Browse Prior Art Database

Method And Apparatus To Move Drives In A Raid Array Concurrent With Full Array Access

IP.com Disclosure Number: IPCOM000201006D
Publication Date: 2010-Nov-04
Document File: 6 page(s) / 64K

Publishing Venue

The IP.com Prior Art Database

Abstract

For a enterprise storage system it is important to provide uninterrupted availability to the data in a RAID array. Redundancy and concurrent repairability of components is critical to achieve this. However even in a fully redundant system there are elements such as the disk array backplane which are not replaceable without removing all the drives installed in the enclosure from operation. Eliminating single points of repair force certain design constraints. In the case of the drive backplane we require it to be completely passive, (no active components) and call the failure rate the engineering equivalent of zero. Higher speed device interfaces and packaging constraints are making it difficult to achieve a completely passive backplane. A mechanism to replace a defective backplane is needed. Other mechanisms are to copy the data from one set of drives to another. This is a very long process especially for large drives.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 85% of the total text.

Page 01 of 6

Method And Apparatus To Move Drives In A Raid Array Concurrent With Full Array Access

The diagrams below show the sequence and cabling required to move drives within the array concurrent with operation.

Read operations to drive being moved when it is unavailable are parity reconstructs from the other drives in the RAID array.

Writes to the drive being moved are saved in a nonvolatile write caching memory.

(This page contains 00 pictures or other non-text object)

1


Page 02 of 6

(This page contains 00 pictures or other non-text object)

2


Page 03 of 6

(This page contains 00 pictures or other non-text object)

3


Page 04 of 6

(This page contains 00 pictures or other non-text object)

4


Page 05 of 6

(This page contains 00 pictures or other non-text object)

5


Page 06 of 6

(This page contains 00 pictures or other non-text object)

If a drive fails during the process:
1. Drive being moved fails - Stop process and rebuild drive. Remove defective drive. Resume drive move process on next drive.
2. Some other drive in move group fails. If drive to be moved has not been pulled, De-stage data to the drive. Pause move process to rebuild failed drive. Resume drive movement process.

Potentially separate disclosure content.
1. If write volume exceeds or is expected to exceed write cache capacity for the drive move duration nonvolatile storing of sector identifiers for data not yet updated on drive can be performed. When drive is made available after physical move, the modified sector...