Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Deadlock detecting method, equipment and system

A deadlock detection and deadlock technology, applied in the field of clusters, can solve problems such as poor real-time performance and inapplicability to distributed cluster environments

Active Publication Date: 2014-03-26
HUAWEI TECH CO LTD
View PDF9 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The deadlock detection method in the prior art has poor real-time performance and is not suitable for a distributed cluster environment

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Deadlock detecting method, equipment and system
  • Deadlock detecting method, equipment and system
  • Deadlock detecting method, equipment and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0047] An embodiment of the present invention provides a deadlock detection method for a proxy server, and the proxy server establishes communication connections with the cluster lock manager and the first thread client device respectively, see figure 2 , the method flow includes:

[0048] 201. Receive the first thread client device's application for the first lock;

[0049] 202. Send an authorization request to the cluster lock manager according to the application, and apply for authorization of the first lock;

[0050] 203. Receive a waiting message returned by the cluster lock manager;

[0051] 204. Query the local real-time lock application records, and if it is found that the client device of the first thread holds the second lock, initiate a traceback search of the dependency path of the first lock in the cluster;

[0052] 205. Determine whether a deadlock is formed according to the result of the dependency path backtracking search.

[0053] Further, the real-time re...

Embodiment 2

[0078] The embodiment of the present invention provides a deadlock detection method, see image 3 , the method flow includes:

[0079] 301. The local Proxy receives the request for lock a from the thread client device T1, suspends the thread client device T1, and applies to the Master for authorization of the lock a.

[0080] 302. The local Proxy receives the waiting message returned by the Master;

[0081] 303. The local Proxy queries the real-time lock application record of T1, finds that T1 holds lock x, and executes step 304.

[0082] Specifically, after the Proxy receives the authorization from the Master to the Proxy, it needs to perform secondary authorization (that is, distribution) to the local lock applicant (that is, the thread client device), so the Proxy needs to maintain the lock application and authorization status of the local thread client device The lock application is recorded in real time. The local Proxy can query the locks already held by T1 and the lo...

Embodiment 3

[0102] The embodiment of the present invention also provides a deadlock detection method, see Figure 6 , to explain the commonly used ABBA locks in the cluster. The thread client device T1 and Proxy1 are on node 1, and the thread client device T2 and Proxy2 are on node 2. The method flow includes:

[0103] 601. The thread client device T1 applies to the Proxy1 for the authorization of the lock 1, and the Proxy1 suspends T1, sends a Request (request) message to the Master, and applies for the authorization of the lock 1.

[0104] 602. The Master sends a Grant (authorization) instruction to Proxy1, and sends the authorization of lock 1 to Proxy1, and Proxy1 activates T1 (that is, performs a second authorization of lock 1 to T1). At this time, T1 holds lock 1.

[0105] 603. The thread client device T2 applies to the Proxy2 for the authorization of the lock 2, and the Proxy2 suspends T2, and sends a Request message to the Master, requesting to apply for the authorization of the l...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a deadlock detecting method, equipment and system and belongs to the technical field of clusters. The method includes: receiving an application of a first lock from first thread client equipment; transmitting an authorization request to a cluster lock manager according to the application so as to apply for the authorization of the first lock; receiving the waste message returned by the cluster lock manager; inquiring the local lock application real-time records, and starting backtracking search of the dependence path of the first lock in a cluster if the fact that the first thread client equipment holds a second lock is discovered; judging whether deadlock occurs or not according to the search result. The method has the advantages that when the thread client equipment applies to hang a certain lock, a proxy server searches whether the client equipment holds another lock or not, if so, the backtracking search of the dependence path of applied lock is started, real-time detection of deadlock is achieved, and the method is simple and efficient.

Description

technical field [0001] The invention relates to the field of cluster technology, in particular to a deadlock detection method, device and system. Background technique [0002] A cluster is a single, unified computing resource with high availability, high performance, and manageability. In the pursuit of cluster performance, distributed clusters came into being. To ensure the consistency of the cluster, cluster locks are essential. There are three entities in the implementation of the two-level authorized distributed lock: cluster Master (lock manager), local Proxy (proxy server), and thread client device. Taking an n-node cluster as an example, each node has a Proxy (that is, a local Proxy); the thread client device is the applicant for the lock, identified by the Tread (thread) number, and can apply for multiple locks from the Proxy of the node; The Master is responsible for the distribution authorization of each lock in the cluster, arbitrates the lock applications of d...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06F11/36G06F9/46
Inventor 岳亮
Owner HUAWEI TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com