing node.
16. The system of claim 11 , wherein the request from the container invokes a hostpath driver in the container orchestration engine.
17. The system of claim 16 , wherein the hostpath driver communicates with the storage provision engine and wherein the storage provision engine is further configured to communicate directory information back to the hostpath driver to give access to the local mount point.
18. The system of claim 16 , wherein the distributed filesystem is mounted as a persistent volume without invoking any native persistent volume driver within the container orchestration engine.
19. The system of claim 11 , wherein the distributed filesystem is also mounted to a container on at least one other processing node.
20. The method of claim 1 , wherein the container includes a set of software applications, configuration files, and workload datasets and is instantiated by a container orchestration engine.