Browse Prior Art Database

A method to mark cache contents invalid by tracking when data gets modified at source location

IP.com Disclosure Number: IPCOM000247122D
Publication Date: 2016-Aug-08
Document File: 4 page(s) / 67K

Publishing Venue

The IP.com Prior Art Database

Related People

Satya Ainaparthi: INVENTOR [+3]

Abstract

Due to hardware failure or for any other reason, the contents of the volume can be modified without involvement of cache. In this case cache contents will be stale. This invention describes a way to identify and invalidate the cache contents by using an identifier which will be associated with data volume and cache contents. The identifier is timestamp/chksum of the data volume. Here data volume means any file or any data object or portion of block device whose contents can be saved in a faster IO device.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 43% of the total text.

Page 01 of 4

A method to mark cache contents invalid by tracking when data gets modified at source location

Satya Ainaparthi

Rajesh Chepuri

Shailesh Marathe

Abstract

Due to hardware failure or for any other reason, the contents of the volume can be modified without involvement of cache. In this case cache contents will be stale. This invention describes a way to identify and invalidate the cache contents by using an identifier which will be associated with data volume and cache contents. The identifier is timestamp/chksum of the data volume.

Here data volume means any file or any data object or portion of block device whose contents can be saved in a faster IO device.

Problem Statement

When caching solution which provides warm cache support is used to store data volume contents for improving I/O performance, the data present on the data volume will be cached.

The contents of the volume can be modified in two ways without involving cache solution.

1) Data is restored using snapshot solution which does not involve caching layer (like replacing with cloned disks or restoring the contents from snapshot.

2) The data volume has migrated to different node due to HA failure/offline analysis requirement and application has written new content on it and the data volume migrated/on- lined back on its original node.

In either of the above two cases, when the data volume is available, it is necessary to find if we can use the cache contents or not. If no data is modified, we can use contents in cache for the data volume.

Publication Description

There are two ways of solving the issue: using time stamp as identifier for each data volume

1

© 2016 Veritas Technologies LLC. All rights reserved. Veritas and the Veritas Logo are trademarks or registered trademarks of Veritas Technologies LLC or its affiliates in the U.S. and other countries. Other names may be trademarks of their respective owners.


Page 02 of 4

object or using checksum of the data volume object as identifier.


a) Using time stamp:

We will write a time stamp whenever there is first access, at both the data volume and cache contents. We will update the time stamp when the application closes its access to data volume, both on data volume and on cache content. When the data volume object is getting associated.

This solves both the problems described in 5. Cloning/snapshot: As restore from snapshot/replacing with cloned disk requires application stop and as access close triggers a time stamp update on cache contents and data volume, we can easily identify if the data contents are modified as time stamp will be different on cloned data volume or snapshot.

HA failover/Offline analysis: For both HA failover/offline analysis, application will do last access close and the time stamp will be updated on both the data volume and cache contents. So if no data is modified in other node, and data volume is migrated back time stamp matches and will use warm cache.

If content gets modified, time stamp will be up...