Browse Prior Art Database

Redundant Console Attachment with High Availability

IP.com Disclosure Number: IPCOM000013626D
Original Publication Date: 2001-May-16
Included in the Prior Art Database: 2003-Jun-18
Document File: 4 page(s) / 72K

Publishing Venue

IBM

Abstract

Background

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 29% of the total text.

Page 1 of 4

Redundant Console Attachment with High Availability

Background

The Alternate SE (Support Element) feature was initially released on the IBM 9672 G6 series of processors (S/390 Parallel Enterprise Server - Generation 6 processors). The 9672 previously had only one SE per system to provide such functions as System Activation and Loading of an operating system (ie., OS/390, VM, etc.). If this SE would break, the operating system(s) running on the 9672 would continue to run, but if the user wanted to Activate a new partition or reload an operating system, he/she would be unable to do this until the SE was repaired which could take several hours to have a service representative to come onsite with the repair parts.

The Alternate SE feature provides a second SE (called the Alternate SE) in addition to the Primary SE (the active SE). Periodically, the Primary SE mirrors its customized data to the Alternate SE to keep the Alternate SE harddisk at the same running state as the Primary SE. The data is mirrored across the customer network (the LAN (Local Area Network) which the HMC(s) (Hardware Management Console(s)) and SE use for communication). However, only the Primary SE is the only SE used for communication to the 9672 processor.

9672 G6 Processor

For the G6 Processor, both the Primary and Alternate SEs have cables attached to their parallel port adapter on the SE Thinkpad hardware, and these cables go to a physical switch which then have a single cable output which is then connected to the 9672 processor. If the Primary SE breaks, it should trigger an error being reported to the service provider which in turn signal the service provider to send someone onsite to throw the Alternate SE switch to cause the Alternate SE to now become the Primary SE and the former Primary SE to be the Alternate SE. The service provider could also ask the customer to go to the 9672 to throw the switch as well, but the 9672 may or may not be at the same location as the customer since sometimes data centers are isolated in their locations.

Once the Alternate SE switch is thrown, the SEs will switch no matter what state each of the SEs are in. If the Primary SE is busy doing something including service actions, the switch will occur. If the Alternate SE is at a code level that may not be compatible with 9672 processor microcode, the switch will execute. There is no way to stop the switch since it is a physical switch.

Someone has to physically go onsite to the throw the switch which is part of the 9672 frame. In many cases the 9672 systems are locked unattended in secure rooms, and it can take time to get someone to the location of that system and obtain access to them. This is unfortunate since the HMC (Hardware Management Console) allows the customer to operate the 9672 remotely from not only another location in a building but also from almost any location around the world.

z 900 Processor

For the z900 processors (zSeries 900), we invented solutions to each...