JP3513484B2 - Management system for parallel computer system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [0001] The present invention relates to a parallel computer system.
In particular, a parallel computer system is configured for the system management device.
Main processor of multiple nodes not working
Perform maintenance and management of the plurality of nodes even in the case
Effective technology applied to the management system of parallel computer systems
It is about. [0002] 2. Description of the Related Art Conventionally, a computer comprising a plurality of computers
Regarding the system operation management method and its execution device,
Several have been proposed. The control of a plurality of UNIX (registered trademark) machines
Operation and monitoring problems that occur when the soles are combined into one
Centralized operation and operation of multiple UNIX machines to prevent
For the monitoring console display, see
No. 2,147,633. [0004] The outline is to concentrate multiple UNIX machines
Center console, which is a server to operate and manage
Destination table that stores command destinations for each operation purpose
File according to the destination table.
Is to execute. [0005] A composite computer system comprising a plurality of computers
Connected by a single system console on the system
Error when performing maintenance and operation by switching computers
Console in compound computer system to prevent operation
The switching control method is disclosed in JP-A-5-120247.
It is stated in the report. [0006] An overview of this is based on service programs in a plurality of computers.
Connect the processors to a switching device, and connect the
Connect the system console and from the system console
Outputs a message using an identifier that identifies the computer.
By switching the target computer sequentially, multiple
In this method, one system console is shared by computers.
When performing maintenance and operation using the system console
The identifier of the computer on which the operation is to be performed and the system
Compare the identifiers of computers connected to the console
Perform the operation if the identifiers match.
You. [0007] From each computer constituting the distributed processing system
Centralized message management system for centralized management of messages
This is described in JP-A-5-20281.
You. [0008] The outline is that the
A central management node is determined in multiple computers, and the central management
Operation status messages issued by the monitored node
This is a method for centrally managing the data. [0009] SUMMARY OF THE INVENTION The inventor of the present invention
As a result of studying the technology, the following problems were found. That is, the conventional computer comprises a plurality of computers.
The management device of the computer system
Use the functions of network software running on
If the computer to be managed is not
If the operating system is not operating and the network
If the network software is not running,
There was a problem that management could not be performed. Centralized operation of the conventional plurality of UNIX machines
And how to use the monitoring console display
The managed computers are operating system
Operating system UNIX is assumed to be operating.
If the rating system is not running,
Central operation and monitoring cannot be performed from the sole display
There was a problem. [0012] The core in the conventional complex computer system is
Console switching control, the system console and each
There is a switching device between the service processor
Therefore, special hardware such as a switching device will be required
There was a problem. In the conventional message centralized management system,
Messages are sent from multiple computers to the central management node.
The central management node goes down.
The centralized management of messages
And the message is the network to which the node is connected.
Operation via each node
If the operating system and network are not running
In this case, the status of each node must be managed from the central management node.
There was a problem that can not be. [0014] [0015] [0016] [0017] [0018] [0019] [0020] [0021] [0022] The present inventionEyesThe goal is to run parallel computer systems.
Provide technology that enables remote management of
And there. The above and other objects of the present invention and a novel
Features will be apparent from the description of this specification and the accompanying drawings.
Will be. [0025] SUMMARY OF THE INVENTION The present invention is disclosed.
Of the inventions, a brief description of the typical ones is as follows:
It is as follows. [0026] [0027] [0028] [0029] [0030] [0031] [0032] [0033] [0034] [0035] [0036] [0037] [0038] [0039] [0040] [0041] [0042] [0043] [0044] [0045] [0046] [0047] [0048] [0049] [0050] [0051] [0052] [0053] [0054] [0055] [0056] [0057] [0058] [0059] [0060] [0061] [0062] [0063] [0064] [0065] [0066] [0067] [0068] [0069] [0070] [0071] [0072] [0073] [0074]Parallel connection of multiple computer nodes
Equipped with a management terminal device for managing the computer systemParallel meter
In the management device of the computer system,The plurality of nodes
Operates on the main power supply of the node and executes parallel processing.
Powered by the main processor and the auxiliary power supply of the node.
System control frame for managing the main processor
And a sub-processor for executing commands.
Operate and communicate with the management terminal device, the management terminal
Control based on system control commands issued from the device
System control mechanism for generating data and system control
Connect to the management terminal on a different network
The application of the node under the control of the management terminal device.
LAN control mechanism to manage application software
Prepare,The management terminal device,A plurality of said plurality of nodes
A system control mechanism that communicates with the system control mechanism;Supplement
Operates on the auxiliary power supply, and when a specific signal is input, the management terminal
Power-on logic for turning on the main power of the device, and power-on
When the main power is turned on by logic, the plurality of nodes
Turn on the main power supply to the sub-processors in batch or individually
System control commandsVia the system control mechanism
soMeans to sendAnd a plurality of systems of the plurality of nodes.
Stem control mechanism and system control mechanism of the management terminal device
And a plurality of main nodes of the plurality of nodes.
The system control command for managing the processor is transmitted to the management terminal.
Device to a plurality of system control mechanisms of the plurality of nodes
System control interface to send toWith
is there. BeforeOfWith the management device of the parallel computer system
Is operated by an auxiliary power supply and transmits a specific signal to the management terminal device.
Input the main power of the management terminal device.
Connect the power-on logic and connect the power-on logic to the network
And other communication lines. Further, the main power of the management terminal device is turned on.
At the end of the system startup process that is executed when
Collectively or individually to the sub processors of the plurality of nodes
A program that sends a system control command to turn on the main power
Add a gram. Next, the network and other communication lines are connected.
A specific signal from another terminal device to the power-on logic.
And the main power supply of the management terminal device is turned on. The main power of the management terminal device is turned on.
And the system startup processing of the management terminal device was performed.
Later, collectively or individually to the sub-processors of the plurality of nodes.
Send a system control command to turn on the main power separately
The program is executed and the parallel computer system is operated.
To start, the operator directly operates the management terminal device
It can be done without. As described above, the parallel computer system
According to the management device, the management terminal
Since the main power of the device is turned on, the operation of the parallel computer system
Management can be performed from a remote location. [0080] BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, the present invention will be described with reference to embodiments.
This will be described in detail with reference to the drawings. Note that, in all drawings for explaining the embodiment, FIG.
And those having the same function are given the same reference
The description of the return is omitted. (Embodiment 1) The parallel computing of the present invention will be described below.
Configuration of Embodiment 1 for Implementing Management Device for Machine System
Will be described. FIG. 1 is a diagram showing a pipe of a parallel computer system according to the present invention.
1 is a diagram showing a schematic configuration of a first embodiment for implementing a processing device.
You. In FIG. 1, 100a to 100c are nodes, 10
1 is a communication processing device, and 102a to 102c are communication interfaces.
Ace mechanism, 103a to 103c are communication cables, 10
4 is a management terminal device, 105a to 105d are system control
Mechanism, 106a to 106d are communication cables, 107 is
Network concentrators 108a-108d are LAN (L
ocal Area Network) control mechanism, 10
9a to 109d are communication cables, 110 is a network
It is a line concentrator. As shown in FIG. 1, the parallel computer of this embodiment
System management device constitutes a parallel computer system
Nodes 100a to 100c and node 10 in parallel processing
Communication processing device 101 for controlling communication in 0a to 100c
And a pipe for performing system management of the nodes 100a to 100c.
Management terminal device 104, management terminal device 104 and node 10
Network concentrator 10 for connecting 0a-100c
7 and a network concentrator 110.
The management terminal device 104 includes a system control mechanism 105a and L
The node 100a has an AN control mechanism 108a.
Communication interface mechanism 102a and system control mechanism 1
05b and the LAN control mechanism 108b, and the node 1
00b is a communication interface mechanism 102b and a system.
System control mechanism 105c and LAN control mechanism 108c.
And the node 100c communicates with the communication interface
c, the system control mechanism 105d, and the LAN control mechanism 1
08d. Further, as shown in FIG.
In the management device of the computer system, the nodes 100a to 10a
0c through the communication interface mechanisms 102a to 102c.
Communication cables 103a to 103c and communication processing device 101
Through the nodes 100a to 100c
The control mechanisms 105b to 105d are connected to the communication cables 106a to 106d.
Management via 106d and network concentrator 107
Connect to the system control mechanism 105a of the terminal device 104,
LAN control mechanism 108b of nodes 100a to 100c
108d is a communication cable 109a-109d and a network
L of the management terminal device 104 via the work concentrator 110
It is connected to the AN control mechanism 108a. The management system of the parallel computer system of the present embodiment
System control interface, as described above,
System control mechanism 105a on terminal device 104 side and node
System control mechanisms 105b-1 on the 100a-100c side
05d to a communication cable such as Ethernet (registered trademark)
Concentration of network such as 106 and multiport repeater
Implemented by interconnecting using device 107
Interface. The parallel computer system according to the present embodiment
The system operation support interface of the management device is
LAN control mechanism 108a on the terminal 104 side and the node 10
LAN control mechanisms 108b to 108d on the side of 0a to 100c
And communication cable 109 such as Ethernet
-Using a network concentrator 110 such as a repeater
An interface realized by interconnecting
You. The system operation support interface includes:
An interface that manages the operation of a conventional parallel computer system
And the main process of the nodes 100a to 100c.
The node 100a-1 is used when the server is operating.
Application running on main processor 00c
Management software to output messages output by the
System management such as display on the display 104 is performed. The parallel computer system according to the present embodiment will be described below.
Of the parallel computer system in the management device
The nodes 100a to 100c will be described. FIG. 2 shows a parallel computer system according to the present embodiment.
Of the parallel computer system in the management device
FIG. 2 is a diagram illustrating a schematic configuration of nodes 100a to 100c. In FIG. 2, reference numeral 200 denotes a main power supply, and 201 denotes a main power supply.
Auxiliary power supply, 202 is main processor, 203 is software
Hardware, 204 is the main memory, 205 is the processor
Memory control mechanism, 206 is system bus, 207 is system
Disk, 208 is an I / O control mechanism, 209 is an RS-
232C control mechanism, 210 is bootstrap ROM
(Read Only Memory), 211 is cis
System support mechanism, 212 is a sub processor, 213 is
ROM and 214 are SRAM (Static Lando)
m Access Memory; nonvolatile memory), 2
15 is a local bus, 216 is a power on / off signal, 2
17 is a processor reset signal, 218 is LAN control
Section, 219 is an RS-232C control section, and 220 is a processor.
221 is a ROM, 222 is a RAM (Random)
Access Memory) 223 is a data interface.
A face 224 is a control interface. As shown in FIG. 2, the parallel computer of this embodiment
The nodes 100a to 100c of the management device of the system
Communication interface mechanism 102a-102c and system
Control mechanisms 105b to 105d and LAN control mechanism 108
b to 108d, and the nodes 100a to 100c
Run application software that performs column processing
A main processor 202 and a sub-processor 212
System support mechanism 211, main power supply 200,
An auxiliary power supply 201 is provided. The parallel computer system according to the present embodiment
The nodes 100a to 100c of the management device
Operating system executed by the server 202
And network software 20
3 and main memory 20 for storing software 203
4, the main processor 202 and the main memory 204
Memory Control Mechanism for Interface Control
205, a system bus 206, and a system disk 2
07 and I / O control for controlling the system disk 207
Mechanism 208, node message output and system control
Interaction with the operator via mechanisms 105b-105d
RS-232C control mechanism 20 for active exchange
9 and system startup processing of the main processor 202
Bootstrap that contains the bootstrap program
And a strap ROM 210. The management system of the parallel computer system of the present embodiment
In the configuration, a sub processor 212 is provided,
Performs system control such as status management of the processor 202
The system support mechanism 211 includes a sub-processor 212
ROM 21 storing a control program that operates on
3 and SR storing hardware-dependent information
AM 214 is provided. The management system of the parallel computer system of the present embodiment
System control mechanism 10 of each of the nodes 100a to 100c
5b to 105d are Ethernet communication with the management terminal device 104.
LAN control unit 218 for controlling transmission and reception of net packets
, RS-232C control mechanism 209 and sub-processor
Control the transmission and reception of RS-232C packets between
Controlling the RS-232C control unit 219 and the Ethernet
Protocol conversion between packet and RS-232C packet
And a processor that operates on the processor 220.
ROM 221 for storing a control program
Sent from the processor 212 and the RS-232C control mechanism 209
RAM 222 for storing incoming node messages
It has. As shown in FIG. 2, the parallel computer of this embodiment
In the nodes 100a to 100c of the system management device,
The system control mechanisms 105b to 105d are
C control unit 219, data interface 223, and R
S-232C control mechanism 209 and system bus 206
Through the processor memory control mechanism 205
Connected to the processor 202;
05b to 105d are transmitted to the RS-232C control unit 219 and
System support mechanism via the control interface 224
Connected to the sub-processor 212 of the
The sub processor 212 of the
215 and the processor memory control mechanism 205
It is connected to the main processor 202. In addition,
The processor 212 responds to the processor reset signal 217.
Reset the main processor 202 and turn on / off the power.
The main power supply 200 is controlled by the disconnection signal 216. Note that, as shown in FIG.
To the nodes 100a to 100c of the management system of the computer system
In addition, the system control mechanisms 105b to 105d
-232C control unit 219 and data interface 22
3 and the main unit via the RS-232C control mechanism 209.
Connected to the processor 202 is a system controller
Between the structures 105b to 105d and the main processor 202
With a serial interface such as RS-232C.
Communication software
And when a failure occurs in the main processor 202.
If any, the system control mechanisms 105b to 105d and the main
Communication with the processor 202 is impaired as much as possible.
This is to ensure that they are not lost. The management system of the parallel computer system of the present embodiment
Nodes 100a to 100c operate on the main power supply 200
And a part operated by the auxiliary power supply 201.
ing. The parts operated by the main power supply 200 include:
The main processor 202 of the
Main memory 204 for storing software 203,
Interface between the main processor 202 and the main memory 204.
Processor memory control mechanism 205 for performing
Of the main processor 202 of each of the cards 100a to 100c.
Bootstrap program that performs system startup processing
Bootstrap ROM 210 that stores
These are connected to the communication interface via the system bus 206.
Interface mechanisms 102a to 102c, LAN control mechanism 1
08b to 108d are connected.
207 is connected via an I / O control mechanism 208. The parts operated by the auxiliary power supply 201 include:
Control of main power supply 200 of nodes 100a to 100c and main
Control such as status management of the processor 202
And a system support mechanism 211 that performs
Communication between the nodes 100a to 100c and the management terminal device 104
There are system control mechanisms 105b to 105d for controlling. [0101] The sub-processor 212
The power on / off signal 216 is output according to the instruction from the controller 04
By controlling the main power supply 200, the processor
By outputting the set signal 217, the main processor
It has a function to reset 202. The node messages of the nodes 100a to 100c
The main processor 202 operates and the main
An operating system that controls the processor 202;
Software 203 which is network software
In the activated state, the data interface 223
From the RS-232C control mechanism 209 to the RAM 22
2 and the software 203 is not running
In this state, the control interface
Bootstrap message or the like via source 224
Stored in AM 222. The management system of the parallel computer system of this embodiment
Processor of the system control mechanism 105b to 105d
220 is the following in addition to the protocol conversion of the packet.
The following processing is also performed. That is, the e-mail from the management terminal device 104
Interpret the Ethernet packet and process according to the contents of the packet.
And RAM in accordance with an instruction from the management terminal device 104.
Management terminal device storing the node message stored in the storage device 222
A process for transmitting to the sub-processor 212 is performed.
Is the password sent through the control interface 224.
Interpret the packet and perform control according to its contents. The parallel computer system according to the present embodiment will be described below.
The general configuration of the management terminal device 104 of the management device of FIG.
I will tell. FIG. 3 shows a parallel computer system according to the present embodiment.
Is a diagram showing a schematic configuration of a management terminal device 104 of the management device of FIG.
is there. In FIG. 3, reference numeral 300 denotes a processor, and 301 denotes a software.
Software, 302 is main memory, 303 is boots
A trap ROM 304, a processor memory control mechanism,
305 is a system bus, 306 is an I / O control mechanism, 30
7 is a system disk, 308 and 309 are RS-232
C control mechanism, 310 is a graphics control mechanism, 311
Is a LAN controller, 312 is an RS-232C controller, 31
3 is a processor, 314 is a ROM, 315 is a RAM, 3
16 is a control interface, 317 is a data interface
It is a source. As shown in FIG. 3, the parallel computer of this embodiment
The management terminal device 104 of the system management device is a management terminal
A processor that controls and controls all processes in the device 104
300 and the operating system of the management terminal device 104.
System and network software
A main memory 302 in which the memory 301 is stored;
Boots up to perform system startup processing of
Bootstrap RO storing wrap program
M303, processor 300, main memory 302 and
And interface control of bootstrap ROM 303
And a processor memory control mechanism 304 for performing
You. The parallel computer system of the present embodiment
The management terminal device 104 of the management device includes a system bus 305
And an I / O control mechanism for controlling the system disk 307
306, system disk 307, and software 3
01 for the nodes 100a to 100c such as power control.
RS-23 used when issuing a stem control command
The 2C control mechanism 308 is connected to the node message output and system
Interaction with the operator via the system control mechanism 105a
RS-232C control mechanism 30 for active exchange
9 and display terminal, keyboard and mouse
Graphics that control man-machine interfaces such as
Control mechanism 310, the system control mechanism 105a,
It has. The management system of the parallel computer system of this embodiment
System control mechanism 105a of the management terminal device 104
Is an Ethernet path between the nodes 100a to 100c.
A LAN control unit 311 for controlling transmission and reception of the
RS-232C between control mechanisms 308 and 309
RS-232C system for controlling transmission and reception of 232C packets
Control unit 312, Ethernet packet and RS-232C
Processor 313 that performs protocol conversion with packets
And a control program operated by the processor 313 are stored.
ROM 314 and the nodes 100a to 100c
RAM 315 for storing incoming node messages
It has. Further, as shown in FIG.
In the management terminal device 104 of the management device of the computer system,
Processor 300 to processor memory control mechanism 304
Main memory 302, bootstrap ROM 3
03 and the system bus 305,
To the system bus via the I / O control mechanism 306.
305, and the LAN control mechanism 108a and the RS-2
32C control mechanisms 308 and 309 and graphics system
The control mechanism 310 is connected to the system bus 305. Further, as shown in FIG.
In the management terminal device 104 of the management device of the computer system,
RS-232C control unit 31 of system control mechanism 105a
2 with the control interface 316 and the data interface
RS-232C control mechanism 308 via the source 317 and
309. The management system of the parallel computer system of the present embodiment
, The system control interface is node 1
00a to 100c system control mechanisms 105b to 105
d and the system control mechanism 105 a of the management terminal device 104.
Must be interconnected using an Ethernet cable, etc.
Is formed. The system control interface is
The system control mechanism 105a on the terminal device 104 side can operate
And the auxiliary power supply for the nodes 100a to 100c
201 is turned on, and the sub-processor 212 and the
With the stem control mechanisms 105b to 105d operable
If any, the main power supply 200 of the nodes 100a to 100c is turned on.
The main processor 202 is not running.
Control the entire main processor 202
Operating system and network software
Even if the software 203 is not running.
Is available. In contrast, the system operation support interface
Source is the LAN control mechanism 108a of the management terminal device 104.
And LAN control mechanism 108b of nodes 100a to 100c
To 108d using Ethernet cable etc.
The system operation is formed by connecting
The support interface is TCP / IP (Transmi
session Control Protocol / In
telnet protocol)
E of the management terminal device 104 and the nodes 100a to 100c.
Operating system and its network software
Software 203 and software 301
Is started and a network that supports TCP / IP
Available only when the software is running
You. Hereinafter, the parallel computer system according to the present embodiment will be described.
Terminal device 104 and the nodes 100a to 1
The communication sequence with 00c will be described. FIG. 4 shows a parallel computer system according to this embodiment.
Terminal device 104 and the nodes 100a to 1
It is a figure which shows an example of the communication sequence with 00c. FIG.
, 401 is an adapter control command and its
And 402, a system control command and its response.
403 is a node message. As shown in FIG. 4, the parallel computer of this embodiment
In the system management device, the adapter control command and its
Response 401 or a system control command and
And transmission and reception of the packet which is the response 402
Communicates by sending and receiving node messages 403.
U. Adapter control command and its response
Reference numeral 401 denotes the software of the management terminal device 104.
Communication with the system control mechanism 105a of the management terminal device 104
And the sub processor 212 controls the system.
Used to communicate with the mechanisms 105b to 105d and control
Interface 316 or control interface 224
Sent and received via System control command and its response
402, the software 301 of the management terminal device 104
Communication with the sub-processors 212 of the nodes 100a to 100c
Control interface 316 and control
It is transmitted and received via the control interface 224. The node message 403 is software
When the sub-processor 212 is not running,
To the RAM 22 of the system control mechanisms 105b to 105d
2 and stored, and the software 203
When activated, the main processor 202
System control mechanism 10 from S-232C control mechanism 209
5b to 105d RAM 222
You. R of system control mechanisms 105b to 105d
The node message 403 stored in the AM 222 is
In response to a request from the management terminal device 104, the nodes 100a to 100a
RA of the system control mechanism 105b-105d of 100c
From M222, the system control mechanism of the management terminal device 104
105a, the RS-232C of the management terminal device 104
Sent to the control mechanism 309,
Graphics connected to fix control mechanism 310
It is displayed on a display or the like. Hereinafter, the parallel computer system according to the present embodiment will be described.
Control command and its control in the management device
The packet format of the response 401 will be described.
You. FIG. 5 shows a parallel computer system according to this embodiment.
Control command and its control in the management device
FIG. 3 is a diagram showing a packet format of a Ponce 401.
In FIG. 5, reference numeral 501 denotes a type field, and 502 denotes transmission.
Source address field, 503 is the destination address field
Field, 504 is an information field, and 505 is an identifier.
You. As shown in FIG. 5, the parallel computer of this embodiment
An adapter control command in the system management device;
The packet of the response 401 has a type field 5
01, a source address field 502, and a destination address.
Dress field 503 and information field 504
And an identifier 505. The management system of the parallel computer system of the present embodiment
In the type field 501, the adapter control
Command or its response packet
An identifier, for example, “A” is stored in the
Field 502 contains the source address and destination address of the packet.
The destination address of the packet is stored in the dress field 503.
Is stored. In the information field 504, the packet
Different parameters are stored depending on the type of
In addition, the end of the packet has an identification that indicates the end of the packet.
Child 505, for example, "LF" (line feed) is added.
It is. Hereinafter, the parallel computer system according to the present embodiment will be described.
Control command and its control in the management device
The packet format of the Ponce 402 will be described.
You. FIG. 6 shows a parallel computer system according to this embodiment.
Control command and its control in the management device
FIG. 4 is a diagram showing a packet format of a Ponce 402.
In FIG. 6, reference numeral 601 denotes a type field, and 602 denotes transmission.
Source address field, 603 is the recipient address field
Field, 604 is an information field, and 605 is an identifier.
You. As shown in FIG. 6, the parallel computer of this embodiment
A system control command in a system management device; and
The packet of the response 402 has a type field 6
01, a source address field 602, and a destination address
Dress field 603 and information field 604
And an identifier 605. The management system of the parallel computer system of the present embodiment
In the field, the type field 601 includes the system control
A packet that indicates a command or its response
A packet identifier, for example, "d" is stored in the
The field 602 contains the source address and the destination of the packet.
The address field 603 contains the destination address of the packet.
Is stored. The information field 604 contains a packet
Different parameters are stored depending on the type of
At the end of the packet, an identifier indicating the end of the packet
605, for example, “LF” is added. Further, the parallel computer system of the present embodiment
In the management device, the transmission packet from the management terminal device 104 is transmitted.
Hexadecimal number in the destination address field 603 of the
When “0xffffffff” is stored, the packet
Is a broadcast packet, and all nodes 1
00a to 100c. The pipes of the parallel computer system of the present embodiment
In the processor, the number with “0x” added is a hexadecimal number
Shall be shown. Hereinafter, the parallel computer system according to the present embodiment will be described.
Control mechanisms 105a to 105 in the management device of
d, mode transition between packet mode and non-packet mode
Will be described. FIG. 7 shows a parallel computer system according to this embodiment.
Shows the mode transition of the system control mechanism in the management device
FIG. In FIG. 7, reference numeral 701 denotes a packet mode;
702 is a non-packet mode, and 703 is “SET-MOD
E "command. As shown in FIG. 7, the parallel computer of this embodiment
The system control mechanism in the system management device is fixed
A packet mode 701 for transmitting and receiving long packets,
Non-packet for sending and receiving node message 403 of undefined length
A packet mode 701, and a packet mode 702.
The mode transition of the packet mode 702 is performed by the sub processor
"SET-M" which is an adapter control command
This is performed by executing the “ODE” command 703. As described above, the parallel computer system of this embodiment is
System control mechanisms 105a to 105d of system management devices
The operation modes are packet mode 701 and non-packet mode.
There are two types of modes 702, the packet mode 701
Is connected to the management terminal device 104 and the plurality of nodes 100a to 100a.
0c is a mode set when performing communication,
The connection mode 702 is for connection-oriented communication with a specific node.
And sends a node message 4 from the specific node.
03 to be displayed on the management terminal device 104.
Mode. The pipes of the parallel computer system of the present embodiment
In the management device, the management terminal device 104 and the node 100
a to 100c system control mechanisms 105a to 105d
Switches to the packet mode 701 when the auxiliary power supply 201 is turned on.
It works. Hereinafter, the parallel computer system of this embodiment will be described.
Control mechanisms 105a to 105 in the management device of
Transition of connection state in non-packet mode 702 of d
The transfer will be described. FIG. 8 shows a parallel computer system according to this embodiment.
Non-packet mode of the system control
FIG. 10 is a diagram showing transition of the connection state in the node 702.
8, reference numeral 800 denotes a disconnected state, and 801 denotes a disconnected state.
Wait connected state, 802 is connected state, 803
Is "SET-CONNECT" command, 804 is management
System control mechanism 105a and node on terminal device 104
System control mechanism 105b-1 on 100a-100c
05c. As shown in FIG. 8, the parallel computer of this embodiment
Non-package of system control mechanism in system management device
In the connection state in the cut mode 702, the partner
No stem control mechanism is connected and no
Disk message 403 is not stored.
Connected state 800 is connected to partner system control mechanism
The node message 403 is not stored in the RAM 222
A wait connect state 801 that is accumulating in the
Connected state where the partner system control mechanism is connected
802, and the transition of the connection state is “S
ET-CONNECT "command 803 or system
This is performed by call control 804 from the control mechanism. As shown in FIG. 8, the parallel computer of this embodiment
In the system management device, the non-packet mode 702
When setting, disconnect state 800, wait connect
Connection state 801 and connection state 802
In the disconnect state 800, the system state is maintained.
Communication between the system control mechanisms becomes impossible, and wait connection
In the default state 801, communication with the partner system control mechanism is
Although not possible, the node message 403 is stored in the RAM 22
2 are sequentially stored. The system control mechanisms for communication are connected
In the non-packet mode 702 when in the
Connection type communication becomes possible. These states are described in “SET-CONNECT”.
The transition is made by issuing a “T” command 803. Ma
Also, call control 804 from the partner system control mechanism
If there is a connect request, wait connect state 80
The state transits from 1 to the connect state 802. The management system of the parallel computer system of the present embodiment
Control commands used by the device and their responses
Table 1 shows a list of the resources 401. In Table 1, the information section
The information stored in the information field 504 is shown.
The number in byte 0 of the information part is a number indicating the packet type.
No. [0146] [Table 1] The management system of the parallel computer system of the present embodiment
Control commands used by the device and their responses
Table 2 shows a list of the resources 402. In Table 2, the information section is
The information stored in the information field 604 is shown.
The number in byte 0 of the information part is a number indicating the packet type.
No. [0148] [Table 2]The following describes the parallel computer system of the present embodiment.
Of the nodes 100a to 100c in the management device
Of the processor 220 of the system control mechanisms 105b to 105d.
The management procedure will be described. FIG. 9 shows a parallel computer system according to this embodiment.
Of the nodes 100a to 100c in the management device
Of the processor 220 of the system control mechanisms 105b to 105d.
9 is a flowchart showing a part of a management procedure. As shown in FIG. 9, the parallel computer of this embodiment
Nodes 100a to 100c in system management device
Processor 2 of the system control mechanism 105b to 105d
In step 20, in the process of step 900, “SET-CO
NNECT ”command or call control by the management terminal 10
4 to check if there is a connect request. In the process of step 900, “SET-CO
NNECT ”command or call control by the management terminal 10
If there is a connect request from step 4, step 901
To the system control of the nodes 100a to 100c.
Control mechanism 105b-105d is in non-packet mode 702
Check if there is. In the process of step 901, the node 100a
To 100c are not controlled by the system control mechanisms 105b to 105d.
If it is the packet mode 702, the process of step 902
The processing proceeds to step 901, and the nodes 100a-1
00c is a non-packet.
If not in the cut mode 702, the
And a system control command from the sub processor 212.
By “SET-MODE”, the nodes 100a to 100
non-packet with the system control mechanisms 105b to 105d
The mode is set to mode 702, and the process proceeds to step 902. In the process of step 902, the node 100
a to 100c of the system control mechanisms 105b to 105d
The node message 403 stored in the RAM 222 is
To management terminal device 104 via stem control interface
Then, the process returns to step 900. In the processing of step 900, “SET-CO
NNECT ”command or call control by the management terminal 10
If there is no connect request from step 4,
Proceed to the process, and in the process of step 904, the system control
The system control from the management terminal device 104 is executed by a command.
Check to see if you have. In the processing of step 904, the system
A system from the management terminal device 104 is issued by a control command.
If there is a program control, the process proceeds to step 905,
System control mechanism 105b of nodes 100a to 100c
Check if ~ 105d is in packet mode 701
I do. At step 904, the system control
System control from management terminal device 104 by command
If not, the process proceeds to step 909. In the process of step 905, the node 100a
100c are controlled by the system control mechanisms 105b to 105d.
If it is the packet mode 701, the process proceeds to step 906.
Go to the system control mechanism 1 of the nodes 100a to 100c
If 05b to 105d are not the packet mode 701,
In the processing of step 907, the sub-processor 212
By the system control command "SET-MODE" of
Set the system control mechanisms 105b to 105d in packet mode
701, and the process proceeds to step 906. In the processing of step 906, the system
Check the destination address field 603 of the control command.
Check the address of the system control command
Field 603 has its own logical address or "0xffff
ffffff ”, the process proceeds to step 908.
The contents of the system control command
Notification is sent to 212, and the process returns to step 900. In step 906, the system
The destination address field 603 of the control command is
If not a logical address and "0xffffffff"
Returns to the process of step 900. In the processing of step 909, the node 100
Results of processing from sub processors 212 of a to 100c
Is returned and whether the sub-processor 212
If the result of these processes is returned, step 910
The processing proceeds to the processing of
Sends a response packet to the
The process returns to the step 900. Hereinafter, the parallel computer system according to the present embodiment will be described.
Of the nodes 100a to 100c in the management device
Of the sub-processor 212 of the system support mechanism 211
The order will be described. FIG. 10 shows a parallel computer system according to this embodiment.
Of the nodes 100a to 100c in the system management device.
Of the sub-processor 212 of the system support mechanism 211
It is a flowchart which shows a part of procedure. [0164] As shown in FIG.
100a to 100a in the management device of the machine system
Sub-processor 21 of system support mechanism 211 of c
In step 2, when the auxiliary power supply 201 is turned on,
00, the logical add of the nodes 100a to 100c
Address, and provided for the nodes 100a to 100c.
SRAM for storing status code to be displayed on panel
"0000" in the panel status management area in the 214
Set. Next, in the processing of step 1001, no
System control mechanisms 105b-1 of the nodes 100a-100c
05d in non-packet mode 702, and step 1
In the process of 002, the system of the nodes 100a to 100c is
Non-packet mode 70 of the system control mechanisms 105b to 105d
2 connection state to wait connect state 801
Set. System control of nodes 100a-100c
Change the mode of the mechanisms 105b to 105d to the non-packet mode 7
02 for the system control mechanisms 105b to 105d.
Wait for connection state in non-packet mode 702
The connection state 801 is set for the nodes 100a to 100a.
The node message of 100c is transmitted to the system control unit 105.
b to 105d of the RAM 222.
Control from the system control mechanism 105a of the terminal device 104
When a connection request is made by the node 804, the node 10
0a to 100c system control mechanisms 105b to 105d
Terminal message stored in the RAM 222 of the management terminal
This is for sending to the device 104. Also, by doing so, the nodes 100a-1
When the software 203 on 00c is not running
However, the information stored in the RAM 222 from the management terminal device 104
The load message can be read. Next, the system control of the management terminal device 104 is performed.
From the mechanism 105a, the system of the nodes 100a to 100c
System control commands to the system control mechanisms 105b to 105d.
Is sent, the system control command
To the sub-processors 212 of the nodes 100a to 100c,
The sub-processor 212 transmits the system control command.
Execute. In the process of step 1003, the management terminal
From the system control mechanism 105a of the
a to 100c of the system control mechanisms 105b to 105d
Through the main power supply 200 of the nodes 100a to 100c.
System control command for power control instruction to turn on or off
Check to see if a message has been sent. In the process of step 1003, the management terminal 1
As a result of checking whether there is a power control instruction from
When there is a power control instruction from the management terminal device 104
Are the nodes 100a-1
00c power supply control processing for turning on or off the main power supply 200
The power supply control is executed in step 1005.
The execution result of the control process is stored in the system of the nodes 100a to 100c.
Steps after reporting to the system control mechanisms 105b to 105d
It returns to the process of 1003. In the processing of step 1003, the management terminal 1
As a result of checking whether there is a power control instruction from
When there is no power control instruction from the management terminal device 104
Proceeds to the process of step 1006, and the management terminal 10
4 from the system control mechanism 105a to the nodes 100a to 1
00c via the system control mechanisms 105b to 105d.
Control the panels provided in the nodes 100a to 100c.
The system control command of the panel control instruction to be
Find out if it is. In the process of step 1006, the management terminal
It is checked whether there is a panel control instruction from the device 104
As a result, there is a panel control instruction from the management terminal device 104.
In this case, the process proceeds to step 1007 and the panel control is performed.
The process is executed, and in the process of step 1008, the panel
The execution result of the file control process is transmitted to the nodes 100a to 100c.
After reporting to the stem control mechanisms 105b to 105d,
The process returns to step 1003. In the processing of step 1006, the management terminal
It is checked whether there is a panel control instruction from the device 104
As a result, there is no panel control instruction from the management terminal device 104
In this case, the process proceeds to step 1009, where the management terminal
From the system control mechanism 105a of the
a to 100c of the system control mechanisms 105b to 105d
Via the main processors of the nodes 100a to 100c
System control frame for reset instruction to reset 202
Find out if a command has been sent. In the process of step 1009, the management terminal
It is checked whether there is a reset instruction from the device 104.
As a result, when there is a reset instruction from the management terminal device 104
Proceeds to the processing of step 1010,
~ 100c of the main processor 202
Then, in the process of step 1011, the reset process is executed.
System execution results of the nodes 100a to 100c
After reporting to the control mechanisms 105b to 105d, step 10
It returns to the process of 03. System control in the process of step 1012
There is a mode switching request from the mechanisms 105b to 105d.
If there is a mode switching request as a result of
If not, the process proceeds to step 1013, where the adapter control
Execute the command “SET-MODE” and control the system.
The operation modes of the mechanisms 105b to 105d are switched, and
The process returns to step 1003. As described above, the parallel computing of this embodiment
According to the management device of the machine system, the nodes 100a to 100a
0c operated by the auxiliary power supply 201, and the main processor 20
Network software and communication cable used by 2
Network software that is different from
Management using communication cables 106b to 106d
System control mechanism 105b for communicating with terminal device 104
To 105d, the system control command from the management terminal device 104.
Command to send the system control command to the auxiliary power
To be executed by the sub-processor 212 operating on the server 201
The main process of the plurality of nodes 100a to 100c
Control the processor 202, so that the main
Operation of main processor 202 and main processor 20
2. Operating system and network software
Irrespective of the operation of the software 203,
A plurality of nodes 100a to 100a constituting a parallel computer system
100c is collectively managed by the management terminal device 104.
It is possible to (Embodiment 2) The parallel computing of the present invention is described below.
The management terminal device 104
From the main power supply 200 to the plurality of nodes 100a to 100d.
And the status of the nodes 100a to 100d.
Monitor the nodes and monitor the main programs of the nodes 100a to 100d.
Implementation to manage whether the processor 202 has started operation
Mode 2 will be described. FIG. 11 shows a parallel computer system according to this embodiment.
From the management terminal device 104 to the node 1
Power supply that instructs 00a to 100d to turn on main power supply 200
It is a figure showing an example of a throwing sequence. Figure 11
100d is a node, 1101-1112 is power on
Is a sequence showing each stage. As shown in FIG. 11, the parallel computing of this embodiment
From the management terminal device 104 in the management device of the machine system
Instruct nodes 100a-100d to turn on main power supply 200
Power on sequence, the sequence 1101
And the auxiliary power supply 201 of the nodes 100a to 100d is turned on.
Have been. Auxiliary power supply 20 for nodes 100a-100d
1 is input, the sub-programs of the nodes 100a to 100d
The processor 212 executes the system
Initialize the support mechanism 211, and
Command "SET-ADDRESS"
Initialization of control mechanisms 105b to 105d and management terminal
For the device 104 to manage the nodes 100a to 100d
Of the logical address which is an address necessary for the operation. Here, for example, the logical address “0x00”
“SET-ADDRESS” to set “000001”
An example of the format of a command and its response is
It is as follows. <Command>: A0x00000001: (Recipient address field 503 is omitted.
Omitted): 0x01 LF <Response>: A0x00000001: 0x00000001: 0x01 (status information) LF In sequence 1103, the power of the management terminal device 104 is
When turned on, the bootstrap of the management terminal device 104
Bootstrap program stored in ROM 303
RAM executes the system startup process of the management terminal device 104.
Do. In sequence 1104, management terminal device 1
When the system start-up process of 04 is completed, sequence 1
At 105, the software 301 of the management terminal device 104
Sets the logical address of the management terminal device 104 to “SET-A
DDRESS ”. The management terminal device 104 and the nodes 100a to
When the logical address of 100d is set, the sequence 1
At 106, the software 301 of the management terminal device 104
Is the broadcast packet of the system control command
Is used to indicate the status of the nodes 100a to 100d.
Read the status code. The status codes correspond to nodes 100a-1
00d Panel status management area in SRAM 214
, For example, nodes 100a to 100
d when the auxiliary power supply 201 is turned on normally,
The status code is written in the panel status management area.
And the status code is
Can be read out by the
Code "0000" in the management device of the parallel computer system
Can be read. ). [0186] Here, the management terminal device 104 sets "STA
Using the “TUS-READ” command, the node 100
Broadcast is performed for a to 100d. The logical address is “0xa00000000”
Is the STATUS-REA.
"D" command and the frame
The logical address for the command is "0x00000000
One of the formats of the response from the node that is "5"
An example is as follows. <Command>: d0xa0000000: 0xffffffff: 0x4 LF <Response> d0x00000005: 0xa0000000: 0x04 0000 LF In sequence 1107, nodes 100a to 100d
The “STATUS-READ” command is received,
The status code “000” is output by the sub processor 212.
After “0” is read out,
From the terminals 100a to 100d to the management terminal device 104,
A response is returned as described above. Here, the software of the management terminal device 104
A 301 is the theory of the node that returned a normal response.
Management address and software of the management terminal device 104 in advance.
Parallels stored in 301 or in a specific file
Of the nodes 100a to 100d constituting the computer system
Checking the configuration information and returning a normal response
For a node that does not come,
Send a STATUS-READ command again
Perform a try process. In the sequence 1109, the management terminal device 1
04 software 301 is correct in sequence 1108.
Main power supply 2 of the node that returned a normal response packet
00 is input by a “P-ON” command. For example, if the logical address is “0xa0000”
000 ”from the management terminal device 104,
Is "0x00000005" for the node "P
-ON ”command and its response format
An example is as follows. <Command>: d0xa0000000: 0x00000005: 0x01LF <Response>: d0x00000005: 0xa0000000: 0x01 (completion code) LF At this time, the software 301 of the management terminal device 104
By the control, “PO” is performed at predetermined time intervals set in advance.
N "while shifting the command
Power to the entire parallel computer system.
Inrush current to the power supply equipment being supplied can be kept low.
come. In the sequence 1110, the “P-ON”
Of the nodes 100a to 100d that received the command
The processor 212 outputs a power-on signal 216 and outputs
After the source 200 is turned on, a "P-ON" command
Is returned to the management terminal device 104. Main power supply 200 of nodes 100a-100d
Is turned on, the main processor 202 boots
Bootstrap stored in the strap ROM 210
The backup program is executed and the system start-up process starts.
Begun. During the system start-up process, the boot
Status code that the lap program increments
For example, there are the following types. Where the real
In the management device of the parallel computer system of the embodiment, the status
The code is shown in hexadecimal. [0196] [Table 3]The main process of the nodes 100a to 100d
Bootstrap program of the node 202
Panel stator in SRAM 214 of 00a-100d
Write a status code to the
When the start-up process proceeds, the status code is periodically
Update. Further, the panel status management area includes:
From the sub-processor 212 of the nodes 100a to 100d
Can also be referred to, for example, nodes 100a to 100d
Display on a display device such as a panel provided in
The status code to the operator.
It is also possible. Software 301 of management terminal device 104
Is the status code of these nodes 100a to 100d.
Mode using the “STATUS-READ” command.
By periodically reading, the nodes 100a to 100a
Monitor the state of d. At sequence 1110, management terminal device 1
04 software 301 is a system control command
Monitor from transmission to reception of the response in a fixed time
As shown in FIG.
And a normal response is not returned within a certain period of time.
For the code 100d, the sequence 1111
System control command at a set time interval.
Perform a retry process to send the password. As shown in FIG. 11, the parallel computing of this embodiment
A certain number of times (in this embodiment,
As a result of the retry process of 3 times), from the node 100d
If a normal response is not returned, the sequence
At 1112, the software 30 of the management terminal device 104
1 recognizes that a failure has occurred in node 100d
I do. The software 301 of the management terminal device 104
Is, as described above, a specific system control command.
When a normal response is not received within a certain time
At a predetermined time interval, the specific system
By performing control to transmit the system control command again,
The software 203 of the nodes 100a to 100d is activated.
If not, the main program of the nodes 100a to 100d
After the system start-up process of the
Can be managed. As described above, the parallel calculation of the present embodiment
According to the management device of the machine system, the management terminal device 104
The main power supply of the plurality of nodes 100a to 100d is
Since the power source 200 is turned on or off, the parallel computer system
Of the plurality of nodes 100a to 100d configuring the system
ON / OFF of the power source 200 is collectively performed by the management terminal device 104.
Or it can be done individually. The parallel computer system of the present embodiment
According to the management device, the main power supply to the nodes 100a to 100d
An instruction to turn on the power source 200 is given at a predetermined time interval.
Power supply that supplies power to the parallel computer system
It is possible to keep the rush current of the equipment low. The parallel computer system of the present embodiment
According to the management device, in response to an instruction from the management terminal device 104,
Read the status codes of the nodes 100a to 100d.
The status of the nodes 100a to 100d.
Can be collectively managed by the management terminal device 104.
You. Further, the parallel computer system of the present embodiment
According to the management device, a specific system from the management terminal device 104 is specified.
When the normal response to the stem control command is constant
To check if it is received in the middle
The multiple nodes that make up the system are operating normally
Can be monitored by the management terminal device 104. (Embodiment 3) The parallel calculation of the present invention will be described below.
In the management device of the machine system, the management terminal device 104
Display node message 403 from node 100a
Embodiment 3 where maintenance is performed as necessary will be described.
You. FIG. 12 shows a parallel computer system according to this embodiment.
Node 10 in the management terminal device 104 in the management device
Sequence displaying node message 403 from 0a
FIG. 3 is a diagram illustrating an example of a service. In FIG.
Reference numeral 1217 denotes each step of displaying the node message 403.
4 shows a sequence. [0209] As shown in FIG.
Is added to the management terminal device 104 in the management system
Display node message 403 from node 100a
In the sequence, in the sequence 1201, the node
An auxiliary power supply 201 is turned on in advance to 100a.
System control mechanism 105b (operation mode is packet mode
701), the sub-processor 212 and the processor 22
0 is in an operable state. Node 1 to which auxiliary power supply 201 is turned on
00a, the sub-processor 212
In the "SET-ADDRESS" command, node 1
00a logical address is set. Next, the sub processor 2 of the node 100a
12 is a sequence 1203, “SET-MODE”
The operation mode of the system control mechanism 105b is changed by a command.
Non-packet mode 702 (disconnect state 800)
Set to. Sub processor 212 of node 100a
Is a sequence 1204, and further “SET-CONN
ECT ”command to change connection status to non-packet
Set to wait connect state 801 in mode 702
You. On the other hand, the management terminal 104
At 1205, the management terminal device 104 is turned on.
Starts the system startup process of the management terminal device 104
I do. System start-up process of management terminal device 104
When the management is completed, in a sequence 1206, the management terminal
104 has the same software 301 as the node 100a.
And use the “SET-ADDRESS” command
The logical address of the management terminal device 104 is set, and the
In case 1207, the operation is performed using “SET-MODE”.
Operation mode is disconnected in non-packet mode 702
State 800. In sequence 1208, management terminal device 1
04 software 301 is “STATUS-REA
D ”command causes the status code of node 100a to be
Status code "0000"
Then, in the sequence 1209, the “P-ON” command
Node 100a to the node 100a.
Instruct 200 to be inserted. The “P-ON” command from the management terminal device 104
Node 100 receiving the command and turning on the main power supply 200
a is stored in the bootstrap ROM 210
Bootstrap program to main processor 202
And the system start-up processing of the node 100a
I do. At this time, the bootstrap of the node 100a is
Message 403 output from the backup program
Is transmitted to the node 100a via the sub-processor 212.
Stored in the RAM 222 of the system control mechanism 105b
You. Software 301 of management terminal device 104
In the sequence 1210, “SET-CONNECT
T ”command, the system control of the management terminal device 104 is performed.
The connection state of the control mechanism 105a is changed to the connection state 80.
2, the system control mechanism 1 of the node 100a
Node 100a stored in the RAM 222 of the device 05b
During the system start-up process of the main processor 202
The monitoring of the node message 403 starts. Management receiving “SET-CONNECT”
The system control mechanism 105a of the terminal device 104
The system control mechanism 1 of the node 100a
05b and the call control 804.
0a connection state of the system control mechanism 105b
Changes from wait connect state 801 to connect state 80
Transition to 2. At the same time, the system control mechanism of the node 100a
105b is a sequence 1212 in which "REPORT-
CONNECT ”command to the sub-process of the node 100a.
Issued to the processor 212 and received from the management terminal 104
Inform that there was a nect request. At this time, the "REPORT-CONNECT"
Of "T" command and its response format
Is as follows. In addition, the following "REPORT-C
ONNET "command and its response
The communication address is omitted. <Command>: A :: 0xA (connect status change status) LF <Response>: A :: 0xA LF In sequence 1213, the node message of node 100a is
The message 403 is the system control mechanism 1 of the node 100a.
05b received a connect request by call control 804
At this point, the system control mechanism 105b of the node 100a
Management terminal from RAM 222 via LAN control unit 218
Transmitted to the terminal device 104. Although not specifically shown in FIG. 12,
If the system of the main processor 202 of the node 100a
During the system startup process, the main
Inconsistency was found in the file system used by
If the system startup process is interrupted,
In sequence 1214, the operator operates the management terminal 1
Of the node 100a output to the display device
Node messages during the system startup process of the processor 202
The message 403 causes the node 100a to fail.
File system such as UNIX fsck.
By issuing a maintenance command to check the stem,
The maintenance of the node 100a is also possible. The main processor of the node 100a
Although the system start-up process of 202 has been completed normally,
In a later normal operation, the main processor 2 of the node 100a
02, while using panic message
100a main processor 202
If it does, the operator will
The information stored in the RAM 222 of the stem control mechanism 105b
Is displayed on the management terminal device 104.
System down according to the contents of the node message 403.
It is also possible to consider the factors of the investment. Disconnect connection with node 100a
In the case where the
Similarly, the software 301 of the management terminal device 104 is
In sequence 1215, "SET-CONNECT"
Issue a command. Software 301 of management terminal device 104
Issues a “SET-CONNECT” command
As a result, the system control mechanism 105 of the management terminal device 104
a goes to the disconnect state 800 and the sequence 12
At 16, the system control mechanism 10 of the management terminal device 104
5a is transmitted to the system control mechanism 105b of the node 100a.
Call control 804 is performed. At sequence 1217, node 100a
The system control mechanism 105b of the management terminal device 1
04 to the call control 804 from the system control mechanism 105a.
From the other party
To the sub-processor 212 of the node 100a at the same time.
On the other hand, this is referred to as “REPORT-CONNECT”
Report in Mand. As described above, the management terminal 104
Commands to the nodes 100a to 100c.
By transmitting the information, the management terminal device 104
Power control of main power supply 200 and nodes 10a to 100c
Status monitoring of 0a to 100c becomes possible. As described above, the parallel computing of the present embodiment
According to the management apparatus of the communication system, the packet mode 701
And the non-packet mode 702
Since the processing is performed by the processor 212, the
System control commands and their responses to multiple nodes 1
00a to 100c and a specific node
Set up a connection for
The communication that continuously receives pages 403 is like a switching device.
Same management terminal without using any special hardware
This can be done with the device 104. The parallel computer system according to the present embodiment
According to the management device, the main processor 2 of the specific node
02 or the sub-processor 212 outputs the
Mode message 403, and the management terminal device 104
To read the stored node message 403
And a plurality of nodes 100 constituting a parallel computer system
a to 100c of the node processor 202
When the operation is stopped after outputting page 403
However, the node message 403 is
Can be managed collectively. (Embodiment 4) The following is a parallel calculation of the present invention.
In the management device of the machine system, the nodes 100a to 100a
0c main processor 202 system startup processing
Run and optionally its bootstrap device
Embodiment 4 in which system startup processing is performed by changing
Will be described. The management system of the parallel computer system of the present embodiment
The loading process from the bootstrap device
If the operation ends abnormally, the nodes 100a to 100c
By reading and writing the contents of the memory 204,
Change the contents and change the main program of nodes 100a-100c.
Resetting the processor 202 causes another boot
Loading process from strap device is possible
It is. The management system of the parallel computer system of this embodiment
Main memory of the nodes 100a to 100c
When reading / writing the contents of the MS-REA 204
Use "D" command and "MS-WRITE" command
I do. These system control commands are
Normal business in main processor 202 of a to 100c
When a failure occurs during parallel processing,
The contents of the main memory 204 of the nodes 100a to 100c.
It can also be used for inspection. Also, the parallel computer system of the present embodiment
In the management device, the main processor of the nodes 100a to 100c
Processor 202, sub-processor 212 and processor
In the hardware module such as the memory control mechanism 205
It is also possible to read and write the contents of the
In this case, the “REG-READ” command and the “REG-READ”
WRITE command. For example, the names of the nodes 100a to 100c
Parallel processing, which is a normal business,
If a failure occurs during execution, the nodes 100a to 100a
The error log collected by each hardware resource in 0c
Save the data in the register and use the "REG-READ"
The failure log is sent from the management terminal device 104 by a command.
By reading the information, the trouble from the management terminal device 104 is required.
The cause can be specified. Also, the parallel computer system of the present embodiment
In the management device, the "STATUS-READ" command
From the status codes of the nodes 100a to 100c.
If there is a node that has caused readout and system down
To send a “PROC-RESET” command
The main node of the node that caused the system
Resets and restarts the processor 202
Is also possible. The following describes the parallel computer system of the present embodiment.
Of the nodes 100a to 100c in the management device
Explanation of system start-up processing of processor 202
I do. FIG. 13 shows a parallel computer system according to this embodiment.
Of the nodes 100a to 100c in the system management device.
2 shows a system start-up processing procedure of the processor 202.
It is a flowchart. FIG. 14 shows a parallel computer system according to this embodiment.
Of the nodes 100a to 100c in the system management device
It is a figure showing a memory map of AM214. In FIG.
1400 is the primary bootstrap path information.
Information, 1401 is alternate bootstrap path information
It is. As shown in FIG. 14, the parallel computing of this embodiment
100a to 100a in the management device of the machine system
c SRAM 214 memory map is
First storing software 203 such as
Bootstrap device showing primary bootstrap device
Trap path information 1400 and first bootstrap
Bootstrap used when device is unavailable
Bootstrap path information indicating the boot device
Information 1401. As shown in FIG. 13, the parallel computing of this embodiment
100a to 100a in the management device of the machine system
c: System startup processing operator of main processor 202
In order, in response to a power-on instruction from the management terminal device 104,
When the main power supply 200 is turned on, the main processor 202
Stored in the bootstrap ROM 210
The bootstrap program is executed, and step 13
00, the panel status in the SRAM 214
Set the status code of the management area to "1000"
In the nodes 100a to 100c in the process of
Initial diagnosis and initialization of each hardware module
U. In the processing of step 1302, step 1
Initial diagnosis and initialization of hardware in the process of 301 are correct.
It is checked whether or not the process has been completed normally, and
Initial diagnosis and initialization of hardware are completed normally
If yes, the process proceeds to step 1303. In the processing of step 1301, the hardware
If the initial diagnosis and initialization have been abnormally completed,
In the process of step 1313, the panel provided for the node
The status code "1FFF" is displayed on the
The system startup process of the main processor 202
Abnormal termination. In the processing of step 1303, the status
Set the code to "2000" and go to step 1304
Depending on the hardware stored in the SRAM 214
Of the information, the bootstrap path information shown in FIG.
Refer to the primary bootstrap path information 1400
Bootstrap device (for example,
Operating system)
Software 203 such as a system to the main memory 204.
To load. Bootstrap path in SRAM 214
The information is stored in the bootstrap ROM 210.
The main memory when executing the bootstrap program
204 is copied to a specific area, and the system starts up.
Then, it can be referred to by the software 203. The parallel computer system of the present embodiment
On the management device, the bootstrap device
Via Ethernet (system) in addition to local files in the
System control interface) bootp protocol
(Before the Request For Connect boot
Connection to Ethernet using
Bootstrap that can be obtained from other nodes
Files are also applicable. In the processing of step 1305, the primary
Bootstrap specified in bootstrap path information 1400
From the strap device to the operating system, etc.
Software 203 to the main memory 204
Check whether the loading process has completed normally.
If the above-mentioned loading process is successful, step 1306
If the process fails, the process proceeds to step 1314.
No. In the process of step 1306, the status
Set the code to "3000"
The activated software 203 is started, and step 13 is executed.
07, the status code is set to "A000"
Various system parameters are set in the process of step 1308.
File system in step 1309.
Is initialized, and in the process of step 1310, the TCP
Initializes the network such as / IP. The management system of the parallel computer system of this embodiment
The main processors of the nodes 100a to 100c
Operating system and net operating on 202
Use the functions of software 203 such as work software
The system operation support interface to use at this point
It can be used. In the process of step 1311, the application
The application software is started, and the
Set the status code to "F000" by processing
The system start-up process of the processor 202 ends. On the other hand, in the processing of step 1314, the SR
Alternate bootstrap path information in AM214
1401 and the alternate bootstrap pass
Bootstrap device specified by information 1401
(Especially in the management device of the parallel computer system of the present embodiment,
Although not disclosed, DAT (Digital Audio
o Operating from input / output devices such as Tape)
Software 203 such as a
Load 4 In the processing of step 1315, the alternate
Bootstrap path information 1401
Successful loading from the smart strap device
Is checked, and if successful, the process proceeds to step 1306.
move on. In the processing of step 1315, the alternate
Bootstrap path information 1401
If loading from a smart strap device is not successful
In step 1316, the operator
Load processing is performed by specifying the strap device. In the processing of step 1317, step 1
The operator's bootstrap device in the process of 316
Check whether the load processing by specifying the
If the operation has been completed normally,
Proceeds to the processing, and if not completed normally, step 1
In the process of 318, the status code is set to “2FFF”
Then, the system startup process of the main processor 202 is performed.
Abnormal termination. Nodes 100a to 100a-1 performed as described above
00c system startup process of the main processor 202
If the process ends abnormally,
System start-up processing with changed strap device
I do. The software 301 of the management terminal device 104
Uses the “MS-READ” command to
Via the control interface, the nodes 100a to 100
c, the bootstrap path information in the main memory 204 is
Referring to the specific area stored, the main processor
Boots that failed to start the system
Check the wrap device. Next, the software of the management terminal device 104
301 uses the “MS-WRITE” command,
Via the stem control interface, nodes 100a-
Bootstrap path of main memory 204 of 100c
In the specific area where the information is stored, the main processor
Boots that failed to start the system
Write the name of the bootstrap device other than the wrap device.
Get in. The software 301 of the management terminal device 104
Is the main menu of the nodes 100a to 100c as described above.
Bootstrap path information of the specific area of the memory 204
After rewriting the “PROC-RESET” command
Used and the main processor of the nodes 100a to 100c
202 is reset and the system of the main processor 202 is reset.
By restarting the system startup process, the bootstrap destination
Can be performed to change the system. Rewriting bootstrap path information
The main program of the nodes 100a to 100c
The system startup process of the processor 202 has been completed normally.
In this case, the following method is also possible. That is, the S of the nodes 100a to 100c
The bootstrap path information of the RAM 214 is
Rewritten from software 203 of 00a to 100c
Since it is possible, the software 3 of the management terminal device 104
01 via the system operation support interface
For the software 203 of the code 100a to 100c,
Specify that bootstrap path information be rewritten, and
Installed software 203 is
Rewrite pass information. Software of the nodes 100a to 100c
The system 203 updates the bootstrap path information.
Of the management terminal device 104 via the system control interface.
Notify the software 301, and
Software 301 via the system control interface
Use the “PROC-RESET” command at
Processor 202 of each of the nodes 100a to 100c
Reset, immediately updated bootstrap
Load processing from the path is performed. As described above, the parallel computing of the present embodiment
According to the management device of the machine system, the management terminal device 104
The main memo of nodes 100a to 100c
To refer to or update the contents of registers 204 or registers
And a plurality of nodes 100 constituting a parallel computer system
Contents of the main memory 204 at the time of occurrence of a failure from a to 100c
Can be collectively managed by the management terminal device 104.
is there. The parallel computer system according to the present embodiment
According to the management device, in response to an instruction from the management terminal device 104,
Processor 202 of each of the nodes 100a to 100c
To configure a parallel computer system
Main processor 2 of a plurality of nodes 100a to 100c
02 is collectively performed from the management terminal device 104.
It is possible. Further, the parallel computer system of this embodiment
According to the management device, the management terminal device 104
Use different interfaces between 0a and 100c
Is possible, and by an instruction from the management terminal device 104
The boot of the main memory 204 of the nodes 100a to 100c
Change the trap strap path information and change the main processor 20
Since the reset of 2 is performed, the parallel computer system is configured.
Bootstrap of a plurality of nodes 100a to 100c
When a failure occurs in the wrap device, the management terminal
According to the instruction from 104, the bootstrap device is
Main processor of nodes 100a-100c
It is possible to perform the system
You. (Embodiment 5) The parallel computing of the present invention is described below.
Multiple management terminal devices in the management system
The schematic configuration of Embodiment 5 in which the reliability is improved by using
Will be explained. FIG. 15 shows a parallel computer system according to the present invention.
An embodiment in which the management terminal is duplicated in the management device
5 is a diagram showing a schematic configuration of No. 5. FIG. In FIG.
e is a system control mechanism, 106e is a communication cable, 10
8e is a LAN control mechanism, 109e is a communication cable, 11
Reference numeral 1 denotes a management terminal device. As shown in FIG. 15, the parallel computing of this embodiment
The management device of the communication system communicates with the management terminal device 111.
A cable 106e and a communication cable 109e,
The management terminal device 111 includes a system control mechanism 105e,
A management terminal device having a LAN control mechanism 108e.
111 to the communication cable 10
6e to the network concentrator 107
The LAN control mechanism 108e of the
Connected to the network concentrator 110 via the
Has continued. As described above, the parallel computer system of this embodiment
The system management device includes a plurality of management terminal devices 104 and 1
11 so that one management terminal device
Operation of the parallel computer system by another management terminal.
Management can continue, but multiple management terminals
Operation and management of a parallel computer system using devices simultaneously
When this is done, the system control commands transmitted by the
Commands and adapter control commands conflict with each other
When using multiple management terminal devices
Processing to prevent contention between management terminal devices
You. The following describes the parallel computer system of the present embodiment.
If multiple management terminal devices are used in the management device of
Procedure to prevent conflicts in operation of the management terminal
Will be described. FIG. 16 shows a parallel computer system according to this embodiment.
Conflict between the operations of multiple management terminal devices in the management device
5 is a flowchart showing a processing procedure for preventing the problem. The management system of the parallel computer system of this embodiment
If the management terminal device is duplicated,
For example, the management terminal device 1
04 and the management terminal device 111
By setting in advance to a storage area that can be referenced from 1,
It is possible to prevent contention between redundant management terminal devices.
Become. As shown in FIG. 16, the parallel computing of this embodiment
Management terminal device in the management system
In the processing procedure when the
Thus, both the management terminal device 104 and the management terminal device 111
Software 301, the management terminal device is duplicated
Refer to the duplicated bit to indicate that the bit is up
If the management terminal is duplicated,
recognize. In the processing of step 1601, the network
(For example, via a system operation support interface)
Then, the IP address of the partner management terminal device is obtained. In the processing of step 1602, the main management
Backup the terminal device and the main management terminal device
In order to determine the sub-management terminal device to be
The younger one in the dress is the main management terminal, and the other one is
A sub-management terminal device. At this time, only the main management terminal operates.
And a failure has occurred in the main management terminal.
Operation method to switch to sub-management terminal immediately
And the main management terminal device and the sub management terminal device
It is possible to do the operation method to operate, but the latter
In the case of, control the nodes 100a to 100c from both sides
System control commands and adapter control commands
Because the content may conflict,
Transmittable system control commands and adapter control frames
Some restrictions. For example, in the process of step 1603,
Determines whether the management terminal is the main management terminal
If it is not the main management terminal,
The system control command (“P-
ON, P-OFF, etc.)
Commands (such as “SET-CONNECT”)
By controlling the nodes 100a to 100c.
Content of the system control command and the adapter control command
Even so, maintain consistency as a parallel computer system
Is possible. As described above, the parallel computing of this embodiment
According to the management device of the machine system, a plurality of management terminal devices
Failure occurred in one management terminal device
Even if the operation management of the parallel computer system is continued,
It is possible to improve the reliability of computer systems.
You. The parallel computer system according to the present embodiment
According to the management device, the main management terminal is connected to a plurality of management terminal devices.
Since the terminal device and sub-management terminal device are set,
If the management system is managed by multiple management terminal devices,
It is possible to prevent conflicts in operation of multiple management terminal devices.
Noh. (Embodiment 6) The parallel calculation of the present invention will be described below.
In the management device of the machine system, the management terminal device 104
A power-on logic that operates on the auxiliary power supply is added, and the management terminal
Parallel by turning on the main power of the device 104 from a remote location
Embodiment 6 for turning on the main power supply of the computer system
Will be explained. FIG. 17 shows a parallel computer system according to this embodiment.
With an auxiliary power supply to the management terminal device 104 in the system management device.
In the management terminal when the power-on logic that operates is added
FIG. 2 is a diagram showing a schematic configuration of hardware of FIG. FIG.
1700 is an auxiliary power supply, 1701 is a power-on logic,
1702 is a power control signal, 1703 is a main power supply, 1704
Is a terminal device, and 1705 is a network. As shown in FIG. 17, the parallel computing of this embodiment
The management terminal device 104 in the management device of the machine system
Auxiliary power supply 1700, power-on logic 1701, main power supply
1703, and supplies power from the auxiliary power supply 1700.
The received power-on logic 1701 is
2 to the main power supply 1703 and the network
Connected to another terminal device 1704 via the
You. As shown in FIG. 17, the parallel computing of this embodiment
The management terminal device 104 in the management device of the machine system
A power-on logic 1701 operated by the auxiliary power supply 1700 is provided.
And power-up logic 1701 is specifically shown here.
Network control unit, power control unit, and
It is composed of microprocessors, etc.
When a power supply control instruction is received via 1705, the main power supply 17
03 is provided. With this power-on logic 1701, for example,
Remote operation of the management terminal device 104 as follows
Becomes possible. The management system of the parallel computer system of this embodiment
At another end connected by a network 1705
The terminal device 1704 uses, for example, the telnet protocol.
To log in to the power-on logic 1701. (this
At this time, the auxiliary power supply 1700 is
It is in the state that has been done. Next, the terminal device 1704 is
A power-on command is issued to the power-on logic 1701.
The power on logic 1701 receives the power on command.
Recognizes that there has been an instruction to turn on the power from outside,
Output the power source control signal 1702, and
The power supply 1703 is turned on. When the main power supply 1703 of the management terminal device 104 is
When inserted, it is stored in the bootstrap ROM 303.
Bootstrap program is the management terminal device 10
4 performs the system start-up processing, and
Start FIG. 18 shows a parallel computer system according to this embodiment.
Shell program of the management terminal device 104 in the system management device
It is a figure showing an example of a gram. Where the shell program
Is a universal operating system, UNI
Multiple command names or program names executed in X
Refers to a program that performs a series of procedures described.
In step 18, the UNIX command name or program
Instead of a program name, simply describe the function of the command.
I have. [0287] As shown in FIG.
System of the management terminal device 104 of the management device of the machine system
The shell program executed during the startup process
To turn on the main power supply 200 of the nodes 100a to 100c.
Describe the system control commands to be displayed, and
When the main power supply 1703 of the device 104 is turned on,
Make sure that the shell program is executed. By doing so, the management terminal device 104
Of the nodes 100a to 100c
The source 200 is turned on, and the mains of the nodes 100a to 100c are
The system startup process of the processor 202 is automatically executed.
It is possible to As described above, the parallel computation of this embodiment
According to the management system of the machine, access from remote locations
To turn on the main power supply 1703 of the management terminal device 104,
Further, the nodes 100a to 100c are activated (main power supply 200
Operation of the parallel computer system
Can be performed from a remote location. The parallel computing of the present embodiment described above
Although not shown in the illustration,
The present invention is also applicable to such a system. (1) General-purpose operating system for each node
Dedicated to perform certain functions without a system
General-purpose operating system
System that does not use the network function of the
Operation management is performed using only the system control interface. The management apparatus of the parallel computer system of the present invention
According to the system control interface, the operation management
Auxiliary power supply and network independent of the target processor
Work function and a general-purpose operating system.
Not use network functions such as TCP / IP of the system
So, with the general-purpose operating system
It does not need to be applied to dedicated parallel computer systems.
It is possible. (2) Function of operating each node with an auxiliary power supply
Or has limited functions that operate with auxiliary power supply
In a parallel computer system like this,
Only the system operation support interface that is not used
Use to manage. In this case, an auxiliary power supply such as turning on the main power supply
System control interface except for functions that require
Functions by the system operation support interface
By doing so, the management system of the parallel computer system of the present invention
It is possible to apply the arrangement. As described above, the present invention is based on the above-described embodiment.
Although described physically, the present invention is limited to the above-described embodiment.
Species within the scope of the gist.
Of course, it can be changed variously. [0296] The invention disclosed in the present application is representative.
To briefly explain the effects that can be achieved by the traditional,
It is as described. [0297] [0298] [0299] [0300] [0301] [0302] [0303] [0304] [0305] [0306]FarManagement terminal device by access from remote place
Turn on the main power of all nodes, and then turn on the main power of multiple nodes.
Operation management of the parallel computer system from a remote location.
It is possible to do.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing a schematic configuration of a first embodiment for implementing a management device for a parallel computer system according to the present invention. FIG. 2 is a diagram illustrating a schematic configuration of a node configuring the parallel computer system in the management device of the parallel computer system according to the first embodiment. FIG. 3 is a diagram illustrating a schematic configuration of a management terminal device in the management device of the parallel computer system according to the first embodiment. FIG. 4 is a diagram illustrating an example of a communication sequence between a management terminal device and each node in the management device of the parallel computer system according to the first embodiment. FIG. 5 is a diagram illustrating a packet format of an adapter control command and its response in the management device of the parallel computer system according to the first embodiment. FIG. 6 is a diagram illustrating a packet format of a system control command and its response in the management device of the parallel computer system according to the first embodiment. FIG. 7 is a diagram illustrating a mode transition of a system control mechanism in the management device of the parallel computer system according to the first embodiment. FIG. 8 is a diagram illustrating a transition of a connection state in a non-packet mode of a system control mechanism in the management device of the parallel computer system according to the first embodiment. FIG. 9 is a flowchart illustrating a part of a processing procedure of a processor of a system control mechanism in the management device for the parallel computer system according to the first embodiment; FIG. 10 is a flowchart illustrating a part of a processing procedure of a sub-processor of the system support mechanism in the management device of the parallel computer system according to the first embodiment. FIG. 11 is a diagram showing an example of a power-on sequence for instructing each node to turn on main power from a management terminal device in the management device of the parallel computer system according to the second embodiment. FIG. 12 is a diagram illustrating an example of a sequence of displaying a node message of each node on a management terminal device in the management device of the parallel computer system according to the third embodiment. FIG. 13 is a flowchart illustrating a system startup processing procedure of a main processor of a node in the management device for the parallel computer system according to the fourth embodiment. FIG. 14 is a diagram showing a memory map in an SRAM in a node in the management device of the parallel computer system according to the fourth embodiment. FIG. 15 is a diagram showing a schematic configuration of a fifth embodiment in which a management terminal device is duplicated in the management device of the parallel computer system of the present invention. FIG. 16 is a flowchart illustrating a processing procedure for preventing contention between operations of a plurality of management terminal devices in the management device of the parallel computer system according to the fifth embodiment. FIG. 17 is a diagram illustrating a schematic configuration of hardware in the management terminal device when a power-on logic operating on an auxiliary power supply is added to the management terminal device in the management device of the parallel computer system according to the sixth embodiment. FIG. 18 shows a shell program of a management terminal device in the management device of the parallel computer system according to the sixth embodiment. [Description of References] 100a to 100c: nodes, 101: communication processing device,
102a to 102c: communication interface mechanism, 103
a to 103c: communication cable, 104: management terminal device,
105a to 105d: System control mechanism, 106a to 1
06d: communication cable, 107: network concentrator, 108a to 108d: LAN control mechanism, 109a to
109d communication cable, 110 network concentrator, 200 main power supply, 201 auxiliary power supply, 202 main processor, 203 software, 204 main memory, 205 processor memory control mechanism, 206
System bus, 207: system disk, 208: I
/ O control mechanism, 209 ... RS-232C control mechanism, 21
0: bootstrap ROM, 211: system support mechanism, 212: subprocessor, 213: ROM, 2
14 SRAM, 215 local bus, 216 power on / off signal, 217 processor reset signal, 2
18 LAN control unit, 219 RS-232C control unit,
220: Processor, 221: ROM, 222: RA
M, 223: data interface, 224: control interface, 300: processor, 301: software, 302: main memory, 303: bootstrap ROM, 304: processor memory control mechanism, 305:
System bus, 306 I / O control mechanism, 307 System disk, 308 and 309 RS-232C control mechanism, 310 Graphics control mechanism, 311 LA
N control unit, 312 ... RS-232C control unit, 313 ... processor, 314 ... ROM, 315 ... RAM, 316 ...
Control interface, 317 ... data interface,
401 ... Adapter control command and its response, 4
02: system control command and its response, 40
3: node message, 501: type field, 50
2 ... source address field, 503 ... destination address field, 504 ... information field, 505 ... identifier, 601 ... type field, 602 ... source address field, 603 ... destination address field, 6
04: information field, 605: identifier, 701: packet mode, 702: non-packet mode, 703:
"SET-MODE" command, 800: disconnect state, 801: wait connect state, 802: connect state, 803: "SET-CONNECT" command, 804: call control between system control mechanisms, 1400 ...
Primary bootstrap path information, 1401 alternate bootstrap path information, 105e system control mechanism, 106e communication cable, 108e LA
N control mechanism, 109e communication cable, 111 management terminal device, 1700 auxiliary power supply, 1701 power-on logic, 1702 power control signal, 1703 main power supply, 17
04: terminal device, 1705: network.

──────────────────────────────────────────────────続 き Continuation of front page (56) References JP-A-7-129282 (JP, A) JP-A-5-324133 (JP, A) JP-A-3-12772 (JP, A) JP-A 64-64 25270 (JP, A) Japanese Utility Model Showa 60-5541 (JP, U) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06F 15/16-15/177 G06F 1/26-1/32

Claims (1)

(57) [Claim 1] In a management apparatus for a parallel computer system having a management terminal device for managing a parallel computer system to which a plurality of nodes are connected, wherein the plurality of nodes are A main processor that is operated by a main power supply of the node and executes parallel processing; a sub-processor that is operated by an auxiliary power supply of the node and executes a system control command for managing the main processor; and the management terminal device that is operated by the auxiliary power supply A system control mechanism that communicates with the management terminal apparatus and generates control data based on a system control command issued from the management terminal apparatus; and LAN control mechanism for managing application software of the node under control of a terminal device The management terminal device is operated by an auxiliary power supply and a system control mechanism that communicates with a plurality of system control mechanisms of the plurality of nodes, and turns on the main power supply of the management terminal device when a specific signal is input. a power-on logic that, when the main power is turned on by the power on logic, to the sub processor of the plurality of nodes transmits a system control command for collectively or individually after the main power is turned on via the system controls Means for connecting a plurality of system control mechanisms of the plurality of nodes and a system control mechanism of the management terminal device, wherein the management terminal issues a system control command for managing a plurality of main processors of the plurality of nodes. characterized in that the device comprises a system control interface you sent to a plurality of system control mechanism of the plurality of nodes Management unit of column computer system.