Browse Prior Art Database

Method and apparatus to implement high availability for a hybrid application server in a clustered environment

IP.com Disclosure Number: IPCOM000240547D
Publication Date: 2015-Feb-09
Document File: 6 page(s) / 87K

Publishing Venue

The IP.com Prior Art Database

Abstract

A method for initiating transaction recovery for every transaction co-ordinator located in a hybrid application server in a clustered environment.

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

Page 01 of 6

Method and apparatus to implement high availability for a hybrid application server in a clustered environment

Background

A hybrid application server consists of multiple execution environments (such as cloud), each capable of hosting application modules written in specific technologies, for eg: CICS execution environment, Java EE execution environment etc...

In such a hybrid application server, many of the execution environments would have the capability to act as transaction coordinators. So whenever a hybrid application server crashes due to hardware or software failures, transaction recovery needs to happen in each of the individual execution environments. Some execution environments may not support active recovery, in which case, a separate instance of the execution environment may have to be brought up carry out the recovery for the instance which went down. These two requirements bring up the following challenges:

A mechanism is needed to record and convey the location of transaction logs of each of the execution environment so that another hybrid application server may carry out the recovery for a failed instance
A mechanism which could efficiently carry out the recovery, even for those execution environments which don't support active-active recovery through an already running instance

In this particular case, recovery may fail if there are not enough resources. If the deployment is on a cloud, recovery candidate needs to be chosen intelligently such that it always succeeds and does not take much time.

So a method is required which could address all of the above requirements and provide a way to carry out the recovery for hybrid application servers in an efficient manner in a cloud environment. Though we have stated this problem for cloud, it is widely applicable to normal clustered environments as well.

No known mechanisms available today through which a hybrid application server, consisting of multiple execution environments acting as transaction coordinators, can be recovered in an efficient and graceful manner. This features intends to address these requirements.

Problem explained in detail:

Consider a clustered deployment of a hybrid application server as shown in Figure 1.1. For simplicity, consider the case where there are only two cluster members. Each of these hybrid application server consist of multiple execution environments, many of which could act as transaction co-ordinators. Usually, the transaction logs of these transaction coordinators are kept in a centrally accessible location like a NAS, so that, in case any of the execution environment fails, another one in a different cluster member could recover its transactions. High availability manager (sometimes also called as HAManager) takes care of the responsibility of choosing the hybrid app server which should do the recovery etc. This is simple and straightforward for a static cluster consisting normal application server, which usually consists of only o...