Data processing: software development – installation – and managem – Software program development tool – Testing or debugging
Reexamination Certificate
2001-03-26
2004-12-28
Courtenay, III, St. John (Department: 2126)
Data processing: software development, installation, and managem
Software program development tool
Testing or debugging
C710S260000
Reexamination Certificate
active
06836881
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to the remote tracing of a plurality of data processing nodes, which are connected to each other via a network.
BACKGROUND OF THE INVENTION
As computer systems increase in size and complexity, it has become common to distribute data processing applications over a plurality of data processing nodes, communicating over a data communications network, for example, the Internet. This allows a large data processing task to be distributed over several data processing nodes whilst also allowing exchange of data messages, which can consist of requests for processing or replies containing the results of processing. The data processing nodes can be arranged so that a local data processing node can control a remote data processing node. However, it is difficult to monitor operation of remote nodes to facilitate error detection and correction.
Tracing operations are known in the prior art for the purpose of tracing the path of execution of an application to assist in locating errors therein. The tracing operation assists in problem determination by providing a snapshot record in storage of certain types of states existing when a location in an application is reached by the data processing node that is running the application. Such events or states are often stored in a trace table or in trace files in memory.
Static or local tracing operations in response to commands issued within a node is known in the prior art. However, these commands cannot be used to initiate remote traces or dynamic alteration of commands at runtime. Extending tracing operations to remote nodes may involve large overheads and is often intrusive. It is difficult to achieve an acceptable trade off between response times and existing hardware, as the link between data processing nodes may be slow, limiting the efficiency of data exchange. Additionally, if the results from the remote operations are written into a file for subsequent perusal or if the remote data processing node only periodically reports to the local data processing node, this will further delay response times.
Tracing operations are often implemented prior to debugging operations in order to isolate a problem area. Debuggers are tightly coupled to the processes which are targeted and are therefore effective in dealing in detail with the problem identified by the tracing operation. Debugging generally involves stepping a program one step at a time, through all possible paths of execution and monitoring its behaviour. This differs from monitoring the normal execution of the program in a production mode, when execution events are traced in real-time.
Remote debugging to facilitate problem determination and replication of real-time conditions on remote systems is known, for example, as described in U.S. Pat. No. 5,630,049. Although avoiding problems of prior remote debuggers arising from the fact that the data processing nodes typically must be in close physical proximity to each other, which limits the flexibility of the testing environment, the debugging process carries a large overhead and is relatively intrusive in comparison with a trace process. Problems with data processing nodes on a customer site often occur in unpredictable circumstances and the fine control and detailed activity of a debugger is not required for first pass problem analysis.
U.S. Pat. No. 5,630,049 also employs a method of asynchronous messaging in which messages between nodes of a network may be transmitted using process-private interrupts known as Asynchronous System Traps (ASTs). ASTs enable a faster and more reliable communication than with general asynchronous messaging.
SUMMARY OF THE INVENTION
However, there is still a need for a remotely controlled tracing operation in a network of data processing nodes, which executes with minimum intrusion into the data processing nodes. There is also a need for the remote tracing operation to respond dynamically in an acceptable time frame and without limiting the physical proximity of nodes from each other.
Accordingly, the present invention provides a method for remote tracing from a local data processing node of the execution of a process within an application program running on a remote data processing node in a distributed data processing network, said application program including its own local trace facility, said nodes communicating by asynchronous messaging via a data exchange means and each node including process-private interrupt handling means for indicating the presence of a command for the respective process in said data exchange means, said method comprising the steps of sending a trace command from a trace process running on said local data processing node into a data exchange means of said remote data processing node; in response to said trace command, causing a process-private interrupt of a target process running on said remote data processing node; in response to said process-private interrupt, said target process writing trace information from said trace facility to said data exchange means; transmitting said trace information across said network; receiving in a data exchange means on said local data processing node, said transmitted trace information; in response to receiving said trace information, causing a process-private interrupt of said trace process; and in response to said process-private interrupt, reading said trace information by said trace process, from said local data exchange means.
Specifically, the tracing operation is advantageous as it can be executed with networked data processing nodes. This allows an end user on a local data processing node to dynamically perform tracing and diagnostic operations of remote data processing nodes. For example, a service provider can perform online diagnostics of data processing nodes located on a customer site, assuming a secure architecture.
In a further preferred aspect of the present invention, there is provided a distributed data processing system comprising a plurality of data processing nodes connected via a network, each node having a processor, memory and operating system capable of executing application programs, each of said operating systems including data exchange means and interrupt handling means, a first of said nodes comprising means for sending a trace command from a trace process running on said first data processing node into the data exchange means of a second data processing node, said second data processing node including a trace facility for tracing the execution of a process within an application program running on said second node, said interrupt handling means of said second data processing node in response to said trace command, causing a process-private interrupt of a target process running on said second data processing node; said second data processing node further including means for writing trace information from the trace facility, to said second data exchange means in response to the process-private interrupt; and means for transmitting said trace information across said network whereby said first data exchange means receives said transmitted trace information at said first data processing node; and in response to receipt of said trace information, said first interrupt handling means causes a process-private interrupt of said trace process; said first data processing node further including means for reading said trace information, from said first data exchange means in response to said process-private interrupt.
In other aspects, the present invention provides a computer program for remote tracing of data processing nodes in an asynchronous messaging network
In a further preferred aspect of the present invention, there is provided a method in which the step of sending a trace command further comprises the following steps. Firstly, a trace command is written from a trace process into the local data exchange means. In response to the trace command, the process private interrupt on the local data processing node is initiated. Next, the trace command is transmitted across the netwo
Beynon Stephen John
Mihell Brian Kenneth
Swinson Andrew Mark
Winn Geoffrey Martin
Courtenay III St. John
Herndon Jerry W.
International Business Machines - Corporation
Nguyen Van
Synnestvedt & Lechner LLP
LandOfFree
Remote tracing of data processing nodes in an asynchronous... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Remote tracing of data processing nodes in an asynchronous..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Remote tracing of data processing nodes in an asynchronous... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3337528