CN103312564B - InfiniBand network detecting method - Google Patents
InfiniBand network detecting method Download PDFInfo
- Publication number
- CN103312564B CN103312564B CN201310253119.5A CN201310253119A CN103312564B CN 103312564 B CN103312564 B CN 103312564B CN 201310253119 A CN201310253119 A CN 201310253119A CN 103312564 B CN103312564 B CN 103312564B
- Authority
- CN
- China
- Prior art keywords
- port
- equipment
- makeing mistakes
- port number
- infiniband network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000005516 engineering process Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000547 structure data Methods 0.000 description 1
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
- Small-Scale Networks (AREA)
Abstract
The invention discloses a kind of InfiniBand network detecting method, including the second corresponding relation of the device name of equipment obtained in the InfiniBand network physical port number with each port in first corresponding relation of No. LID and equipment and logical port number；No. LID of the equipment at the port place of makeing mistakes in acquisition InfiniBand network and the logical port number of port of makeing mistakes；The device name of the equipment at port place of makeing mistakes and the physical port number of port of makeing mistakes is obtained according to the first corresponding relation, the second corresponding relation, No. LID of equipment of port place of makeing mistakes and the logical port number of port of makeing mistakes.Present invention may be broadly applicable to the cluster of different scales, once run and just can find out the whole problematic link of IB network, it is possible to carry out integrated with other adjustment methods, at utmost reduce human error, substantially increase the debugging efficiency of large-scale IB network, and save time and manpower.
Description
Technical field
The invention relates essentially to network communication field, more specifically, relate to a kind of InfiniBand network detecting method.
Background technology
It is a kind of high-speed network technology and standard that InfiniBand (is called for short IB), there is the advantage such as high bandwidth, low latency, being widely used in fields such as high-performance calculation, big data process, HPCC generally uses InfiniBand as supercomputing network or storage network.The all normal transceiving data of each equipment that network connects it is necessary to ensure that before the node of cluster uses InfiniBand network, additionally due to physical contact when hardware electric property, cable quality and equipment are installed is likely to cause that certain port of some equipment can not keep zero error rate when sending data, it is accomplished by debugging InfiniBand network this time, find problematic equipment, and progressively go contacting between parameter or equipment with the equipment of adjustment equipment so that keep zero error rate or an extremely low error rate when data send.
Existing IB adjustment method complex steps, it is necessary to use network stress test program that whole InfiniBand network is pressed, then manually find out problematic equipment, then problem device or problem port are processed.
General step is: (1) first checks that confirmation opensm service is in running status, and each node IB subcard is in normal operating condition；(2) error count that whole IB nets all IB equipment resets；(3) IB network pressure program is run；(4) diagnostic message of IB network is exported by ibdiagnet, if it find that there is equipment to recorded mistake, need to remove to search corresponding equipment according to equipment No. lid network structure compareing ibnetdiscover output with port numbers, and be in another equipment of same link with this port；(5) problematic equipment is carried out malfunction elimination.Wherein, step (1), (2), (3) perform ratio faster, and what whole debugging process was main is operated in (4th) step.If the scale of IB net is smaller, such as tens, the number of devices every time gone wrong also can be fewer, generally only several.This time, problem device can also be quickly found in the output of manual two orders of comparison.But when the scale of cluster is relatively larger, such as more than 1000 nodes, this time, the output of ibdiagnet diagnostic message was relatively more, and the network structure data of ibnetdiscover output are more, simultaneously relative to small-scale cluster, the IB number of devices made mistakes after first installation also can be on the high side.If it will be a job very consuming time that this time still goes to contrast the equipment of searching problem by hand one by one, at large-scale Chao Suan center, once such debugging even can consume staff's half a day to time；Simultaneously because searching work is manually to search contrast, in search procedure, also can be very easy to that omission occurs.
Summary of the invention
For the defect of above-mentioned prior art, the present invention proposes a kind of InfiniBand network detecting method, addresses how to improve the technical problem of the debugging efficiency of IB network.
According to an aspect of the present invention, provide a kind of InfiniBand network detecting method, including: step S1: obtain the second corresponding relation of the physical port number with each port in first corresponding relation of No. LID and described equipment of the device name of equipment in described InfiniBand network and logical port number；Step S2: obtain No. LID of equipment of port place of makeing mistakes in described InfiniBand network and the logical port number of described port of makeing mistakes；Step S3: according to described first corresponding relation, described second corresponding relation, described in make mistakes No. LID of equipment of port place and the logical port number of described port of makeing mistakes obtain described in make mistakes the device name of equipment at port place and the physical port number of described port of makeing mistakes.
In the process, described step S1 includes: obtained the second corresponding relation of the physical port number with each port in first corresponding relation of No. LID and described equipment of the device name of equipment in described InfiniBand network and logical port number by Ibnetwork order.
In the process, described step S2 includes: by the logical port number of No. LID of the equipment at the port place of makeing mistakes in the Ibdiagnet order described InfiniBand network of acquisition and described port of makeing mistakes.
In the process, described method also includes: obtained the equipment at non-working port place and the physical port number of described non-working port by the physical port number of the device name of the equipment at described port place of makeing mistakes and described port of makeing mistakes, wherein, described non-working port and described port of makeing mistakes are in two ports on same link.
In the process, described method also includes: exported by the physical port number of the device name of the equipment at described non-working port place and described non-working port.
In the process, before described step S1, described method also includes: confirms that the Opensm service of described Infiniband network is in running status, and confirms that each node daughter card of described InfiniBand network is in normal operating condition；The error count of all devices of described InfiniBand network is reset；And to described Infiniband network operation InfiniBand network pressure program.
Present invention may be broadly applicable to the cluster of different scales, once run and just can find out the whole problematic link of IB network, it is possible to carry out integrated with other adjustment methods, at utmost reduce human error, substantially increase the debugging efficiency of large-scale IB network, and save time and manpower.
Accompanying drawing explanation
Accompanying drawing is for providing a further understanding of the present invention, and constitutes a part for description, is used for together with embodiments of the present invention explaining the present invention, is not intended that limitation of the present invention.In the accompanying drawings:
Fig. 1 is the flow chart of the overview embodiment according to the present invention；
Fig. 2 is flow chart according to a particular embodiment of the invention.
Detailed description of the invention
Below in conjunction with accompanying drawing, the preferred embodiments of the present invention are illustrated, it will be appreciated that preferred embodiment described herein is merely to illustrate and explains the present invention, is not intended to limit the present invention.
Fig. 1 is the flow chart of the overview embodiment according to the present invention, in FIG:
Step S100: obtain the second corresponding relation of the physical port number with each port in first corresponding relation of No. LID and equipment of the device name of equipment in InfiniBand network and logical port number.
Step S102: obtain No. LID of equipment of the port place of makeing mistakes in InfiniBand network and the logical port number of port of makeing mistakes.
Step S104: obtain the device name of the equipment at port place of makeing mistakes and the physical port number of port of makeing mistakes according to the first corresponding relation, the second corresponding relation, No. LID of equipment of port place of makeing mistakes and the logical port number of port of makeing mistakes.
Wherein, each step can carry out without artificial participation automatically.
In the present embodiment, corresponding relation due to the corresponding relation of device name Yu No. LID obtaining equipment and the physical port number of this each port of equipment and logical port number, therefore when detecting equipment LID and the logical port number of port of makeing mistakes, can automatically get device name and physical port number, thus decreasing artificial participation, improve Infiniband Networked E-Journals efficiency.
Fig. 2 flow chart according to a particular embodiment of the invention, in fig. 2:
Step S200: confirm that the Opensm service of Infiniband network is in running status, and confirm that each node daughter card of InfiniBand network is in normal operating condition.
Step S202: the error count of all devices of InfiniBand network is reset.
Step S204: to Infiniband network operation InfiniBand network pressure program.
Step S206: obtained the second corresponding relation of the physical port number with each port in first corresponding relation of No. LID and equipment of the device name of equipment in InfiniBand network and logical port number by Ibnetwork order.Such as, corresponding with equipment by No. LID, then the port of device interior is carried out the numbering of 1～18.
Step S208: by the logical port number of No. LID of the equipment at the port place of makeing mistakes in Ibdiagnet order acquisition InfiniBand network and port of makeing mistakes.Wherein, port of makeing mistakes here refers to the port receiving wrong data, and following non-working port refers to the port broken down producing wrong data, wherein, makes mistakes port and non-working port is for two ports on same link.
Step S210: obtain the device name of the equipment at port place of makeing mistakes and the physical port number of port of makeing mistakes according to the first corresponding relation, the second corresponding relation, No. LID of equipment of port place of makeing mistakes and the logical port number of port of makeing mistakes.Wherein it is possible to here just the information of port of makeing mistakes is exported, it is also possible to be obtained in following steps again non-working port information is exported to after the information of non-working port.
Step S212: obtained the equipment at non-working port place and the physical port number of non-working port by the physical port number of the device name of the equipment at port place of makeing mistakes and port of makeing mistakes.Wherein it is possible to here just the information of non-working port is exported, it is also possible in following steps
Step S214: the physical port number of the device name of the equipment at non-working port place and non-working port is exported.For example, it is possible to print in the way of device name adds the physical port number of port or show over the display, thus staff can be clear from recognizing which port there occurs fault.
In the present embodiment, corresponding relation due to the corresponding relation of device name Yu No. LID obtaining equipment and the physical port number of this each port of equipment and logical port number, therefore when detecting equipment LID and the logical port number of port of makeing mistakes, can automatically get device name and physical port number, just the whole problematic link of IB network can be found out thus once running, and can carry out integrated with other adjustment methods, at utmost reduce human error, substantially increase the debugging efficiency of large-scale IB network, and save time and manpower.
The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for a person skilled in the art, the present invention can have various modifications and variations.All within the spirit and principles in the present invention, any amendment of making, equivalent replacement, improvement etc., should be included within protection scope of the present invention.
Claims (5)
1. an InfiniBand network detecting method, it is characterised in that including:
Step S1: obtain the second corresponding relation of the physical port number with each port in first corresponding relation of No. LID and described equipment of the device name of equipment in described InfiniBand network and logical port number；
Step S2: obtain No. LID of equipment of port place of makeing mistakes in described InfiniBand network and the logical port number of described port of makeing mistakes；
Step S3: according to described first corresponding relation, described second corresponding relation, described in make mistakes No. LID of equipment of port place and the logical port number of described port of makeing mistakes obtain described in make mistakes the device name of equipment at port place and the physical port number of described port of makeing mistakes；
Wherein, described step S1 includes: obtained the second corresponding relation of the physical port number with each port in first corresponding relation of No. LID and described equipment of the device name of equipment in described InfiniBand network and logical port number by Ibnetwork order.
2. method according to claim 1, it is characterised in that described step S2 includes:
Logical port number by No. LID of the equipment at the port place of makeing mistakes in the Ibdiagnet order described InfiniBand network of acquisition and described port of makeing mistakes.
3. method according to claim 2, it is characterized in that, described method also includes: obtained the equipment at non-working port place and the physical port number of described non-working port by the physical port number of the device name of the equipment at described port place of makeing mistakes and described port of makeing mistakes, wherein, described non-working port and described port of makeing mistakes are in two ports on same link.
4. method according to claim 3, it is characterised in that described method also includes: the physical port number of the device name of the equipment at described non-working port place and described non-working port is exported.
5. method according to claim 1, it is characterised in that before described step S1, described method also includes:
Confirm that the Opensm service of described Infiniband network is in running status, and confirm that each node daughter card of described InfiniBand network is in normal operating condition；
The error count of all devices of described InfiniBand network is reset；And
To described Infiniband network operation InfiniBand network pressure program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310253119.5A CN103312564B (en) | 2013-06-24 | 2013-06-24 | InfiniBand network detecting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310253119.5A CN103312564B (en) | 2013-06-24 | 2013-06-24 | InfiniBand network detecting method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103312564A CN103312564A (en) | 2013-09-18 |
| CN103312564B true CN103312564B (en) | 2016-07-06 |
Family
ID=49137365
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310253119.5A Active CN103312564B (en) | 2013-06-24 | 2013-06-24 | InfiniBand network detecting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103312564B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1554055A (en) * | 2001-07-23 | 2004-12-08 | �Ƚ�װ�ù�˾ | High-availability cluster virtual server system |
| CN1647466A (en) * | 2002-04-18 | 2005-07-27 | 国际商业机器公司 | A method for providing redundancy for channel adapter failure |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8681606B2 (en) * | 2011-08-30 | 2014-03-25 | International Business Machines Corporation | Implementing redundancy on infiniband (IB) networks |
-
2013
- 2013-06-24 CN CN201310253119.5A patent/CN103312564B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1554055A (en) * | 2001-07-23 | 2004-12-08 | �Ƚ�װ�ù�˾ | High-availability cluster virtual server system |
| CN1647466A (en) * | 2002-04-18 | 2005-07-27 | 国际商业机器公司 | A method for providing redundancy for channel adapter failure |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103312564A (en) | 2013-09-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108667725A (en) | A kind of industrial AnyRouter and implementation method based on a variety of accesses and edge calculations | |
| US8793351B2 (en) | Automated configuration of new racks and other computing assets in a data center | |
| CN105015565B (en) | A kind of train fault information automatic displaying method | |
| CN102880990B (en) | Fault processing system | |
| CN103095475B (en) | The method for inspecting and system of multimode communication device | |
| CN105740121A (en) | Log text monitoring and early-warning method and apparatus | |
| CN109905268B (en) | Network operation and maintenance method and device | |
| CN106568168A (en) | Debugging method, debugger and system | |
| CN107402797A (en) | A kind of software compilation method and device | |
| CN108322318B (en) | Alarm analysis method and equipment | |
| CN106776226A (en) | The monitoring method and device of self-aided terminal | |
| CN106600120A (en) | Economic management cost control system | |
| CN112583644A (en) | Alarm processing method, device, equipment and readable storage medium | |
| CN104202328B (en) | A kind of method, configuration module and the subscription end of subscription GOOSE/SMV messages | |
| CN103312564B (en) | InfiniBand network detecting method | |
| CN105071970A (en) | Failure analysis method, failure analysis system and network management equipment | |
| CN111948993A (en) | Sample processing pipeline control system | |
| CN108512675B (en) | Network diagnosis method and device, control node and network node | |
| CN108712306A (en) | A kind of information system automation inspection platform and method for inspecting | |
| CN104517082B (en) | Electric power data acquisition apparatus and method | |
| CN103810085B (en) | A kind of method and device that module testing is carried out by comparing | |
| CN106294126B (en) | The automation formula correctness management method and device of SEN ion injection machine table | |
| CN106547697B (en) | A kind of the automation formula correctness management method and device of NISSIN ion injection machine table | |
| CN106776236A (en) | The method and apparatus of the execution of monitoring program | |
| CN106126364A (en) | A kind of fault event memory collection method based on Linux system and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220729 Address after: 100089 building 36, courtyard 8, Dongbeiwang West Road, Haidian District, Beijing Patentee after: Dawning Information Industry (Beijing) Co.,Ltd. Patentee after: DAWNING INFORMATION INDUSTRY Co.,Ltd. Address before: 100193 No. 36 Building, No. 8 Hospital, Wangxi Road, Haidian District, Beijing Patentee before: Dawning Information Industry (Beijing) Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20240407 Address after: 266100 room 211, area a, software park, 169 Songling Road, Laoshan District, Qingdao City, Shandong Province Patentee after: Zhongke Shuguang International Information Industry Co.,Ltd. Country or region after: China Address before: 100089 building 36, courtyard 8, Dongbeiwang West Road, Haidian District, Beijing Patentee before: Dawning Information Industry (Beijing) Co.,Ltd. Country or region before: China Patentee before: DAWNING INFORMATION INDUSTRY Co.,Ltd. |