Interactive access to headless cluster managers
A system and method for providing automatic interactive access to headless cluster managers by combining two different resource managers with different capabilities. Hybrid cluster deployments are created, and interactive access is automatically provided to headless cluster managers by using one framework, which supports interactive access, to access the other framework, which does not natively support interactive access. A client in a first framework may create a secure connection from the first framework to a node in a second framework and then pass interactive users requests such as requests for data from storage through the first framework and the secure connection to the node in the second framework, which can access storage.
1. A method for accessing resources comprising:
accessing a client on a first cluster with a first framework having software-defined networking;
creating a secure connection from the first cluster to a node on a second cluster with a second framework, wherein the second framework does not have software-defined networking; and
passing interactive users requests from a user through the first cluster using the client and the secure connection to the node on the second cluster.
2. The method of claim 1 , wherein the secure connection is ssh.
3. The method of claim 1 , further comprising accessing storage from the node, wherein data from the accessed storage is passed to the client and on to the user.
4. The method of claim 1 , wherein the first framework is Kubernetes.
5. The method of claim 1 , wherein the second framework is Slurm.
6. A non-transitory, computer-readable storage medium storing instructions executable by a processor of a computational device, which when executed cause the computational device to:
access a client on a first cluster with a first framework having software-defined networking;
create a secure connection from the first cluster to a node on a second cluster, with a second framework, wherein the second framework does not have software-defined networking; and
pass interactive users requests from a user through the first cluster using the client and the secure connection to the node on the second cluster.
7. The non-transitory, computer-readable storage medium of claim 6 , wherein the instructions, when executed, further cause the computational device to access storage from the node; wherein data from the storage access is passed to the client and on to the user.
8. The non-transitory, computer-readable storage medium of claim 6 , wherein the secure connection is ssh.
9. The non-transitory, computer-readable storage medium of claim 6 , wherein the first framework is Kubernetes.
10. The non-transitory, computer-readable storage medium of claim 6 , wherein the second framework is Slurm.
11. The non-transitory, computer-readable storage medium of claim 6 , wherein the instructions, when executed, further cause the computational device to configure the second cluster to include at least one headless cluster manager and at least one worker node.
12. The method of claim 1 further comprising configuring the second cluster to include at least one headless cluster manager and at least one worker node.
13. A system, comprising:
at least one processor configured to:
access a client on a first cluster with a first framework having software-defined networking;
create a secure connection from the first cluster to a node on a second cluster, with a second framework, wherein the second framework does not have software-defined networking; and
pass interactive users requests from a user through the first cluster using the client and the secure connection to the node on the second cluster.
14. The system of claim 13 , wherein the secure connection is ssh.
15. The system of claim 13 , wherein the first framework is Kubernetes.
16. The system of claim 13 , wherein the second framework is Slurm.
17. The system of claim 13 , wherein the at least one processor is further configured to configure the second cluster to include at least one headless cluster manager and at least one worker node.