HP Integrity NonStop H-Series - Manual (Page 18)

2 Features of NSASJ
This chapter describes the following features of NSASJ inherited from NonStop platform:
•
“High Availability” (page 18)
•
“Scalability” (page 20)
High Availability
High Availability (HA) is the ability of NSASJ to maintain the business continuity and service
availability at all times by surviving hardware or a network failure. In other words, HA is the ability
of NSASJ to achieve the following:
•
Ensure that the business services are available and survive any single point of failure in any
of the system components.
•
Have zero planned downtime, which keeps the business services available even during
maintenance time frame.
•
Keep the business services available even when software components on which NSASJ is
dependent fail.
HA for NonStop Application Server for Java (NSASJ)
NSASJ integrates with NonStop TS/MP to achieve HA feature. The components comprising NSASJ
are configured as TS/MP serverclass and hence, TS/MP performs process management. The
following list describes the functionalities of the different components of NSASJ which helps in
achieving HA:
•
Post Master: The main purpose of the Post Master is to listen on the public interfaces of NSASJ
which are remoting port, http and https connector ports and transfer all data to the NSASJ
serverclass. Post Master does not participate in data processing. It just acts as an agent that
transfers data between the remoting client and NSASJ serverclass. If the TCP/IP provider is
IP-CIP, then there is no limit on the number of instances of Post Master that can be started. If
the TCP/IP provider is TCPIPv6, then one instance of Post Master can be started on each CPU.
As the EJB client application invokes the remote method, the EJB client library marshals the
data required for remote invocation and sends the data over the established TCP connection
to Post Master. The Post Master then transfers data to the NSASJ serverclass instance with
which it has already established a connection. When a web client accesses a servlet or a JSP,
the HTTP requests are sent to the Post Master which then forwards it to NSASJ Instances. The
HTTP responses from the NSASJ Server instances are sent back to the client by the Post Master.
NOTE: Component availability is referred to with respect to HA and not the business logic.
The ongoing transactions might be lost.
In case of a failure of a Post Master, though the connections that it handled are dropped, the
new connection from the client is handled by some other instance of Post Master. This
contributes to the continuity of the business services offered by NSASJ.
•
NSASJ Server: A server is a single instance of NSASJ Java EE container that provides an
implementation of a specific set of Java EE technologies. It is configured as a TS/MP serverclass
and all the instances of the server share the same configuration and file system resources.
Requests from Post Master are routed to the NSASJ sever class using Pathsend API. Since the
requests are routed using Pathsend API, TS/MP ensures that there are instances available and
running to serve the requests.
In case there are no instances available to serve these requests, new instances are created to
service the requests coming through Pathsend API. Thus, TS/MP ensures that a specific number
18 Features of NSASJ