CN105634813A - Method for automatically switching nodes under dual-computer environment based on network - Google Patents

Method for automatically switching nodes under dual-computer environment based on network Download PDF

Info

Publication number
CN105634813A
CN105634813A CN201610000774.3A CN201610000774A CN105634813A CN 105634813 A CN105634813 A CN 105634813A CN 201610000774 A CN201610000774 A CN 201610000774A CN 105634813 A CN105634813 A CN 105634813A
Authority
CN
China
Prior art keywords
node
network
mds
mdt
ost
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.)
Pending
Application number
CN201610000774.3A
Other languages
Chinese (zh)
Inventor
宋辰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Inspur Electronic Information Industry Co Ltd
Original Assignee
Inspur Electronic Information Industry Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Inspur Electronic Information Industry Co Ltd filed Critical Inspur Electronic Information Industry Co Ltd
Priority to CN201610000774.3A priority Critical patent/CN105634813A/en
Publication of CN105634813A publication Critical patent/CN105634813A/en
Pending legal-status Critical Current

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06Management of faults, events, alarms or notifications
    • H04L41/0654Management of faults, events, alarms or notifications using network fault recovery
    • H04L41/0668Management of faults, events, alarms or notifications using network fault recovery by dynamic selection of recovery network elements, e.g. replacement by the most appropriate element after failure

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Hardware Redundancy (AREA)

Abstract

The invention discloses a method for automatically switching nodes under a dual-computer environment based on a network, belongs to a method for automatically switching nodes, and solves the problem that how to avoid the unavailability of the whole Lustre file system caused by the downtime of a single-point metadata server. The technical scheme is as follows: the management node, the standby management node and the login node are all connected to the mdt node and the ost node through the Ethernet switch, and the storage server is respectively connected to the management node, the standby management node, the login node, the mdt node and the ost node through the Ethernet switch; (1) deploying heartbeat service at all mds nodes and oss nodes; (2) modifying ha.cf file codes according to the actual environment of the cluster; (3) starting heartbeat service, and checking whether all IO nodes run the service; (4) manually dropping the Ethernet port of the MDS node, and observing the switching process; (5) and confirming the residual recovery time, and confirming that the Lustre partition is still normal after the time _ remaining is timed out.

Description

A kind of method that network two-shipper environment lower node automatically switches
Technical field
The present invention relates to a kind of method that node automatically switches, a kind of method that specifically network two-shipper environment lower node automatically switches.
Background technology
Instantly HPC high-performance computing sector, the requirement of I/O bandwidth is increased by be skyrocketed through and the computational tasks of data volume day by day, and NFS file system can not meet the demand of NFS. Lustre is as a parallel file system increased income, and its powerful scalability has been widely used in HPCC environment.
But while capacity and bandwidth disclosure satisfy that calculating I/O bandwidth demand along with Lustre file system, the pressure of Lustre server is also gradually increased, especially meta data server (MDS). As the node of storage Lustre metadata, pressure is more big, and fault rate is also more high. High availability is self-evident for the importance of cluster, not only safeguards stablizing of cluster hardware structure, reduces the generation of fault, and can ensure that stablizing of file system. Once cluster file system breaks down, being catastrophic for cluster, bring the interruption even loss of data of production environment, risk is self-evident.
Summary of the invention
The technical assignment of the present invention is to provide a kind of method that network two-shipper environment lower node automatically switches, and solves the disabled problem how avoiding the single-point meta data server machine of delaying to cause whole Lustre file system.
The technical assignment of the present invention realizes in the following manner,
A kind of method that network two-shipper environment lower node automatically switches, involved hardware includes storage server, InfiniBand switch, Ethernet switch, management node, standby management node, logs in node, mds node and oss node, management node, standby management node, logging in node and connect to mdt node and ost node each through Ethernet switch, storage server via Ethernet switch is connected respectively to management node, standby management node, logs in node, mdt node and ost node; Described method comprises the steps:
(1), service at all mds nodes and oss node deployment heartbeat;
(2), ha.cf document code is revised according to cluster actual environment;
(3), open heartbeat service, check whether that all I/O node have all run this service;
(4), do not unload Lustre subregion, manually the Ethernet interface down of MDS node is fallen, observe handoff procedure;
(5), confirm to remain recovery time, after treating time_remaining timing, confirm that Lustre subregion is still normal.
Mds node includes MDS01 node and MDS02 node, and MDS01 node is mdt host node, and MDS02 node is mdt secondary node.
Oss node includes OSS01 node, OSS02 node, OSS03 node and OSS04 node; OSS01 node, OSS02 node, OSS03 node and OSS04 node are ost carry node.
OSS01 node carry ost00 and ost01; OSS02 node carry ost02 and ost03; OSS03 node carry ost04 and ost05; OSS04 node carry ost06, ost07.
The method that a kind of network two-shipper environment lower node of the present invention automatically switches has the advantage that
1, by the method monitor in real time network Heartbeat, under two-shipper environment when host node is due to malfunction and failure, host node fault-signal is informed that secondary node, secondary node take over the service of host node or the carry of memory space by heartbeat mechanism automatically. By writing script the MDS node being deployed in Lustre file system and OSS node, by the service redundant of both nodes, it is achieved the non-stop run of mdt, it is ensured that the normal operation of Lustre file system;
2, this deployment way is disposed based on script, and by installing related service under assigned catalogue, timing detects network environment, and self only takes up a small amount of system resource. And by the amendment to script, can be applicable to multiple different HA environment, colony environment;
3, after this application is disposed, not affecting storage and file system performance, take storage server resource little, after MDS active node switches, mdt recovers availability automatically, it is not necessary to manual operation; When, after OSS single point failure, another OSS being mutually redundant takes over the ost of inefficacy, automatic carry, and checks availability. To be checked complete, recover the read-write of former ost.
Accompanying drawing explanation
Below in conjunction with accompanying drawing, the present invention is further described.
Accompanying drawing 1 is the hardware block diagram of a kind of method that network two-shipper environment lower node automatically switches.
Detailed description of the invention
The method a kind of network two-shipper environment lower node of the present invention automatically switched with reference to Figure of description and specific embodiment is described in detail below.
Embodiment 1:
The method that a kind of network two-shipper environment lower node of the present invention automatically switches, involved hardware includes storage server, InfiniBand switch, Ethernet switch, management node, standby management node, logs in node, mds node and oss node, management node, standby management node, logging in node and connect to mdt node and ost node each through Ethernet switch, storage server via Ethernet switch is connected respectively to management node, standby management node, logs in node, mdt node and ost node; Described method comprises the steps:
(1), service at all mds nodes and oss node deployment heartbeat;
(2), ha.cf document code is revised according to cluster actual environment;
(3), open heartbeat service, check whether that all I/O node have all run this service;
(4), do not unload Lustre subregion, manually the Ethernet interface down of MDS node is fallen, observe handoff procedure;
(5), confirm to remain recovery time, after treating time_remaining timing, confirm that Lustre subregion is still normal.
Mds node includes MDS01 node and MDS02 node, and MDS01 node is mdt host node, and MDS02 node is mdt secondary node.
Oss node includes OSS01 node, OSS02 node, OSS03 node and OSS04 node; OSS01 node, OSS02 node, OSS03 node and OSS04 node are ost carry node.
OSS01 node carry ost00 and ost01; OSS02 node carry ost02 and ost03; OSS03 node carry ost04 and ost05; OSS04 node carry ost06, ost07.
In step (2), ha.cf document code is:
keepalive2
deadtime30
initdead120
#definedifferentudpportfordifferentpairs
#
udpport694
bcasteth0
use_logdoff
logfile/var/log/ha-log
auto_failbackoff
#
#youmustchangehere
#
nodemds01mds02
ping11.11.11.111.11.11.2
respawnhacluster/usr/lib64/heartbeat/ipfail
#addstonith
#stonith_hostmd2external/rackpdu
#stonithexternal/rackpdu/etc/ha.d/rackpdu.conf��
In step (4), observing handoff procedure is check MDS node or OSS node:
/ proc/fs/lustre/mdt/lustre-MDT0000/recovery_status and
/proc/fs/lustre/obdfilter/lustre-OST0000/recovery_status��
By detailed description of the invention above, described those skilled in the art can be easy to realize the present invention. It is understood that the present invention is not limited to above-mentioned detailed description of the invention. On the basis of disclosed embodiment, described those skilled in the art can the different technical characteristic of combination in any, thus realizing different technical schemes.
Except the technical characteristic described in description, it is the known technology of those skilled in the art.

Claims (4)

1. the method that a network two-shipper environment lower node automatically switches, it is characterized in that involved hardware includes storage server, InfiniBand switch, Ethernet switch, management node, standby management node, logs in node, mds node and oss node, management node, standby management node, logging in node and connect to mdt node and ost node each through Ethernet switch, storage server via Ethernet switch is connected respectively to management node, standby management node, logs in node, mdt node and ost node; Described method comprises the steps:
(1), service at all mds nodes and oss node deployment heartbeat;
(2), ha.cf document code is revised according to cluster actual environment;
(3), open heartbeat service, check whether that all I/O node have all run this service;
(4), manually the Ethernet interface down of MDS node is fallen, observe handoff procedure;
(5), confirm to remain recovery time, after treating time_remaining timing, confirm that Lustre subregion is still normal.
2. the method that a kind of network two-shipper environment lower node according to claim 1 automatically switches, it is characterised in that mds node includes MDS01 node and MDS02 node, and MDS01 node is mdt host node, and MDS02 node is mdt secondary node.
3. the method that a kind of network two-shipper environment lower node according to claim 1 automatically switches, it is characterised in that oss node includes OSS01 node, OSS02 node, OSS03 node and OSS04 node; OSS01 node, OSS02 node, OSS03 node and OSS04 node are ost carry node.
4. the method that a kind of network two-shipper environment lower node according to claim 3 automatically switches, it is characterised in that OSS01 node carry ost00 and ost01; OSS02 node carry ost02 and ost03; OSS03 node carry ost04 and ost05; OSS04 node carry ost06, ost07.
CN201610000774.3A 2016-01-04 2016-01-04 Method for automatically switching nodes under dual-computer environment based on network Pending CN105634813A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610000774.3A CN105634813A (en) 2016-01-04 2016-01-04 Method for automatically switching nodes under dual-computer environment based on network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610000774.3A CN105634813A (en) 2016-01-04 2016-01-04 Method for automatically switching nodes under dual-computer environment based on network

Publications (1)

Publication Number Publication Date
CN105634813A true CN105634813A (en) 2016-06-01

Family

ID=56049351

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610000774.3A Pending CN105634813A (en) 2016-01-04 2016-01-04 Method for automatically switching nodes under dual-computer environment based on network

Country Status (1)

Country Link
CN (1) CN105634813A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107291390A (en) * 2017-06-20 2017-10-24 郑州云海信息技术有限公司 A kind of data classification storage and device
CN109445709A (en) * 2018-11-05 2019-03-08 郑州云海信息技术有限公司 The management method and device of storage resource in virtualization system

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103095837A (en) * 2013-01-18 2013-05-08 浪潮电子信息产业股份有限公司 Method achieving lustre metadata server redundancy
CN104023061A (en) * 2014-06-10 2014-09-03 浪潮电子信息产业股份有限公司 High availability cluster scheme of OSS (Open Storage service) for LUSTRE
CN104135513A (en) * 2014-07-24 2014-11-05 浪潮集团有限公司 A method for realizing high availability cluster by PowerPC cloud storage platform using Heartbeat
CN104679907A (en) * 2015-03-24 2015-06-03 新余兴邦信息产业有限公司 Realization method and system for high-availability and high-performance database cluster
CN105117300A (en) * 2015-08-12 2015-12-02 浪潮(北京)电子信息产业有限公司 Apparatus for realizing high availability of heartbeat

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103095837A (en) * 2013-01-18 2013-05-08 浪潮电子信息产业股份有限公司 Method achieving lustre metadata server redundancy
CN104023061A (en) * 2014-06-10 2014-09-03 浪潮电子信息产业股份有限公司 High availability cluster scheme of OSS (Open Storage service) for LUSTRE
CN104135513A (en) * 2014-07-24 2014-11-05 浪潮集团有限公司 A method for realizing high availability cluster by PowerPC cloud storage platform using Heartbeat
CN104679907A (en) * 2015-03-24 2015-06-03 新余兴邦信息产业有限公司 Realization method and system for high-availability and high-performance database cluster
CN105117300A (en) * 2015-08-12 2015-12-02 浪潮(北京)电子信息产业有限公司 Apparatus for realizing high availability of heartbeat

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
张晓波: "基于高性能集群计算的并行文件系统关键技术研究", 《西安电子科技大学硕士学位论文》 *
李江昀等: "基于Linux平台的过程控制双机热备综合解决方案", 《计算机工程与应用》 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107291390A (en) * 2017-06-20 2017-10-24 郑州云海信息技术有限公司 A kind of data classification storage and device
CN107291390B (en) * 2017-06-20 2020-05-15 苏州浪潮智能科技有限公司 Data hierarchical storage method and device
CN109445709A (en) * 2018-11-05 2019-03-08 郑州云海信息技术有限公司 The management method and device of storage resource in virtualization system

Similar Documents

Publication Publication Date Title
US11360854B2 (en) Storage cluster configuration change method, storage cluster, and computer system
CN107111457B (en) Non-disruptive controller replacement in cross-cluster redundancy configuration
EP2435916B1 (en) Cache data processing using cache cluster with configurable modes
CN106776121B (en) Data disaster recovery device, system and method
CN106815097A (en) Database disaster tolerance system and method
CN104536971A (en) High-availability database
CN105069160A (en) Autonomous controllable database based high-availability method and architecture
CN103763155A (en) Multi-service heartbeat monitoring method for distributed type cloud storage system
JP2005209201A (en) Node management in high-availability cluster
CN102467508A (en) Method for providing database service and database system
CN104717077B (en) A kind of method, apparatus and system for managing data center
CN111045602A (en) Cluster system control method and cluster system
CN107395771A (en) Full redundancy balancing the load production process data acquisition system
CN111399766B (en) Data storage method, data reading method, device and system in storage system
CN113254275A (en) MySQL high-availability architecture method based on distributed block device
CN105634813A (en) Method for automatically switching nodes under dual-computer environment based on network
CN113961397A (en) High-availability cluster disaster tolerance method based on backup disaster tolerance system
CN105323271B (en) Cloud computing system and processing method and device thereof
CN102984009A (en) Disaster recovery backup method for VoIP (Voice overInternet Protocol) system based on P2P
CN105607872A (en) Storage apparatus
CN117421158A (en) Database fault processing method, system and storage medium
CN112231399A (en) Method and device applied to graph database
CN103384267A (en) Parastor200 parallel storage management node high availability method based on distributed block device
CN107819619A (en) A kind of continual method of access for realizing NFS
CN106777238B (en) A kind of self-adapted tolerance adjusting method of HDFS distributed file 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
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20160601