Browse Prior Art Database

A system service to determine and remove zombie file storages of Network File System in Cloud environment

IP.com Disclosure Number: IPCOM000248957D
Publication Date: 2017-Jan-24
Document File: 8 page(s) / 214K

Publishing Venue

The IP.com Prior Art Database

Abstract

This article describe an automatic method to determine and remove zombie file storages of Network File System in Cloud environment. The method is running on a virtual server within the same Local Area Network of NFS storage, monitor all the NFS request over TCP/IP network traffic, parse source and target information, set virtual servers and file storages with specified state and save to a table according to monitor rule, and remove all zombie file storages if the virtual server was removed, no matter the virtual server was removed intentionally or destroyed accidentally.

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

1

A system service to determine and remove zombie file storages of Network File System in Cloud environment

Nowadays Cloud Computing technology is widely used in many domains, the storage requirements increased along with applications in Cloud environment.

User connect bare metal and virtual servers to durable, fast, and flexible Network File System (NFS) based file storage. In this Network Attached Storage

(NAS) environment, user has total control over file shares function and perform.

In the relationship between the file storage and the virtual server, the file storage is in a passive position, this means even a file storage is not used by any

virtual server, user will never know the broken relationship unless manual investigation.

In most of cases, these file storages usually mounted by specified bare metal and virtual servers, if the virtual server was removed or destroyed for some

reason, whether intentional or accidental, the relationship between a host inventory object and associated objects living on data stores has become

inconsistent, the file storages will leftover there as orphan or zombie images. From storage perspective, all orphan or zombie images are actually 'active', it

is hard for user to determine which one is zombie image and can be deleted permanently. This can cause significant storage problems, including wasted

resources, increased costs, and reduced performance.

To resolve this problem, there are two methods currently used in popular Cloud environment:

• "Label" file storage: each file storage can only be accessed over TCP/IP (Transmission Control Protocol/Internet Protocol) by a host in a limited list, related action named 'Authorize', the authorized list stored along with the storage itself. User only can check if the associate host is existed or not

through revert path from the authorized list. If a virtual server with mount point was removed, user can remove the file storage manually.

• Virtual server configuration: maintain a specified configuration file for virtual server with associated information between virtual server and file storage. User cannot query if the file storage is associate directly, compare configuration file of all the hosts with the storage lists is the only way to

confirm if a storage is zombie or not.

In either cases, user have to verify if a file storage is zombie or not manually, then select keep or remove it, there is no automatically way to do it by a

system method or service.

This invention introduce a system service running on a virtual server within the same Local Area Network (LAN) of NFS storage, it will maintain a table,

monitor all the NFS request over TCP/IP network traffic, record source and target information, set virtual servers and file storages with specified state and

save to table according to monitor rule, and remove all zombie file storages if the virtual server was removed, no matter the virtual server was removed

intentionally or destroyed accidentally. Or list all pot...