Browse Prior Art Database

Method to speed up failover of applications in cluster

IP.com Disclosure Number: IPCOM000248150D
Publication Date: 2016-Nov-02

Publishing Venue

The IP.com Prior Art Database

Abstract

Proposed idea here will have faster fail over of applications and critical work load running on cluster environment. Present Idea will deal with file system locking and unlocking way using cluster wide lock to release application from one cluster node to next fail over node, Operation for moving application from node failure to next priority node will be at file system level, where as many method currently use operations till volume group level for failing nonconcurrent resource in cluster. This idea will help to achieve fail over with faster rate and high availability of application can be more speedy, with less logical volume manager operations..

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 30% of the total text.

Page 01 of 10

Method to speed up failover of applications in cluster

In present cluster fail over techniques, during failure of node, action will be performed till volume groups level in lvm, so that tends to take more time to fail over application to other node in cluster,
Proposed idea here will have faster fail over of applications and critical work load running on cluster environment. Present Idea will deal with file system locking and unlocking way to release application from one cluster node to next fail over node, Operation for moving application from node failure to next priority node will be at file system level, where as many method currently use operations till volume group level for failing nonconcurrent resource in cluster. This idea will help to achieve fail over with faster rate and high availability of application can be more speedy, with less logical volume manager operations.

Major advantage from this idea is, Fail over time will be reduced as we are not performing varyoff and varyon volume group operations. If cluster consists of more number of resources, it will take less time for failure in compare to present technique available. This will result into faster fail over and reduction in less lvm operations, Also the security level can be increased as many proposed idea have scsipr capability, so here apart from file system locking and unlocking scsi pr capability can also be added.

Claims :


1) This will speed up failover process for applications in cluster environment, failover rate will be faster compare to other clustering technologies


2) Most of cluster based product for high availability have method of taking action on volume group and file system both , but this idea will take action only on file system level, which decrease the operations to get application fail over faster


3) security can be increased using this idea for application if volume group are created on scsi pr enabled disk. Disk itself will have PR shared keys and second layer will be file system lock, if application is access on node file system lock will be release where as disk will be locked with PR shared key.

4 ) Idea can also be applied for concurrent resource group to get work load balancing on a sit.

5 ) Work load for concurrent resource group for cluster with more that one site can also be failed over with faster speed to other site.

In proposed idea cluster is created with resource group( serially action) as per the node priority and policies ( startup policies, fall over policies and fallback policies), under this resource group service ip will be added. Concurrent volume group will be created. File system will be created on top of volume group and it will have normal logical volume manager operations. Logical volume will be created itself by creating file system. File system will contain applications to be used for high availability. Once cluster configured, synchronization will be done to propagate cluster wide same configuration on all the no...