Browse Prior Art Database

Method or Apparatus for having a Dynamic panel manager algorithm for management of Operator Panel(s) on next generation Multi node Power hybrid cloud enterprise server and its like

IP.com Disclosure Number: IPCOM000253190D
Publication Date: 2018-Mar-13
Document File: 9 page(s) / 389K

Publishing Venue

The IP.com Prior Art Database

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 33% of the total text.

Method or Apparatus for having a Dynamic panel manager algorithm for management of Operator Panel(s) on next generation Multi node Power hybrid cloud enterprise server and it’s likeThis is in context of multi node servers, wherein each server node will have its own service processor(s). They can also use different stacks for the service processor and example would be the Next Generation High End Multi node Power hybrid cloud enterprise servers, using an Open service processor stack, and in a high-end enterprise class box there would be four worker nodes. One possible configuration of high end enterprise class Power servers is multiple worker nodes each managed locally by their own Service Processor(s). earlier systems also had a control node which used to house the service processor(s). In earlier system we would have the operator panel being housed in the control node, and so we had only one instance of it on the system and this was managed by the service processor on the control node. In the new server design, we could have housing slots for the Operator panel in each of the worker nodes. we would need a mechanism to use them wisely and that is what is given as a solution in this disclosure. This could further be elaborated as below:

Since current systems employ only one panel and not multiple ones the below problems will need to be take care of 1) Panels in each node need to discover and talk to all the other panels. 2) Role arbitration is needed to determine a master. 3) Failure handling and role switching.

we would have an enterprise redundant Panel manager running on all the service processors , each instance of the manager running on the different service processors would talk to each other and reach a consensus on who would be the manager and would manage the external Operator panel, the cases could be that there is a single operator panel present , there could be multiple panels present , user may use any panel , or the operator panel has failed due a hardware issue . It would find and assign one of the panel manager as the master and all others as slave if present and the slave would be in an ordering of taking over the master-ship in case of a need. One way is that they would then enable the panel connected on the CEC on which the panel manager is master to be working and all other slave ones would be in just display mode only. Another way It could be done is in which all of the panels would work but the master panel manager is the one who would handle all the incoming and outgoing requests routed via the slaves for the slave managers. The setup of the method of usage can be done via configurations used for the system. It could also be changed dynamically.

Implementation:

Assuming we have WORKER NODE-B, WORKER NODE-C, WORKER NODE-A and WORKER NODE-D as the worker nodes and they are assigned the node position by the service processor as WORKER NODE-B => node 1, WORKER NODE-C => node 2, WORKER NODE-A => node 3 and WORKER NODE-D => n...