CN105607954B - A kind of method and apparatus that stateful container migrates online - Google Patents
A kind of method and apparatus that stateful container migrates online Download PDFInfo
- Publication number
- CN105607954B CN105607954B CN201510973448.6A CN201510973448A CN105607954B CN 105607954 B CN105607954 B CN 105607954B CN 201510973448 A CN201510973448 A CN 201510973448A CN 105607954 B CN105607954 B CN 105607954B
- Authority
- CN
- China
- Prior art keywords
- node
- host node
- physical host
- container
- snapshot
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5083—Techniques for rebalancing the load in a distributed system
- G06F9/5088—Techniques for rebalancing the load in a distributed system involving task migration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
- G06F3/0608—Saving storage space on storage systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
- G06F3/0646—Horizontal data movement in storage systems, i.e. moving data in between storage devices or systems
- G06F3/0647—Migration mechanisms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0668—Interfaces specially adapted for storage systems adopting a particular infrastructure
- G06F3/067—Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Software Systems (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention provides a kind of method and apparatus that stateful container migrates online, comprising: the snapshot of creation stateful container data volume to be migrated;By the way that the virtual router on the first physical host node and on the second physical host node is arranged in, across physical host capacitor network is established, the snapshot is transmitted to the second physical host node;The first physical host node is the host node where stateful container to be migrated;The second physical host node is the destination host of the stateful container to be migrated;Determine data newly-increased after the snapshot creation starts;The newly-increased data are transmitted to the second physical host node by the capacitor network;The snapshot and the newly-increased data are written on the second physical host node in the book for newly creating container, stateful container is rebuild.Using technical solution of the present invention, migrating online across physical host for stateful container can be realized.
Description
Technical field
The present invention relates to computer field of cloud computer technology, in particular to a kind of method that stateful container migrates online and
Device.
Background technique
As linux container technology graduallys mature, especially Container Management tool Docker's yields unusually brilliant results,
OpenShift, Cloud Foundry and Baidu's cloud application engine etc. start to execute cloud application as safety using Docker container
Sandbox.Amazon AWS, Google App Engine and Microsoft Azure also release one after another container service.
In order to facilitate the resilient expansion for realizing application, the application being deployed on cloud computing application engine is all usually stateless
, i.e., the memory of these applications and file system only serve data cached, and really realize lasting data preservation is rear end
Storage service, including database server, key-value storage service and message system etc..Therefore these rear end storage services are claimed
For stateful cloud service.
Stateful container is the container for having data storage function, can save data.And stateless container cannot save number
According to.The migration of data volume container is referred to, container is transferred to another host system from a host system.Stateless container
The migration of service is fairly simple, it might even be possible to realize online migration.However, existing cloud computing application engine not branch held states
The migration of container service.
Summary of the invention
The embodiment of the present invention proposes a kind of method and apparatus that stateful container migrates online, to realize stateful appearance
The online migration of device service.
The embodiment of the invention provides a kind of methods that stateful container migrates online, include the following steps:
Create the snapshot of stateful container data volume to be migrated;
By the way that the virtual router on the first physical host node and on the second physical host node is arranged in, establish across object
The capacitor network for managing host, is transmitted to the second physical host node for the snapshot;The first physical host node is
Host node where stateful container to be migrated;The second physical host node is the mesh of the stateful container to be migrated
Mark host;
Determine data newly-increased after the snapshot creation starts;
The newly-increased data are transmitted to the second physical host node by the capacitor network;
The snapshot and the newly-increased data are written on the second physical host node to the data for newly creating container
In volume, stateful container is rebuild.
The embodiment of the invention provides a kind of devices that stateful container migrates online, comprising:
Snapshot creating unit, for creating the snapshot of stateful container data volume to be migrated;
Transmission unit, for by the way that the virtual road on the first physical host node and on the second physical host node is arranged in
By device, across physical host capacitor network is established, the snapshot is transmitted to the second physical host node;First physics
Host node is the host node where stateful container to be migrated;The second physical host node is described to be migrated to have shape
The destination host of state container;
Data determination unit, for determining that the snapshot creation starts data newly-increased later;
The transmission unit is further used for that the newly-increased data are transmitted to described second by the capacitor network
Physical host node;
Data write unit, for the second physical host node to be written in the snapshot and the newly-increased data
In the book of new creation container, stateful container is rebuild.
The present invention has the beneficial effect that:
The method and apparatus that a kind of stateful container provided in an embodiment of the present invention migrates online, by being arranged to be migrated
Virtual router where stateful container on host node and on destination host node establishes capacitor network, transmits to be migrated
The snapshot of stateful container data volume transmits newly-increased number after determining the data increased newly after snapshot creation starts to destination host
According to destination host, in the book of newly created container, will be rebuild stateful on snapshot and newly-increased data write-in destination host
Migrating online across physical host for stateful container may be implemented in container.
Detailed description of the invention
Specific embodiments of the present invention are described below with reference to accompanying drawings,
Fig. 1 is the method flow schematic diagram that stateful container migrates online in the embodiment of the present invention;
Fig. 2 is that stateful container data rolls up the flow diagram migrated online in the embodiment of the present invention;
Fig. 3 is cluster management schematic diagram in the embodiment of the present invention;
Fig. 4 is cluster period distribution schematic diagram in the embodiment of the present invention;
Fig. 5 is primary node status transition diagram in the embodiment of the present invention;
Fig. 6 is daily record reproduction schematic diagram in the embodiment of the present invention;
Fig. 7 is OpenContrail system framework figure in the embodiment of the present invention;
Fig. 8 is that cluster network manages schematic diagram in the embodiment of the present invention;
Fig. 9 is scheduling of resource schematic diagram in the embodiment of the present invention;
Figure 10 is the apparatus structure schematic diagram that stateful container migrates online in the embodiment of the present invention.
Specific embodiment
In order to which technical solution of the present invention and advantage is more clearly understood, below in conjunction with attached drawing to of the invention exemplary
Embodiment is described in more detail, it is clear that and the described embodiments are merely a part of the embodiments of the present invention, rather than
The exhaustion of all embodiments.And in the absence of conflict, the feature in the embodiment and embodiment in this specification can be with
It is combined with each other.
Fig. 1 is the method flow schematic diagram that stateful container migrates online in the embodiment of the present invention, as shown in Figure 1, there is shape
The method that state container migrates online may include steps of:
Step 101: the snapshot of creation stateful container data volume to be migrated;
Step 102: by the way that the virtual router on the first physical host node and on the second physical host node is arranged in,
The capacitor network across physical host is established, the snapshot is transmitted to the second physical host node;The first physics master
Machine node is the host node where stateful container to be migrated;The second physical host node is described to be migrated stateful
The destination host of container;
Step 103: determining data newly-increased after the snapshot creation starts;
Step 104: the newly-increased data are transmitted to by the second physical host node by the capacitor network;
Step 105: the snapshot and the newly-increased data being written newly to create on the second physical host node and held
In the book of device, stateful container is rebuild.
Snapshot is a completely available copy to specified data acquisition system, the copy include corresponding data at some time point
The image at (time point that copy starts).That is, snapshot can be a copy of the data represented by it, it is also possible to data
One duplicate.The effect of snapshot is mainly able to carry out online data backup and restores.When application failure occurs for storage equipment
Or quick data recovery can be carried out when file corruption, data are restored to the state at some available time point.Snapshot
Another effect is to provide another data access channel for storage user, when former data carry out application on site processing,
The accessible snapshot data of user can also carry out the work such as testing using snapshot.
It determines that the snapshot creates and starts data newly-increased later, in specific implementation, can be and determine the snapshot transmission
To described to be migrated after completing the stateful container to be migrated out of service, and determine since snapshot creation
Newly-increased data before stateful container is out of service in this period;It is also possible to determining that snapshot creation starts
Afterwards, out of service after enabling the stateful container to be migrated continue to run a period of time, determine since the snapshot creation it
Newly-increased data to the stateful container to be migrated before out of service in this period afterwards.Wherein, described to be migrated to have
The time that status container continues to run after determining that the snapshot creation starts can be by those skilled in the art according to practical feelings
Condition determines, is not specifically limited here.
In specific implementation, the snapshot is written on the second physical host node with the newly-increased data and is newly created
In the book of container, it can be the snapshot that will be first transmitted and be first written on the second physical host node and newly create container
Book in, then newly-increased data are written on the second physical host node in the book for newly creating container;It can also
To be to merge the snapshot and newly-increased data the data for newly creating container are written on the second physical host node together
In volume;It is also possible to that first newly-increased data are written on the second physical host node in the book for newly creating container, then
The snapshot is written on the second physical host node in the book for newly creating container.The specific snapshot and newly-increased
The write-in data of data and be merge after be written on the second physical host node in the book for newly creating container, here not
It is specifically limited.
The method that a kind of stateful container provided in an embodiment of the present invention migrates online, by being arranged to be migrated stateful
Virtual router where container on host node and on destination host node establishes the capacitor network across physical host, transmission
The snapshot of stateful container data volume to be migrated is to destination host, and after determining the data increased newly after snapshot creation starts, transmission is new
The data of increasing in the book of newly created container, will be rebuild to destination host on snapshot and newly-increased data write-in destination host
Migrating online across host for stateful container may be implemented in stateful container.
In implementation, the snapshot for creating the book of stateful container to be migrated be can specifically include:
The snapshot of book based on Btrfs system creation stateful container to be migrated, the file system of the container are
Btrfs system.
In specific implementation, due to the good scalability of Btrfs system, B-tree and dynamic inode (information node,
Information node/index node) characteristics such as creation ensure that Btrfs system still has brilliant performance, entirety in large machines
Performance is reduced without the increase with power system capacity.And when the unpredictable hardware fault of systems face, Btrfs system
It can guarantee the consistency of file system using COW (Copy On Write, copy-on-write) technology.Btrfs system is also supported always
And check code, avoid the appearance of no alarm failure.In addition, Btrfs system is supported creation snapshot (snapshot) and is cloned
(clone), multiple physical equipments can easily be managed.Btrfs system can also significantly improve time/space of file system
Performance, including delayed allocation, the storage optimization of small documents, directory index etc..
So can choose Btrfs system in the embodiment of the present invention as physical host storage number where stateful container
According to file system.Fig. 2 is the flow diagram of stateful container data volume migration in the embodiment of the present invention, as shown in Fig. 2,
Before starting a stateful PostgresSQL database server container from node 1, it is connected first with from node 1
Storage pool "/data " in create " pg " catalogue, for storing data data of library server.Then official's mirror image is used
Postgres runs a PostgresSQL container A 1, and wherein operating parameter "-v " is mapped as "/the data/pg " of host hold
The book "/var/lib/postgresql/data " of device A1.If stateful container A 1 is moved to from node 2, need
It executes the following steps: (1) creating the snapshot "/data/pg.snapshot " of "/data/pg " from node 1 and transmitting from section
The snapshot "/data/pg.snapshot " of "/data/pg " that creates on point 1 arrives the storage pool from node 2.According to database
Type, file it is of different sizes, the step for may need to continue for some time;(2) out of service after snapshot is sent
Stateful container A 1 finds out the difference from "/the data/pg.snapshot " and "/data/pg " of snapshot on node 1, i.e. snapshot is created
Newly-increased data after building, are sent to from node 2, with "/the data/pg.snapshot " of node 2 merge into one it is new "/
data/pg".Since incremental data volume is smaller, this step can be completed faster.
In implementation, by the way that the virtual router on the first physical host node and on the second physical host node is arranged in,
Establishing across physical host capacitor network can also include: before the snapshot is transmitted to the second physical host node
The first virtual router is set on the first physical host node;
The second virtual router is set on the second physical host node;First and second virtual router is established
Connection, institute's virtual router have kernel state forwarding capability and/or User space control function.
In specific implementation, since existing stateful Container Management tool, such as Docker etc. are not supported across physical host
Stateful container between networking.Network technology can be defined with application software in the embodiment of the present invention, in each stateful appearance
One virtual router is set on the host where device, and the virtual router of setting can be support MPLS (Multiprotocol
Label Switching, multiprotocol label switching) over GRE (Generic Routing Encapsulation, general road
By encapsulating)/UDP (User Datagram Protocol, User Datagram Protocol) or VXLAN (Virtual
EXtensible Local Area Network, virtual expansible local area network) etc. agreements gateway router, for connecting not
With the physical host where stateful container.
Virtual router can be a kernel module, virtual network and corresponding virtual flow-line needed for being created by it
Table.If virtual machine or stateful container are connected on the virtual router of the machine, a pair of virtual ether is created first
Network interface, one is connected with virtual machine/stateful container, another is connected with the virtual routing tables of virtual network.It is different
The tunnel MPLS over GRE/UDP or VxLAN is established between the virtual router of physical host node.In this way, being located at difference
Physical host node but network communication can be established by belonging to the different virtual machine of same virtual network or stateful container
?.
Wherein, the virtual router of setting may include kernel state forwarding capability and/or User space control function, kernel state
It is mapped with User space with respective memory, i.e. address space.The application program of user's space, is called by system, into interior
Nuclear space.The process is represented by kernel and runs on kernel spacing, user's space and kernel spacing have different address of cache, lead to
With or dedicated register group, and the process of user's space will transmit many variables, parameter to kernel spacing, and kernel spacing is also wanted
Some registers, the variable etc. for saving consumer process, return to user's space after calling so as to system and continue to execute.It operates in
Program under User space cannot direct access operation system kernel data structure and program.When executing a program in systems,
Most of the time operates under User space, its need operating system help through it is certain it there is no power and ability to complete
Work when will be switched to kernel state.
In specific implementation, the first virtual router is set on the first physical host node and in second physics
The step of the second virtual router is arranged on host node has no absolute sequential relationship, as long as guaranteeing in step 102 by setting
The virtual router on the first physical host node and on the second physical host node is set, across physical host container net is established
Network is completed before the snapshot is transmitted to the second physical host node, can guarantee the first physical host node
It can be communicated with the second physical host node, this is easy as known to those skilled in the art, to repeat no more.
It, will with the virtual router on the second physical host node by being arranged on the first physical host node in implementation
The snapshot is transmitted to the second physical host node and can specifically include:
It will be described fast by the way that the virtual router on the first physical host node and on the second physical host node is arranged in
According to tunnel encapsulation at being sent to the second physical host node after data packet.
In specific implementation, since tunneling technique is that a kind of infrastructure by using internet is transmitted between networks
The mode of data.The data (or load) transmitted using tunnel can be the data frame or packet of different agreement.Tunnel protocol by its
Then the data frame or packet of its agreement Reseal to be sent by tunnel.New frame head provides routing iinformation, will pass through interconnection
The packed load data of net transmitting.Tunneling technique can make the network service from many information sources in the same infrastructure
In transmitted by different tunnels.Tunneling technique passes through routing network using the exchange connection of point-to-point communication protocol instead
Network connects data address.Tunneling technique allows that mobile subscriber or the user authorized is authorized to visit at any time and any place
Ask enterprise network.So by being arranged on the first physical host node and the second physical host node in the embodiment of the present invention
On virtual router by the snapshot tunnel encapsulation at being sent to the second physical host node after data packet.
In implementation, determine that the data newly-increased later that are transmitted of the snapshot can specifically include:
The data increased newly during determination is out of service to the stateful container to be transferred since snapshot creation.
In specific implementation, newly-increased data, which can be, to be started when being transmitted of the snapshot to described to be transferred to have
In this period of status container that time out of service, the data that are increased newly in the stateful container to be migrated.It is described to
It is any a certain moment that can choose the time after being transmitted of the snapshot that it is out of service, which to shift stateful container,.In order to
Shorten the time of the stateful container migration to be transferred, can the snapshot transmission once complete it is out of service described in
Stateful container is shifted, the data volume of newly-increased data is minimized.In general, the snapshot transmission once complete stop fortune
The processing mode of the row stateful container to be transferred, newly-increased data volume only have several hundred KB, stateful container service it is offline
Time can be very short.
In implementation, the snapshot and the newly-increased data are written on the second physical host node and newly create container
Book in, rebuild after stateful container, can also include:
The network configuration information of the stateful container to be migrated is sent to the second physical host node, the net
Network configuration information includes: affiliated virtual network information, (Internet Protocol, the iso-ip Internetworking protocol ISO-IP) address IP and/or position
In Hostname;
The routing policy of the router on the second physical host node is updated according to the network configuration information.
In specific implementation, the snapshot is written on the second physical host node with the newly-increased data and is newly created
It in the book of container, rebuilds after stateful container, can also check the network configuration letter of the stateful container to be migrated
Breath, the information such as virtual network, IP address and host name including the stateful container connection to be migrated, be also sent to it is described to
Migrate the destination host of stateful container.(3) on the destination host of the stateful container to be migrated, virtual router is added
The virtual network where the stateful container to be migrated is added to, as shown in Figure 2 with official's mirror database postgre operation one
A Postgre SQL container B 2, wherein operating parameter "-v " "/the data/pg " of host be mapped as container B 2 book "/
Var/lib/postgresql/data ", according to the network of the configuration setting B2 of A1.By above-mentioned steps, from having on node 1
Status container is just successfully migrated to from node 2.
In implementation, the snapshot and the newly-increased data are written on the second physical host node and newly create container
Book in, rebuild after stateful container, can also include:
The configuration information of the second physical host node is obtained by proxy module, the proxy module is arranged described
On second physical host node;The configuration information includes resource information, resource occupation status information and/or creates on this node
The information of container built or run, the first and second physical host node are from node;
The configuration information is sent to main section corresponding with the second physical host node by the proxy module
Point;
The host node is according to the configuration information update and reports global configuration information to corresponding leader node;
By the leader node according to the resource status of the global configuration information update container cluster.
In specific implementation, the method for stateful container migration provided in an embodiment of the present invention can be using in the cluster.Fig. 3
For cluster management schematic diagram in the embodiment of the present invention, as shown in figure 3, physical host cluster is mapped as one by cluster management component
Resource pool.After user submits job task, cluster management component selects suitable physical host to create according to scheduling strategy
With the stateful container of operation execution task.According to the difference of role, the physical host for forming cluster is divided into host node and from section
Point two major classes.The configuration information of the centrally stored all nodes of host node, scheduling fulfil assignment task from node.It receives and leads from node
The dispatch command of node creates and runs stateful container;The configuration of this node and stateful container is reported to host node simultaneously
The case where information and stateful container occupy resource.Host node fails and causes cluster unavailable in order to prevent, generally can be
Multiple host nodes are disposed in cluster, it will usually dispose 3,5 or 7 host nodes, including 1 leader (leader) and more
A follower (follower).From node meeting and only session (TCP connection) can be established with a host node.It is asked when from node transmission
When asking, the host node being attached thereto forwards the request to leader's node first.Leader's node is the order for including in request
It is added in log, is then copied to follower's node.The present invention guarantees the storage state of each host node using Raft agreement
Consistency.
Fig. 4 is cluster period distribution schematic diagram in the embodiment of the present invention, as shown in figure 4, in the cluster, the time, which is divided into, appoints
The period for length of anticipating.The serial number of period is the integer continuously increased since 1.Each period will complete leader when starting
Election is coordinated the consistency of host node by it if successfully electing leader's node, completes configuration management and resource
The functions such as scheduling;If election failure, terminates present period, restarts a new period.
Host node in cluster only can be in one of following three state: follower, candidate or leader, three it
Between there are specific switch condition, Fig. 5 is primary node status transition diagram in the embodiment of the present invention, as described in Figure 5, in cluster
When starting, all host nodes are all in " follower " state.Any one host node is led as long as it is received from one
The far call of people (or candidate) node is requested, it will be always maintained at " follower " state.And leader's node can be determined
Phase sends heartbeat message to all follower's nodes.If follower's node does not receive any letter in heart beat cycle
Breath, it will be considered that leader's node is no longer valid, then open new round leader election.
Follower's node is incremented by the serial number of present period first, the state of oneself is converted to from " follower " " candidate
People ".Then it can vote to oneself, and the information of the host node transmitted in parallel request ballot all other into cluster.The section
Point can be constantly in " candidate " state, and until one of following three condition is set up: (1) it has won election, becomes leader
Node;(2) there are other nodes to win election, become leader's node;(3) the election time expires, but is elected without node
For leader.
The condition that candidate's node wins election is that most of host node is all thrown ballot paper to it.When one fixed
Section, each host node are voted according to the principle of prerequisite variable to candidate's node.Once candidate's node wins choosing
It lifts, it just becomes leader's node, sends heartbeat message at once to other host nodes, to prevent new election.
In election process, there are two types of special circumstances to be treated.The first situation be candidate's node A waiting and
When collecting the ballot paper of other host nodes, it is possible to will receive the data of another node B transmission, and B node claims oneself
It is leader's node.At this point, A node can compare the size of the period serial number of two nodes.If the period serial number of B is greater than A's
The state of oneself then A recognizes that B is a legal leader, and is revised as " follower " from " candidate " by period serial number;
Otherwise, A node can refuse the request of B node, continue to keep candidate's state.
Second of special circumstances is in this period, and no node wins election.For example, if there is many nodes are almost same
When switch to candidate, may cause none of candidate's node can obtain most ballot papers.At this point, all candidate's sections
Point can terminate present period after election time-out, open a new election period.Occurs election failure repeatedly in order to prevent
The case where, it is the different election time-out of each candidate's node sets, i.e., in a regular time section (for example, 150-300
Millisecond) in randomly select a timeout value.This method can efficiently reduce because of election failure time caused by ballot paper dispersion
Number guarantees the leader's node for rapidly selecting cluster.
The snapshot and the newly-increased data are written on the second physical host node to the data for newly creating container
It in volume, rebuilds after stateful container, completes the migration of stateful container, operate in newly created on the second physical host node
After stateful container, actuator reports the occupation condition of container to the proxy module from node 1, by from node 1
Proxy module is reported to leader's node 2, modifies the available resources state of cluster.Actuator monitors newly created stateful container
Operating status, once stateful container operation error, actuator report error message to scheduler, moreover it is possible to according to preparatory
Error handle setting, is restarted automatically container.
In implementation, the host node report updated global configuration information to the corresponding leader node of the host node it
Before, can also include:
The host node saves the configuration information to the cluster global configuration information using memory key-value database
In.
In specific implementation, because key assignments storing data library is a kind of non-relational database, data are with key-value pair
Form carries out tissue, index and storage, is suitble to not be related to the business datum of excessive data relationship business relations, while can effectively subtract
The number for reading and writing disk less, is gathered around than SQL (Structured Query Language, structured query language) database purchase
There is better readwrite performance, so, report updated global configuration information corresponding to the host node in the host node
Before leader node, the configuration information can be saved using memory key-value database into the cluster global configuration information,
To guarantee that the cluster global configuration information has good readwrite performance.
In theory, as long as other skills of the configuration information into the cluster global configuration information can be saved
Art is also that this programme can be implemented, and it is specifically how real that memory key-value Database Systems are only used for introduction those skilled in the art
It applies the present invention, but it is not intended that be only capable of using a kind of mode of memory key-value Database Systems, can be needed in conjunction with practice in implementation process
To determine corresponding mode.
Request is sent from node to the host node being attached thereto, the life that each request needs host node to execute comprising one
It enables.All requests are forwarded to leader's node there.Leader's node adds a new log first, including order,
Receive the serial number of the period of order and the call number of log.Journal entry at this time is in and does not submit state, i.e., there are no execution
Order.Next, leader's node copies to journal entry all follower's nodes.If most follower's nodes are
It is successfully written into the journal entry received, leader's node can submit the journal entry, i.e., the order for including in execution journal item,
Then all follower's nodes is notified also to submit the journal entry.In this way, the state of most host nodes keeps one in cluster
It causes.
Fig. 6 is daily record reproduction schematic diagram in the embodiment of the present invention, as shown in fig. 6, a configuration " X=is written from node 1
5".What is be attached thereto is host node 1, this write request is transmitted to leader's node 2 by it.Host node 2 generate one it is new
Journal entry " (4,2, X=5) ", wherein " 4 " are the call numbers of journal entry, " 2 " are period serial numbers at this time, and " X=5 " is to hold
Capable order.Next, journal entry " (4,2, X=5) " is sent to all follower's nodes, i.e., main section by leader's node 2
Point 1 and 3.If they can return to success there is no the log to conflict with the journal entry in the log of node 1 and node 3 oneself
The message of journal entry is written.After leader's node 2 receives message, the journal entry is submitted, i.e., the value of the X of storage is revised as
" 5 " finally notify follower's node also to submit this log.Node 1 and node 3 are also respectively revised as the value of the X of storage
"5".At this point, the state of 3 host nodes is consistent in cluster, the notice of leader's node 1 has successfully been write from 1 data of node
Enter.In above process, if confirming write-in journal entry without most follower's nodes, leader's node 1 will not submitting day
Will item, and notify to fail from 1 write request of node.
Above-mentioned leader's election and daily record reproduction process can guarantee the configuration information submitted from node to be written to main section
In point cluster, and efficiently solve host node Problem of Failure.
In implementation, believed by the leader node according to the global configuration of the global configuration information update container cluster
Breath.
The container cluster establishes across physical host capacitor network, the Open using Open Contrail system
Node in Contrail system includes analysis node, control node and configuration node;The control node is according to the analysis
The analysis of node is as a result, update the global configuration information of the capacitor network;The configuration node saves the capacitor network
Global configuration information.
Fig. 7 is OpenContrail system framework figure in the embodiment of the present invention, as shown in fig. 7, disposing in the cluster first
One OpenContail system, it is made of a series of node.Each node is one and physically or a virtually services
Device undertakes different roles.Configuration node, analysis node and control node collectively constitute OpenContrail controller, by it
Manage the virtual network that stateful container or virtual machine on multiple nodes separately constitute.Wherein, configuration node saves all
The configuration information of node and network;The information of analysis node collection, storage and analysis virtual network;Control node realizes network
Control plane function, and the data forwarding level function of network, then it is real by the virtual router installed in each calculate node
It is existing.
Fig. 8 is that cluster network manages schematic diagram in the embodiment of the present invention, as shown in figure 8, all belonging to from node 1 and from node 2
Therefore node in OpenContrail system will install and configure virtual router on the two nodes.If from node
Tri- stateful containers of A1, A2 and A3 have been run on 1, and two stateful containers of B1 and B2 have been run from node 2.Each
2 virtual networks net1 and net2 are all created on virtual router, each virtual router there are respective virtual routing tables.?
From node 1, stateful container A 1 and A2 are assigned to virtual network net1, stateful container A 3 is assigned to virtual network
net2;From node 2, stateful container B 1 is assigned to virtual network net1, stateful container B 2 is assigned to network
net2.A pair of of virtual ethernet network interface is created for each stateful container, one is connected with status container, another company
Connect the corresponding virtual routing forwarding of virtual network.(the Dynamic Host Configuration of DHCP built in virtual router
Protocol, dynamic host configuration protocol) server, the stateful container being connected with each virtual network can dynamic acquisition
To a network address.Configuration node collects the virtual network configuration of all virtual routers, and virtual router is notified to exchange
Route table items.It therefore, finally also can include the routing of status container B1 from the routing table of the virtual network net1 of node 1,
It and also can include the routing of status container A1 and A2 from the routing table of the virtual network net1 of node 2.When stateful container
When A1 sends data packet to stateful container B 1, virtual router one layer of tunnel envelope of addition on the basis of raw data packets first
Dress is subsequently sent to from node 2, and the latter is transmitted to stateful container B 1 according to the information of tunnel encapsulation, then initial data.
Under implied terms, the node in different virtual networks can not be communicated.For example, A1 and A3 can not just establish network
Connection.If supporting the communication between virtual network, it is necessary to add corresponding routing policy toward virtual router manually.
In theory, as long as the other technologies for being that by network virtualization are also that this programme can be implemented
, OpenContrail system is only used for how introduction those skilled in the art specifically implement the present invention, but it is not intended that is only capable of making
With a kind of mode of OpenContrail system, can be needed in conjunction with practice in implementation process to determine corresponding mode.
In implementation, it is arranged before virtual router on the first physical host node and on the second physical host node, also
May include:
Determine the second physical host node.
In specific implementation, on the first physical host node and on the second physical host node be arranged virtual router it
Before, it may be predetermined that the resources such as some CPU, memory are suitble to the physical host of the stateful container migration to be migrated as institute
State the second physical host node.
In implementation, determine that the second physical host node can specifically include:
Stateful container migration request is sent to corresponding host node by the first physical host node;
Forward the stateful container migration request to corresponding leader node based on the host node;
The stateful appearance is met according to the determination of the resource status of the stateful container cluster by the leader node
The slave node of device migration request resource requirement is the second physical host node.
In specific implementation, physical host cluster uses two-level scheduler mechanism, and Fig. 9 is that scheduling of resource is shown in the embodiment of the present invention
It is intended to, the money as shown in figure 9, all configuration informations for reporting this node from node to leader's node, on leader's node
Source distribution module provides the resource operation API of standard.Developer calls resource distribution module, writes different operation frames.
Each frame is suitble to run the application program of specific type (for example, residing permanently service, batch processing or stream process operation), mainly
It is made of scheduler and actuator.Scheduler is commonly installed on the primary node, it is according to the scheduling plan such as rotating or have priority
Slightly, the task of operation is assigned to available from node;Actuator is mounted on from node, receives and executes scheduler dispatches
Task, while the executive condition of monitor task.Actuator calls the container of Docker creation and operation execution task, realizes not
Isolation and resource constraint with task.
Configuration from node 1 is 4 core CPU and 4GB memories.Leader's node 2 is received from node 1 with postponing, and is notified that
The scheduler of each frame.The scheduler of frame 1 checks the operation that user submits, and discovery operation 1 includes two tasks, is respectively required for
Want 2 core CPU/2GB memories and monokaryon CPU/1GB memory.From the configuration meet demand of node 1.Then, the scheduler handle of frame 1
This two tasks are distributed to from node 1.The mission bit stream of transmission includes frame title (being frame 1 here), from node ID (this
In be from node 1), resource limit (CPU, memory etc.) and execution required by task Docker mirror image or for constructing mirror image
Dockerfile.The actuator and Docker of frame 1 have been mounted with from node 1.It is held if it is necessary, actuator constructs first
The Docker mirror image of row required by task.Actuator is using Docker mirror image as template, and using the resource limited as start-up parameter, operation is held
The container of row task.
In the embodiment of the present invention when determining the second physical host node, can by first physical host from
Node sends stateful container migration request to corresponding host node, then forwards the stateful container by the host node
Migration request is to corresponding leader node, by the leader node according to the resource status of the stateful container cluster
The slave node for being determined to meet the stateful container migration request resource requirement is the second physical host node, i.e., really
Determining resource distribution to be migrated has shape as described more than or equal to the slave node of the stateful container migration request resource requirement
The target physical host node of state container.
Based on the same inventive concept, a kind of device that stateful container migrates online is additionally provided in the embodiment of the present invention,
Since the principle that the device solves the problems, such as is similar to a kind of method that stateful container migrates online, the implementation of the device can
With referring to the implementation of method, overlaps will not be repeated.
Figure 10 is the apparatus structure schematic diagram of stateful container migration in the embodiment of the present invention, be it is as shown in Figure 10, this has
Status container migration device, may include:
Snapshot creating unit, for creating the snapshot of stateful container data volume to be migrated;
Transmission unit, for by the way that the virtual road on the first physical host node and on the second physical host node is arranged in
By device, across physical host capacitor network is established, the snapshot is transmitted to the second physical host node;First physics
Host node is the host node where stateful container to be migrated;The second physical host node is described to be migrated to have shape
The destination host of state container;
Data determination unit, for determining that the snapshot creation starts data newly-increased later;
The transmission unit is further used for that the newly-increased data are transmitted to described second by the capacitor network
Physical host node;
Data write unit, for the second physical host node to be written in the snapshot and the newly-increased data
In the book of new creation container, stateful container is rebuild.
In implementation, specifically for being based on, B-tree file system Btrfs system creation is to be migrated shape to the snapshot creating unit
The snapshot of the book of state container, the file system of the container are Btrfs system.
In implementation, the device of the stateful container migration can also include:
Setting unit, for by the way that the virtual road on the first physical host node and on the second physical host node is arranged in
By device, across physical host capacitor network is established, before the snapshot is transmitted to the second physical host node, described the
The first virtual router is set on one physical host node;The second virtual flow-line is set on the second physical host node
Device;First and second virtual router establishes connection, and institute's virtual router has kernel state forwarding capability and/or User space control
Function processed.
In implementation, the transmission unit specifically can be used for by being arranged on the first physical host node and the second physics
Virtual router on host node is sent to the second physical host node after the snapshot is carried out tunnel encapsulation
In implementation, the data determination unit is specifically determined for since creating the snapshot to described to be transferred
The data increased newly during stateful container is out of service.
In implementation, the transmission unit can be further used for the snapshot and the newly-increased data write-in described the
It in the book for newly creating container on two physical host nodes, rebuilds after stateful container, by the stateful appearance to be migrated
The network configuration information of device is sent to the second physical host node, and the network configuration information includes: affiliated virtual network
Information, IP address and/or the Hostname being located at;
The device of the stateful container migration can also include:
Routing policy updating unit, for being updated on the second physical host node according to the network configuration information
The routing policy of router.
In implementation, the device of the stateful container migration can also include:
Acquiring unit is newly created for the snapshot to be written on the second physical host node with the newly-increased data
It builds in the book of container, rebuilds after stateful container, matching for the second physical host node is obtained by proxy module
Confidence breath, the proxy module are arranged on the second physical host node;The configuration information includes resource information, resource
Occupied state information and/or the information of container for creating or running on this node, the first and second physical host node are from section
Point;
The transmission unit is further used for being sent to the configuration information and described second by the proxy module
The corresponding host node of physical host node;
Updating unit, for by the host node according to the configuration information update and report global configuration information to
Its corresponding leader node;By the leader node according to the resource status of the global configuration information update container cluster.
In implementation, the device of the stateful container migration can also include:
Storage unit reports updated global configuration information to the corresponding leader of the host node for the host node
Before node, the host node saves the configuration information to the cluster global configuration information using memory key-value database
In.
In implementation, the updating unit is specifically used for establishing by the container cluster using Open Contrail system
Across physical host capacitor network, the node in the Open Contrail system includes analysis node, control node and configuration section
Point;The control node is according to the analysis of the analysis node as a result, updating the global configuration information of the capacitor network;It is described
Configuration node saves the global configuration information of the capacitor network.
In implementation, the device of the stateful container migration can also include:
Destination host determination unit, it is virtual for being arranged on the first physical host node and on the second physical host node
Before router, the second physical host node is determined.
In implementation, the destination host determination unit is specifically used for sending by the first physical host node stateful
Container migration request is to corresponding host node;Based on the host node forward the stateful container migration request to and its
Corresponding leader node;Have according to the determining satisfaction of the resource status of the stateful container cluster by the leader node
The slave node of status container migration request resource requirement is the second physical host node.
For convenience of description, each section of apparatus described above is divided into various modules with function or unit describes respectively.
Certainly, each module or the function of unit can be realized in same or multiple softwares or hardware in carrying out the present invention.
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the present invention
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the present invention, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although preferred embodiments of the present invention have been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications may be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the scope of the invention.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (22)
1. a kind of method that stateful container migrates online, which comprises the steps of:
Create the snapshot of stateful container data volume to be migrated;
By the way that the virtual router on the first physical host node and on the second physical host node is arranged in, across physics master is established
The snapshot is transmitted to the second physical host node by the capacitor network of machine;The first physical host node is wait move
Move the host node where stateful container;The second physical host node is the target master of the stateful container to be migrated
Machine;
Determine data newly-increased after the snapshot creation starts;
The newly-increased data are transmitted to the second physical host node by the capacitor network;
The snapshot and the newly-increased data are written on the second physical host node in the book for newly creating container,
Rebuild stateful container.
2. the method as described in claim 1, which is characterized in that the snapshot of the book of creation stateful container to be migrated is specific
Include:
The snapshot of book based on B-tree file system Btrfs system creation stateful container to be migrated, the file of the container
System is Btrfs system.
3. the method as described in claim 1, which is characterized in that by being arranged on the first physical host node and the second physics
Virtual router on host node establishes across physical host capacitor network, the snapshot is transmitted to the second physics master
Before machine node, further includes:
The first virtual router is set on the first physical host node;
The second virtual router is set on the second physical host node;First and second virtual router establishes connection,
The virtual router has kernel state forwarding capability and/or User space control function.
4. the method as described in claim 1, which is characterized in that by being arranged on the first physical host node and the second physics
Virtual router on host node establishes across physical host capacitor network, the snapshot is transmitted to the second physics master
Machine node specifically includes:
It is shone into fastly by the way that the virtual router on the first physical host node and on the second physical host node is arranged in by described
The second physical host node is sent to after row tunnel encapsulation.
5. the method as described in claim 1, which is characterized in that determine the snapshot is transmitted data tool newly-increased later
Body includes:
The data increased newly during determination is out of service to the stateful container to be migrated since snapshot creation.
6. the method as described in claim 1, which is characterized in that by the snapshot and the newly-increased data write-in described second
In the book for newly creating container on physical host node, rebuild after stateful container, further includes:
The network configuration information of the stateful container to be migrated is sent to the second physical host node, the network is matched
Confidence breath includes: affiliated virtual network information, IP address and/or the Hostname being located at;
The routing policy of the router on the second physical host node is updated according to the network configuration information.
7. the method as described in claim 1, which is characterized in that by the snapshot and the newly-increased data write-in described second
In the book for newly creating container on physical host node, rebuild after stateful container, further includes:
The configuration information of the second physical host node is obtained by proxy module, the proxy module is arranged described second
On physical host node;The configuration information include resource information, resource occupation status information and/or on this node creation or
The information of container of operation, the first and second physical host node are from node;
The configuration information is sent to host node corresponding with the second physical host node by the proxy module;
The host node is according to the configuration information update and reports global configuration information to corresponding leader node;
By the leader node according to the resource status of the global configuration information update container cluster.
8. the method for claim 7, which is characterized in that the host node reports updated global configuration information to institute
Before stating the corresponding leader node of host node, further includes:
The host node saves the configuration information into the cluster global configuration information using memory key-value database.
9. the method for claim 7, which is characterized in that by the leader node according to the global configuration information more
The resource status of new container cluster specifically includes:
The container cluster establishes across physical host capacitor network, the Open Contrail using Open Contrail system
Node in system includes analysis node, control node and configuration node;The control node is divided according to the analysis node
Analysis is as a result, update the global configuration information of capacitor network;The configuration node saves the global configuration information of the capacitor network.
10. the method for claim 7, which is characterized in that on the first physical host node and the second physical host section
It is arranged before virtual router on point, further includes:
Determine the second physical host node.
11. method as claimed in claim 10, which is characterized in that determine that the second physical host node specifically includes:
Stateful container migration request is sent to corresponding host node by the first physical host node;
Forward the stateful container migration request to corresponding leader node based on the host node;
The stateful container migration request is met according to the determination of the resource status of the container cluster by the leader node
The slave node of resource requirement is the second physical host node.
12. a kind of device that stateful container migrates online characterized by comprising
Snapshot creating unit, for creating the snapshot of stateful container data volume to be migrated;
Transmission unit, for by the way that the virtual flow-line on the first physical host node and on the second physical host node is arranged in
Device establishes across physical host capacitor network, the snapshot is transmitted to the second physical host node;The first physics master
Machine node is the host node where stateful container to be migrated;The second physical host node is described to be migrated stateful
The destination host of container;
Data determination unit, for determining that the snapshot creation starts data newly-increased later;
The transmission unit is further used for that the newly-increased data are transmitted to second physics by the capacitor network
Host node;
Data write unit is newly created for the snapshot to be written on the second physical host node with the newly-increased data
It builds in the book of container, rebuilds stateful container.
13. device as claimed in claim 12, which is characterized in that the snapshot creating unit is specifically used for being based on B-tree file
The snapshot of the book of system Btrfs system creation stateful container to be migrated, the file system of the container are Btrfs system
System.
14. device as claimed in claim 12, which is characterized in that further include:
Setting unit, for by the way that the virtual flow-line on the first physical host node and on the second physical host node is arranged in
Device establishes across physical host capacitor network, before the snapshot is transmitted to the second physical host node, described first
First virtual router is set on physical host node;The second virtual router is set on the second physical host node;
First and second virtual router establishes connection, and the virtual router has kernel state forwarding capability and/or User space control
Function.
15. device as claimed in claim 12, which is characterized in that the transmission unit is specifically used for by being arranged in the first object
Virtual router on reason host node and on the second physical host node is sent to institute after the snapshot is carried out tunnel encapsulation
State the second physical host node.
16. device as claimed in claim 12, which is characterized in that the data determination unit is specifically used for determining from described fast
Start the data increased newly during out of service to the stateful container to be migrated according to creation.
17. device as claimed in claim 12, which is characterized in that the transmission unit is further used for the snapshot and institute
Newly-increased data are stated to be written on the second physical host node in the book for newly creating container, rebuild stateful container it
Afterwards, the network configuration information of the stateful container to be migrated is sent to the second physical host node, the network is matched
Confidence breath includes: affiliated virtual network information, IP address and/or the Hostname being located at;
The device of the stateful container migration, further includes:
Routing policy updating unit, for updating the routing on the second physical host node according to the network configuration information
The routing policy of device.
18. device as claimed in claim 12, which is characterized in that further include:
Acquiring unit, for the snapshot and the newly-increased data to be written on the second physical host node, newly creation is held
In the book of device, rebuild after stateful container, by proxy module obtain the second physical host node with confidence
Breath, the proxy module are arranged on the second physical host node;The configuration information includes resource information, resource occupation
Status information and/or the information of container for creating or running on this node, the first and second physical host node are from node;
The transmission unit is further used for being sent to the configuration information and second physics by the proxy module
The corresponding host node of host node;
Updating unit, for according to the configuration information update and reporting global configuration information extremely right with it by the host node
The leader node answered;By the leader node according to the resource status of the global configuration information update container cluster.
19. device as claimed in claim 18, which is characterized in that further include:
Storage unit reports updated global configuration information to the corresponding leader node of the host node for the host node
Before, the host node saves the configuration information into the cluster global configuration information using memory key-value database.
20. device as claimed in claim 18, which is characterized in that the updating unit is specifically used for passing through the container cluster
Across physical host capacitor network is established using Open Contrail system, the node in the Open Contrail system includes
Analysis node, control node and configuration node;The control node is according to the analysis of the analysis node as a result, updating the appearance
The global configuration information of device network;The configuration node saves the global configuration information of the capacitor network.
21. device as claimed in claim 18, which is characterized in that further include:
Destination host determination unit, for virtual flow-line to be arranged on the first physical host node and on the second physical host node
Before device, the second physical host node is determined.
22. device as claimed in claim 21, which is characterized in that the destination host determination unit is specifically used for by described
First physical host node sends stateful container migration request to corresponding host node;Institute is forwarded based on the host node
Stateful container migration request is stated to corresponding leader node;By the leader node according to the money of the container cluster
Source state determines that the slave node for meeting the stateful container migration request resource requirement is the second physical host node.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510973448.6A CN105607954B (en) | 2015-12-21 | 2015-12-21 | A kind of method and apparatus that stateful container migrates online |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510973448.6A CN105607954B (en) | 2015-12-21 | 2015-12-21 | A kind of method and apparatus that stateful container migrates online |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105607954A CN105607954A (en) | 2016-05-25 |
CN105607954B true CN105607954B (en) | 2019-05-14 |
Family
ID=55987908
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510973448.6A Active CN105607954B (en) | 2015-12-21 | 2015-12-21 | A kind of method and apparatus that stateful container migrates online |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105607954B (en) |
Families Citing this family (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106506255B (en) * | 2016-09-21 | 2019-11-05 | 微梦创科网络科技(中国)有限公司 | A kind of method, apparatus and system of pressure test |
CN106502830B (en) * | 2016-10-27 | 2019-01-22 | 一铭软件股份有限公司 | A kind of method for restoring system backup based on Btrfs file system |
CN108268304B (en) * | 2017-01-03 | 2021-06-08 | 中国科学院声学研究所 | Web app migration method based on container |
CN107070691B (en) * | 2017-01-12 | 2020-01-21 | 阿里巴巴集团控股有限公司 | Cross-host communication method and system of Docker container |
CN108628874B (en) * | 2017-03-17 | 2020-12-22 | 北京京东尚科信息技术有限公司 | Method and device for migrating data, electronic equipment and readable storage medium |
CN107193489A (en) * | 2017-05-15 | 2017-09-22 | 郑州云海信息技术有限公司 | A kind of storage level data migration method and device based on storage virtual gateway |
CN107193499A (en) * | 2017-05-26 | 2017-09-22 | 郑州云海信息技术有限公司 | A kind of moving method and device of container data volume |
CN109150941B (en) * | 2017-06-27 | 2021-04-27 | 中国科学院信息工程研究所 | Data center physical resource floating method |
CN109246167B (en) * | 2017-07-11 | 2022-04-01 | 阿里巴巴集团控股有限公司 | Container scheduling method and device |
CN107515782A (en) * | 2017-07-26 | 2017-12-26 | 北京天云融创软件技术有限公司 | Implementation method of the container across host migration under a kind of Docker environment |
US10691504B2 (en) | 2017-08-14 | 2020-06-23 | International Business Machines Corporation | Container based service management |
CN107391318A (en) * | 2017-09-18 | 2017-11-24 | 郑州云海信息技术有限公司 | A kind of method, system and device of container data persistence |
CN107729020B (en) * | 2017-10-11 | 2020-08-28 | 北京航空航天大学 | Method for realizing rapid deployment of large-scale container |
CN107728947A (en) * | 2017-10-13 | 2018-02-23 | 郑州云海信息技术有限公司 | A kind of cold moving method of container and system |
CN108170514A (en) * | 2017-12-19 | 2018-06-15 | 中国科学院声学研究所 | A kind of container operation method and device |
EP3522013B1 (en) | 2018-02-01 | 2020-04-01 | Siemens Aktiengesellschaft | Method and system for migration of containers in a container orchestration platform between compute nodes |
CN108279969A (en) * | 2018-02-26 | 2018-07-13 | 中科边缘智慧信息科技(苏州)有限公司 | Stateful service container thermomigration process based on memory compression transmission |
CN108897557B (en) * | 2018-06-20 | 2022-06-10 | 中国联合网络通信集团有限公司 | Updating method and device of microservice architecture |
CN111343219B (en) * | 2018-12-18 | 2022-08-02 | 同方威视技术股份有限公司 | Computing service cloud platform |
CN111352689B (en) * | 2018-12-21 | 2023-04-07 | 中国电信股份有限公司 | Method and device for realizing seamless migration of application containerized deployment |
CN109857516B (en) * | 2018-12-29 | 2023-12-12 | 航天信息股份有限公司 | Cluster migration method and device based on container |
CN111092770B (en) * | 2019-12-23 | 2021-09-14 | 联想(北京)有限公司 | Virtual network management method and electronic equipment |
CN111190702A (en) * | 2020-01-14 | 2020-05-22 | 山东超越数控电子股份有限公司 | Cross-node hot migration method for Docker |
CN111339061B (en) * | 2020-02-12 | 2023-09-26 | 杭州涂鸦信息技术有限公司 | Data migration method and system for distributed database |
CN111694789A (en) * | 2020-04-22 | 2020-09-22 | 西安电子科技大学 | Embedded reconfigurable heterogeneous determination method, system, storage medium and processor |
CN116204305B (en) * | 2022-12-21 | 2023-11-03 | 山东未来网络研究院(紫金山实验室工业互联网创新应用基地) | Method for limiting number of dock container inodes |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101169740A (en) * | 2007-08-07 | 2008-04-30 | 北京大学 | EJB component dynamic migration method |
EP1636704A4 (en) * | 2003-06-09 | 2008-06-11 | Ind Defender Inc | Event monitoring and management |
CN103024053A (en) * | 2012-12-18 | 2013-04-03 | 华为技术有限公司 | Cloud storage method, resource scheduling system and cloud storage node and system |
WO2013112023A1 (en) * | 2012-01-26 | 2013-08-01 | Samsung Electronics Co., Ltd. | Processing state information |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2007538313A (en) * | 2004-04-29 | 2007-12-27 | インターナショナル・ビジネス・マシーンズ・コーポレーション | System and method for modeling and dynamically deploying services within a distributed networking architecture |
-
2015
- 2015-12-21 CN CN201510973448.6A patent/CN105607954B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1636704A4 (en) * | 2003-06-09 | 2008-06-11 | Ind Defender Inc | Event monitoring and management |
CN101169740A (en) * | 2007-08-07 | 2008-04-30 | 北京大学 | EJB component dynamic migration method |
WO2013112023A1 (en) * | 2012-01-26 | 2013-08-01 | Samsung Electronics Co., Ltd. | Processing state information |
CN103024053A (en) * | 2012-12-18 | 2013-04-03 | 华为技术有限公司 | Cloud storage method, resource scheduling system and cloud storage node and system |
Non-Patent Citations (2)
Title |
---|
《Web应用服务器关键技术研究》;范国闯;《中国博士学位论文全文数据库 信息科技辑》;20040301;全文 |
《一种面向构件化软件的在线演化方法》;王晓鹏;《计算机学报》;20051130(第11期);全文 |
Also Published As
Publication number | Publication date |
---|---|
CN105607954A (en) | 2016-05-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105607954B (en) | A kind of method and apparatus that stateful container migrates online | |
US11120044B2 (en) | System and method for maintaining a master replica for reads and writes in a data store | |
CN109067828B (en) | Kubernetes and OpenStack container-based cloud platform multi-cluster construction method, medium and equipment | |
CN107426034B (en) | Large-scale container scheduling system and method based on cloud platform | |
US11928506B2 (en) | Managing composition service entities with complex networks | |
US11947697B2 (en) | Method and system to place resources in a known state to be used in a composed information handling system | |
CN102103518B (en) | System for managing resources in virtual environment and implementation method thereof | |
CN105074702B (en) | The Database Systems of single tenant and multi-tenant environment are provided | |
CN112099918A (en) | Live migration of clusters in containerized environments | |
WO2020062131A1 (en) | Container cloud management system based on blockchain technology | |
Essa et al. | Mobile agent based new framework for improving big data analysis | |
CN109547512A (en) | A kind of method and device of the distributed Session management based on NoSQL | |
KR102438214B1 (en) | ICT service provision method and system | |
CN113839814B (en) | Decentralized Kubernetes cluster federal implementation method and system | |
TW201434300A (en) | Service migration across cluster boundaries | |
CN112463366A (en) | Cloud-native-oriented micro-service automatic expansion and contraction capacity and automatic fusing method and system | |
CN109462511B (en) | Network establishing method and device | |
WO2020088340A1 (en) | Method and system for providing cloud computing service | |
CN106326226A (en) | Method and system for starting database service on public cloud | |
CN111966482A (en) | Edge computing system | |
CN111343219A (en) | Computing service cloud platform | |
CN109992373B (en) | Resource scheduling method, information management method and device and task deployment system | |
JP2024501005A (en) | Management method and device for container clusters | |
CN116723077A (en) | Distributed IT automatic operation and maintenance system | |
US11290318B2 (en) | Disaster recovery of cloud resources |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |