JPH07141077A - Hot-line inserting/pulling out device - Google Patents

Abstract

PURPOSE:To insert/pull out a sub unit without disconnecting a device power source and the power source of each slot by providing a diagnosis control circuit which instructs self diagnosis to the diagnostic circuit of a sub unit for exchange via a spare slot. CONSTITUTION:When the package 17 of the spare slot is inserted to a memory part 3, a jig 13 for hot-line insertion/pulling-out is mounted on the package of the spare slot, and a memory control circuit 18 is set in a reset state, and also, power source application sequence is set equal to a voltage to be supplied to the package 17 of the spare slot, then, a hot line is inserted/pulled out. At this time, the package 17 of the spare slot confirms an operation by performing the self diagnosis by using a self-diagnostic circuit in the package 17, and when it is judged that the package 17 of the spare slot can be operated normally as a result of confirmation, exchange procedure is executed by performing the alternate operation of the package 17. Therefore, the package 17 can be inserted/pulled out without disconnecting the device power source.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to maintenance during operation of a device, and when a sub-unit in the device is desired to be inserted / removed, the power of the device is not cut off, and after compensating for the data without the sub-unit. The present invention relates to a hot-line insertion / extraction device that can be inserted and removed.

[0002]

2. Description of the Related Art Regarding hot insertion / removal, various publicly known examples of connection sequence of power sources in a connector portion between a conventionally inserted / removed subunit and an apparatus main body and maintenance of a reset state of the subunit have been proposed. For example, JP-A-4-
The techniques disclosed in 181411 and JP-A-4-157514 correspond to this.

In addition, the power supply to the slot is individualized,
It has also been proposed to cut or supply only a designated slot by control from inside and outside, and replace a failed package during system operation. For example, JP-A-4-333118
The technology described in the issue corresponds to this.

[0004]

However, in the former technique, the structure of the connector must be special, and in the latter technique, the power supply can be cut off for each subunit or each slot. However, there is no change in the fact that the power must be turned off for that slot. Further, when the sub unit is exchanged by using these techniques, the data stored in the sub unit to be exchanged is not guaranteed, so that it is necessary to restart the device after the sub unit is exchanged.

An object of the present invention is to insert / remove a subunit without disconnecting the power source of the device and the power supply for each slot, and by guaranteeing the data of the subunit, it is not necessary to restart the device. It is to provide a hot-swap method that has been adopted.

[0006]

The above-mentioned object is constituted by a platter having a spare slot preliminarily wired on the platter (subunit insertion portion) so that the subunit required for normal operation and the replacement subunit can be inserted. It can be realized by using a hot-plugging / disconnecting jig as a means for performing the sub-unit insertion / removal while the power supply of the device is being performed.

Further, as a means for improving the reliability of the device when the subunit is inserted or removed, the controller controls the spare platter in the same manner as the platter to control the normal operation and the address / subunit selection. A control circuit and a data buffer circuit necessary for guaranteeing data, a circuit for freezing and stocking error information, a data transfer / subunit switching control circuit for instructing data transfer and disconnection of the sub unit and replacement sub unit, This can be realized by using a subunit containing a self-diagnosis circuit as a means for preventing data double guarantee and guarantee of data that is possible by configuring a data comparison circuit capable of comparing data and error information.

[0008]

Since the sub-unit is inserted / removed during power supply to the device, when the sub-unit for replacement is inserted as a substitute for the sub-unit, the power-on sequence of the sub-unit is guaranteed and the voltage supplied to the platter and the sub-unit for replacement is the same. In addition, it is necessary to reset the data signal and control signal in the replacement subunit (fixed to low level). Therefore, these can be realized by attaching the hot-swap jigs and jigs to the sub-unit.
It is possible to prevent electrical disturbance when the subunit is inserted or removed.

Further, in order to switch to the replacement sub-unit as an alternative to the sub-unit, the self-diagnosis circuit in the sub-unit receives the self-diagnosis start signal from the diagnosis control circuit of the device, and thereby the self-diagnosis built in the sub-unit is performed. By activating the circuit and performing self-diagnosis, and confirming the normal operation of the replacement subunit, double failure of the device is prevented, and data is guaranteed for the replacement subunit and replacement subunit. The alternative operation to the sub-unit is performed. As a data guarantee for both subunits, the subunit selection signal was stored in the data buffer circuit by selecting the subunit, reading all the data from the subunit to the data buffer circuit, and then selecting the replacement subunit. Data is written and data comparison is performed by the data comparison circuit for reading data from both subunits. The same data line on the platter is used for writing and reading this data. Therefore, the sub-unit selection signal reads the respective data by alternately selecting the sub-unit and the replacement sub-unit, and compares the data of both sub-units with the data comparison circuit via the ECC circuit in the device, It is determined that the contents of the subunit have been copied to the replacement subunit. Further, the error data of the subunit is corrected through the ECC circuit in the apparatus, and the defective address and subunit information of the corrected error data are stored in the error address freeze stack. The corrected data syndrome is stored in the error data stack, and the contents of the error address freeze stack and error data freeze stack of the relevant subunit that occurred while the device was operating are the errors that occurred during this data guarantee. If the contents are the same, it can be verified that the failure location in the subunit is fixed. Therefore, since the replacement subunit does not adversely affect the device even if the replacement subunit performs the replacement operation of the subunit, the replacement subunit can perform the replacement operation even if the subunit is disconnected according to the instruction of the subunit switching control circuit. Since the device can be continuously operated and the device can be maintained in the operating state, it contributes to the improvement of efficiency and reliability of the device.

[0010]

【Example】

(Embodiment 1) Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

In the present embodiment, the subunit will be described as a package, and the device will be described as a system device. Furthermore, it will be explained including the case where measures for improving reliability are taken. FIG. 1 is a schematic diagram showing an embodiment of a system device according to the present invention.

The system unit 1 includes packages 17A to 17F.
A memory unit 3 composed of slots A to F (F is a spare slot) for accommodating data, a data transfer / package switching control circuit 4 for instructing data transfer from a defective package to a replacement package and a package switching instruction, Diagnosis control circuit 5 for activating self-diagnosis control circuit and normal operation of package, memory control circuit 6 for generating timing for driving package, address / package selection signal control circuit 7 for address supply and package selection control, data A data buffer circuit 8 for storing temporary data at the time of comparison, a data comparison circuit 9 for comparing data and defect information, a defective address and package information generated during normal operation, and a defective address and package information generated during data guarantee. For error address freeze consisting of surface configuration Tack 10, as means for correcting the correctable error data, using the method of adding the check bits included (address ECC) ECC
A signal connecting the control unit 2 and the memory unit 3 and the control unit 2 which is composed of the circuit 11, the syndrome generated during the normal operation and the error data freeze stack 12 having the two-sided configuration of the syndrome generated during the data guarantee. As timing, address, package selection signal, data line and self-diagnosis start signal 20, self-diagnosis end signal 21,
It is composed of the self-diagnosis result signal 22.

FIG. 2 is a jig for hot-plugging and unplugging 1 for exchanging a package during power supply of the system unit 1 of FIG.
3 is a schematic view of FIG.

Reset signal / power supply sequence generation circuit 1
Reference numeral 6 denotes a jig / tool starting terminal 14 in a power supply unit (in advance, a jig / tool power supply terminal capable of equalizing the voltages supplied to the memory unit 3 (platter) and the package 17F is provided on the platter). By connecting, supply of operating power of the generation circuit and activation of the jig, and a suppression signal 160 for suppressing an in-package signal to the memory / control circuit 18, a reset for resetting the memory / control circuit when the package is inserted / removed The generation of the signal 161 and the power supply to the package 17 are controlled by the generation circuit via the power supply unit to achieve an ideal power-on sequence sequence (power supply ground, lower power supply voltage V1 ... V4 in order). The power supply control signal 162 for instructing to turn on is supplied.

The connector between the package 17F and the memory section 3 uses a common general component in which all leads have the same length, unlike the known example shown above.

When the package 17F is inserted into the memory unit 3, the hot-swap jig 13 is attached to the package 17F so that the memory / control circuit 18 is reset, and the power-on sequence and the platter and the package 17F. By supplying the same voltage to
Can be hot-plugged and unplugged. When removing the jig / tool after the package 17F is inserted, the jig / tool starting terminal 14 is removed from the power supply unit, and the jig / tool is removed after the reset state of the in-package signal is released. Also, package 1
When 7A is removed from the memory unit 3 (replacement of a defective package), the jig is attached to the package 17A, the memory / control circuit 18 is reset, and the 7A is removed. After the package 17A is removed from the memory unit 3, the jig / tool starting terminal 14 and the jig / tool may be removed in this order. Therefore, by using the hot-swap jig 13 for inserting a replacement package as a substitute for a defective package during power supply to the system device 1, guaranteeing the power-on sequence at the time of package insertion and removal and supplying to the platter and the package 17F. The same voltage is applied, and the signal inside the package of the replacement package to be inserted can be reset, so that electrical disturbance can be prevented.

FIG. 3 is a schematic diagram of a self-diagnosis circuit incorporated in the replacement package. The self-diagnosis circuit 19 built in the package 17 includes a self-diagnosis start signal 20, a self-diagnosis end signal 21, a self-diagnosis result signal 22, and a memory / control circuit 1.
8, a diagnostic signal group 23 for self-diagnosis and a memory /
The control circuit 18 includes a diagnosis pass / fail judgment circuit 24 for verifying an expected value and an output value at the time of diagnosis.

The self-diagnosis circuit 19 is activated by the self-diagnosis start signal 20 from the diagnosis control circuit 5 in the system unit 1 and sends the diagnosis signal group 23 to the memory / control circuit 18 to start the self-diagnosis. To do. After completion of the self-diagnosis, the diagnosis pass / fail judgment circuit 24 compares the expected value sent from the self-diagnosis circuit 19 with the value output from the memory / control circuit 18 to make a pass / fail judgment, and reports the result to the self-diagnosis circuit 19. . The self-diagnosis circuit 19 transmits a self-diagnosis end signal 21 and a self-diagnosis result signal 22 to the diagnosis control circuit 5 in the system unit 1.

From the self-diagnosis result, it is possible to confirm the operation of the replacement package which replaces the defective package.
It is possible to prevent a failure (double failure) caused by inserting the defective replacement package into the system device 1.

FIG. 4 and FIG. 5 are flow charts of the procedure of the package replacement at the time of failure occurrence and the data guarantee at the package replacement. A series of operations based on a procedure for exchanging a defective package when a failure occurs in a package in the memory unit 3 and a data guarantee at the time of exchanging the package will be described below with reference to FIGS.

It is assumed that a failure (for example, a collectable error) has occurred in the data read operation from the memory section 3 of the package 17A. At this time, the defective address and package information are stored in the error address freeze stack 10, and the syndrome is stored in the error data freeze stack 12. If the system unit 1 is in the instruction unexecuted state, the system unit 1 is temporarily stopped and the procedure for inserting the package 17F into the spare slot F is performed.
When the system unit 1 is in the instruction execution state, the procedure for inserting the package 17F is performed after the system unit 1 is temporarily stopped by waiting until the instruction execution is completed. When the error occurs, the package 17F is operated in place of the package 17A. Therefore, a spare slot F wired in advance on the platter is provided so that the same operation as the package 17A can be performed, and the package 17F is inserted.
However, during the power supply of the system unit 1, the package 17
To insert / remove F, use the hot wire insertion / extraction jig 13
Prevents electrical disturbance when inserting and removing the package.

Next, the package 17F is the package 17
If self-diagnosis is performed using the self-diagnosis circuit 19 therein to confirm the operation, and as a result, it is determined that the package 17F can operate normally, the substitute operation of the package 17A will adversely affect the system unit 1. Since there is no package replacement procedure, continue.

If it is determined that the package 17F is abnormal, the substitute operation of the package 17A is disabled, and the process returns to the process of replacing the spare package 17F.

Next, the package 17A and the package 17
As for the data guarantee of F, the system unit 1 is in the instruction unexecuted state (J
OB stop state), the package selection signal selects the package 17A, sequentially reads all the data, stores the data in the data buffer circuit 8, and then the package selection signal selects the package 17F, and stores the data stored in the data buffer circuit 8. Are sequentially written in the package 17F. The same data line on the platter is used for reading and writing this data. Therefore, the package selection signal reads the data of both by alternately selecting the package 17A and the package 17F, and the data comparing circuit 9 is read via the ECC circuit 11 and the data buffer circuit 8 in the system device 1.
Then, the two data are compared to determine that the contents of the package 17A have been copied to the package 17F. Also,
The error data in the package 17A is corrected through the ECC circuit 11 in the system unit 1, the defective address and the package information are stored in the error address freeze stack 10, and the syndrome is in the error data freeze stack 12. To store. The data comparison circuit 9 compares the contents of the error generated at this time with the data stored in the error address freeze stack 10 and the error data freeze stack 12 at the time of the previous failure, and if they are the same, a failure occurs. The location is fixed (eg, solid error). Therefore, the package 17F has no problem even if the package 17A substitutes the package 17A, and the package replacement procedure is continued.

However, the contents of the error generated during the data guarantee this time are the stack 10 for error address freeze and the stack 1 for error data freeze when the previous fault occurred.
If the contents stored in 2 are not the same, package 17
Since the failure point of A is intermittent (for example, soft error), the package 17A is not replaced.

Next, the data transfer / package selection switching control circuit 4 in the system unit 1 disconnects the package 17A, and the package selection signal selects the package 17F instead of the package 17A to enable the alternative operation. As a result, even if the package 17A is removed from the memory unit 3 by using the jig, the system device 1 can be continuously operated.

After repairing the defective package 17A, the defective package 17A is again inserted into the memory section 3 using the jig and the package 17A
From the package 1 to the package 17F
7F is switched to operate the repaired package 17A.

Further, by inserting the package 17F into the spare slot F in advance before supplying the power source of the system unit 1, the data transfer / package selection switching control circuit 4 is automatically operated by a software means when a failure occurs. It is also possible to selectively switch the operations of the package 17A and the package 17F and disconnect the package 17A.

The procedure of the first embodiment is applicable not only to the failure occurrence described in the first embodiment with respect to the insertion / removal of the package 17 but also to the preventive maintenance at the time of regular inspection and the insertion / removal of the package 17 at the time of upgrading the system unit 1. By inserting and removing the package 17 in accordance with the above, the package 17 can be inserted and removed without turning off the power source of the system device 1.

(Embodiment 2) FIG. 6 is a schematic view of an embodiment and a hot-swap jig according to the present invention.

In this embodiment, the module 25 is used as the subunit, and the insertion and removal of the module 25 in the case of a water cooling device will be described.

In this embodiment, in addition to the structure of the first embodiment, in order to supply cooling water when the module 25 is inserted and removed,
In order to provide a sub-water supply route in the water supply route of the hot-swap jig 13 and the module 25 and the system device 1, the sub-water supply device 26 and the sub-water supply hose 28 are provided in the hot-swap jig 13 to the module 25. The dual hose jack 29 is provided by providing the sub hose jack 30, and the bifurcated hose jack 31 is provided for the system device 1.
By providing (the hose jack 29 is bifurcated and either one of the cocks 32 can be selected), the module 25 can be inserted / removed without cutting off the power supply in the water cooling device as in the air cooling device.

If you want to remove the module 25A,
A case in which the module 25B performs an alternative operation while being removed will be described.

When it is desired to remove the module 25A, the sub-water supply hose 28 of the hot-swap jig 13 is attached to the sub-hose jack 30 of the module 25A while the system unit 1 is being supplied with power, and the hot-swap jig 13 is attached. The cooling water of the module 25A is also supplied from the sub water supply device 26 of the system device 1, and then the bifurcated hose jack 31 of the system device 1 is supplied.
The sub water supply route hose 33 is attached to the hose jack 29 on the side of the sub water supply route, and the cock 32 is turned to cause the cooling water to flow to the sub water supply route to shut off the cooling water from the system device 1 to the module 25A. In the meantime, according to the procedure of Example 1, the module 25A is removed.

When the module 25B is inserted into the slot, the sub water supply hose 28 of the hot-swap jig 13 is connected to the sub hose jack 3 of the module 25A.
0, and while the cooling water for the module 25A is being supplied from the sub water supply device 26 of the hot-swap jig 13, the replacement module 25B is inserted into the slot according to the procedure of the first embodiment. Water supply hose 27 of system unit 1
Is attached to the module 25B, and the hot-swap tool 1
This can be done by cutting off the sub-water supply route from 3 and then removing the hot-swap jig 13 according to the procedure of the first embodiment.

[0036]

According to the present invention, when a sub-unit in the device is inserted or removed, the sub-unit can be replaced with a sub-unit that operates as a substitute for the sub-unit while the power is being supplied to the device, and the device can be recovered without restarting. Therefore, replacement of defective parts and maintenance time can be shortened.

Further, when the sub-unit is inserted or removed, the self-diagnosis circuit built in the sub-unit is used to confirm the operation of the replacement sub-unit, so that the failure of the sub-unit which operates as an alternative to the sub-unit can be found. Double failure of the device can be prevented.

[Brief description of drawings]

FIG. 1 is a schematic diagram of a system device showing an embodiment of the present invention.

FIG. 2 is a schematic view of a jig for hot-plugging and unplugging a package for exchanging the package during power supply to the system device.

FIG. 3 is a schematic diagram of a self-diagnosis circuit that is implemented when a package is replaced.

FIG. 4 is a flowchart 1 of a procedure for package replacement and data guarantee when a collectable error occurs in a package in the system unit.

FIG. 5 is a flowchart 2 of a procedure for package replacement and data guarantee when a collectable error occurs in a package in the system unit.

FIG. 6 is a schematic view of an embodiment of a water cooling device and a hot-swap jig.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... System device 2 ... Control part 3 ... Memory part 4 ... Data transfer / package switching control circuit 5 ... Diagnostic control circuit 6 ... Memory control circuit 7 ... Address / package selection control circuit 8 ... Data buffer circuit 9 ... Data comparison Circuit 10 ... Stack for error address freeze 11 ... ECC circuit 12 ... Stack for error data freeze 13 ... Tool for hot-plugging / unplugging 14 ... Tool start terminal 16 ... Reset signal / power sequence generation circuit 17 ... Package 18 ... Memory / control Circuit 19 ... Self-diagnosis circuit 20 ... Self-diagnosis start signal 21 ... Self-diagnosis end signal 22 ... Self-diagnosis result signal 23 ... Diagnostic signal group 24 ... Diagnostic pass / fail judgment circuit 25 ... Module 26 ... Sub water supply device 27 ... Water supply hose 28 ... Sub Water supply hose 29 ... Hose jack 30 ... Sub hose jack 31 ... N Hose jack 32 ... cock 33 ... sub-water supply route for the hose

Claims (1)

[Claims] 1. A device having a sub-unit inserted in a slot, a spare slot for inserting the sub-unit for replacement, and a diagnostic provided for the sub-unit for replacement to confirm operation by self-diagnosis. A hot-swap device having a circuit and a diagnostic control circuit for giving a self-diagnosis instruction to a diagnostic circuit of the replacement subunit through a spare slot.