Browse Prior Art Database

Autonomic Expanding/Increasing number of Virtual Tape devices during On-line State

IP.com Disclosure Number: IPCOM000238322D
Publication Date: 2014-Aug-18
Document File: 6 page(s) / 354K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method for the automatic/autonomic enabling/increasing of the number of devices during an on-line state within a virtualization engine for storage.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 49% of the total text.

Page 01 of 6

Autonomic Expanding/Increasing number of Virtual Tape devices during On-line State

This article addresses the updating of devices managed by a virtualization engine, which is a mainframe virtual tape solution that

implements a fully integrated, tiered storage of disk and tape. This virtualization engine can dynamically increase the number of virtual devices while in an on-line state. In a grid configuration, up to six engine clusters are interconnected, and one of clusters

works with the other clusters to replicate data. However, the updated code releases (i.e. from an older version to a new one) cannot be simultaneously applied to all clusters, as it reduces the availability of production services for customers. Therefore, after updating the new release, the new feature must be disabled until the other clusters' code releases are updated. The process is further complicated because operator intervention is required when new code is applied to all clusters.

A method is needed to more efficiently update program releases during incremental updates across multiple (i.e. an increased number of) supported devices in a virtualization engine.

The novel contribution is a method for the automatic/autonomic enabling/increasing of the number of devices during an on-line state within a virtualization engine.

The method comprises the following actions:

1. Suppress virtual devices at a higher address, even on the new code level, unless all servers are updated. When the host makes a device on-line, which should not be on-line because the associated device address is out of range, the engine rejects the initial command from the host in the device driver. The device driver determines whether the device address is out of range by referring to a domain code level.

2. Notify the clusters that have already been dynamically updated of the event release update of the new cluster. When the remote cluster is updated from the old release level to the new, the supported device address range is expanded. Notify the clusters of the updating event, by utilizing the Library Capabilities Updated events ("2F"x attention) instead of polling the other clusters' code levels.

3. Activate the virtual devices at a higher address with a notification that all other clusters have been upgraded. Start additional virtual device processes on demand. This can minimize usage of system resources.

A virtualization engine is built with two components: Vnode and Hnode. Each component includes some sub-components. VNode

virtualizes the tape device interface. It receives the tape commands from the host and then parses and processes the commands. Hnode accumulates the received data or restores the data from/to hard disks or physical tapes. The grid manages the physical storage resources by selecting appropriate clusters to store the data from the host.

1


Page 02 of 6

Figure 1 illustrates the control flow and data flow between Vnode and Hnode, interconnecting the other cluster. The Vi...