Surety is performing system maintenance this weekend. Electronic date stamps on new Prior Art Database disclosures may be delayed.
Browse Prior Art Database

A method to rescue data from space starvation in rebuilding process

IP.com Disclosure Number: IPCOM000242980D
Publication Date: 2015-Sep-04
Document File: 3 page(s) / 48K

Publishing Venue

The IP.com Prior Art Database


This disclosure proposes a method which can conditionally utilize the unallocated space of the failed module(s) in rebuilding process, and therefore can rescue data from space starvation.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 3

A method to rescue data from space starvation in rebuilding process

Scale out storage system uses management module to decide whether other modules are functioning and alive. If a module is decided as "not functioning" or "not alive", it will be excluded by the management node from the cluster and labeled as "failed". If a module gets to "Failed" unexpectedly, i.e., not by "Phasing out", a data rebuild is required for the cluster because some data has onlyone copy after the module failures. The data rebuild replicates the single data and makes the clusterto be full redundancy again. Meanwhile, the same amount of free capacity is required in the cluster in rebuild. The failed module can't be utilized by the cluster at all in rebuild process. And it can join in the cluster again only when it passes the component test process, which tests all the hardware of the module and downloads the microcode to the module. Then the module can be added into the cluster as a fresh new device. As a result, all the legacy data and metadata are cleaned during the component test process.

The above procedure is the strictest one. It confines the system's available capability in a large content because the "failed" modules are excluded fromthe cluster and the capacity of these failed modules can't be utilized at all. This may impede the rebuild process becauseof insufficient free space and increase the risk of data loss if there is another module failure in the rebuild process.

However, in many filed cases, a module (or modules) is failed only for software reasons which can be r...