CN106201561A - The upgrade method of distributed caching cluster and equipment - Google Patents
The upgrade method of distributed caching cluster and equipment Download PDFInfo
- Publication number
- CN106201561A CN106201561A CN201510216024.5A CN201510216024A CN106201561A CN 106201561 A CN106201561 A CN 106201561A CN 201510216024 A CN201510216024 A CN 201510216024A CN 106201561 A CN106201561 A CN 106201561A
- Authority
- CN
- China
- Prior art keywords
- server
- distributed caching
- current
- internal memory
- caching cluster
- 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.)
- Granted
Links
Abstract
The application provides upgrade method and the equipment of a kind of distributed caching cluster, by by all Data Migrations on server maximum for internal memory quota in sequence to other server of this distributed caching cluster, using server maximum for internal memory quota in described sequence as initial server current to be upgraded, service routine on described current server to be upgraded is upgraded, by described sequence tightly inferior on the Data Migration on current server to be upgraded to current server to be upgraded, described sequence is tightly inferior to current server to be upgraded as new server current to be upgraded, the server of distributed caching cluster can be realized upgrading when, stable transition upgrading can be carried out, escalation process can not be lost data, do not interrupt outside service.
Description
Technical field
The application relates to computer realm, particularly relate to the upgrade method of a kind of distributed caching cluster with
Equipment.
Background technology
To property to whole network system of the access efficiency of data base or disk file in general network system
Can there is vital impact.Due to the efficiency of memory cache disk to be significantly larger than, if to data
The data acquisition operations of storehouse or file all can be postponed to deposit and is read out being greatly improved its reading efficiency.
The data capacity that the caching system of unit often stores and the access ability externally provided are the most limited, and are distributed
Formula cache cluster uses whole computer cluster to come outside provides buffer service, solves buffer memory capacity not
The problem that foot and access ability limit.In existing scheme, when distributed caching cluster is upgraded, need
New service, the party is restarted after directly stopping the operation of server in distributed caching cluster and clearing up environment
Case can stop service in escalation process and lose the data cached of user, simultaneously can be to the number of user rear end
The impact of short time is caused according to storehouse.
Summary of the invention
One purpose of the application is to provide a kind of upgrade method for distributed caching cluster and equipment,
Can be when the server updating of distributed caching cluster, it is impossible to lose data, do not interrupt outside
Service.
An aspect according to the application, it is provided that the upgrade method of a kind of distributed caching cluster, bag
Include:
The size of the internal memory quota of the Servers-all according to distributed caching cluster is by Servers-all
Carry out descending sequence；
All Data Migrations server on maximum by internal memory quota in described sequence are distributed to this
On other server of cache cluster, using server maximum for internal memory quota in described sequence as initially
Server current to be upgraded；
Service routine on described current server to be upgraded is upgraded, described sequence is tightly inferior to
Data Migration on current server to be upgraded is on current server to be upgraded, by the most secondary for described sequence
In current server to be upgraded as new server current to be upgraded, repeat this step until described row
In sequence, the service routine on last server completes upgrading.
Further, in said method, the service routine on described current server to be upgraded is carried out
Upgrading, including:
Service routine on described current server to be upgraded is updated to new service routine；
Service routine after updating is restarted, and waits that the service routine after described renewal is just restarted
Often.
Further, in said method, join according to the internal memory of the Servers-all of distributed caching cluster
Before Servers-all is carried out descending sequence by the size of volume, also include:
Suspend the automatic load balancing function of the Servers-all of distributed caching cluster；
After in described sequence, the service routine on last server completes upgrading, also include:
Recover the automatic load balancing function of the Servers-all of distributed caching cluster.
Further, in said method, the automatic of Servers-all suspending distributed caching cluster is born
Carry equalization function, including:
When the Servers-all of distributed caching cluster is not at automatic data migration state, suspend institute
There is the automatic load balancing function of service；
When there being part server to be in automatic data migration state in distributed caching cluster, suspending should
The automatic load of other server being not at automatic data migration state of distributed caching cluster is equal
Weighing apparatus function, and after the described server being in automatic data migration state completes automatic data migration,
Suspend the automatic load balancing function of this server completing automatic data migration.
Further, in said method, institute server on maximum by internal memory quota in described sequence
There is Data Migration to other server of this distributed caching cluster, including:
Whether judge the current total surplus amount of storage of other server in current distributed caching cluster
The all data on the server that described internal memory quota is maximum can be accommodated；
If so, all data on server maximum for internal memory quota in described sequence are moved to respectively
On other server each of this distributed caching cluster；
If it is not, after increasing new server in current distributed caching cluster, in described sequence
Deposit all Data Migrations on the server that quota is maximum to the server newly increased, according to distributed
The size of the internal memory quota of the current Servers-all of cache cluster Servers-all is re-started by
Big to little sequence, wherein, the internal memory quota of described new server is less than clothes maximum in described sequence
The internal memory quota of business device.
Further, in said method, it is judged that other server in current distributed caching cluster
Whether current total surplus amount of storage can accommodate all numbers on the server that described internal memory quota is maximum
According to before, also include:
Obtain the currently stored amount of the Servers-all of current distributed caching cluster.
Further, in said method, obtain current distributed caching cluster Servers-all work as
After front amount of storage, also include:
Internal memory quota and the currently stored amount of the Servers-all according to current distributed caching cluster are estimated
The upgrading of the Servers-all counting current distributed caching cluster expends the time.
Further, in said method, institute server on maximum by internal memory quota in described sequence
Have data to move to respectively on other server each of this distributed caching cluster to include:
According to the respective residue amount of storage of other server in current distributed caching cluster, by institute
State all data on the server that in sequence, internal memory quota is maximum will migrate in the way of average
On other server each of this distributed caching cluster.
Further, in said method, institute server on maximum by internal memory quota in described sequence
Have data to move to respectively on other server each of this distributed caching cluster to include:
All data server on maximum by internal memory quota in described sequence are according to described sequence
Order moves on other server each of this distributed caching cluster successively.
According to the another side of the application, also provide for the updating apparatus of a kind of distributed caching cluster, including:
First device, for the size of the internal memory quota of the Servers-all according to distributed caching cluster
Servers-all is carried out descending sequence；
Second device, all data on the server that internal memory quota in described sequence is maximum are moved
Move on on other server of this distributed caching cluster, the clothes maximum by internal memory quota in described sequence
Business device is as initial server current to be upgraded；
3rd device, for the service routine on described current server to be upgraded is upgraded, will
Described sequence tightly inferior on the Data Migration on current server to be upgraded to current server to be upgraded,
Using described sequence tightly inferior to current server to be upgraded as new server current to be upgraded, repeat to hold
This device of row, until the service routine on last server completes upgrading in described sequence.
Further, in the said equipment, described 3rd device, for by described current service to be upgraded
Service routine on device is updated to new service routine, restarts the service routine after updating, and
Wait that the service routine after described renewal is restarted normally.
Further, in the said equipment, described equipment also includes the 4th device, is used for suspending distributed
The automatic load balancing function of the Servers-all of cache cluster；
Described equipment also includes the 5th device, for recovering the Servers-all of distributed caching cluster
Automatic load balancing function.
Further, in the said equipment, described 4th device, for when the institute of distributed caching cluster
When having server to be not at automatic data migration state, suspend the automatic load balancing merit of all services
Energy；
When there being part server to be in automatic data migration state in distributed caching cluster, suspending should
The automatic load of other server being not at automatic data migration state of distributed caching cluster is equal
Weighing apparatus function, and after the described server being in automatic data migration state completes automatic data migration,
Suspend the automatic load balancing function of this server completing automatic data migration.
Further, in the said equipment, described second device includes:
Unit the 2nd 1, for judging the current total of other server in current distributed caching cluster
Whether residue amount of storage can accommodate all data on the server that described internal memory quota is maximum, if
It is to forward Unit the two or two to, if it is not, then forward Unit the two or three to；
Unit the two or two, all data on the server that internal memory quota in described sequence is maximum
Unit the two or four is forwarded to after moving to respectively on other server each of this distributed caching cluster；
Unit the two or three, after increasing new server in current distributed caching cluster, by institute
State all Data Migrations on the server that internal memory quota in sequence is maximum to the server newly increased,
The size of the internal memory quota of the current Servers-all according to distributed caching cluster is by Servers-all
Unit the two or four, wherein, the internal memory of described new server is forwarded to after re-starting descending sequence
Quota is less than the internal memory quota of server maximum in described sequence；
Unit the two or four, works as initial for the server that internal memory quota in described sequence is maximum
Front server to be upgraded.
Further, in the said equipment, described equipment also includes the 6th device, is used for obtaining current point
The currently stored amount of the Servers-all of cloth cache cluster.
Further, in the said equipment, described equipment also includes the 7th device, for according to current point
The internal memory quota of the Servers-all of cloth cache cluster and currently stored amount are estimated current distributed slow
The upgrading of the Servers-all depositing cluster expends the time.
Further, in the said equipment, described Unit the two or two, for according to current distributed caching
The respective residue amount of storage of other server in cluster, by internal memory quota maximum in described sequence
All data on server are will migrate into each of this distributed caching cluster in the way of average
On other server.
Further, in the said equipment, described Unit the two or two, for joining internal memory in described sequence
It is distributed that all data on server that volume is maximum move to this successively according to the order of described sequence
On other server each of cache cluster.
Compared with prior art, the application is by by the institute on server maximum for internal memory quota in sequence
There is Data Migration to other server of this distributed caching cluster, by internal memory quota in described sequence
Maximum server is as initial server current to be upgraded, by described current server to be upgraded
Service routine upgrade, by described sequence tightly inferior to the Data Migration on current server to be upgraded
On current server to be upgraded, described sequence is tightly worked as new inferior to current server to be upgraded
Before server to be upgraded, the server of distributed caching cluster can be realized upgrading when, can carry out
The upgrading of transition smoothly, can not lose data, not interrupt outside service in escalation process.
Further, the application emptied by the one hand Servers-all at distributed caching cluster and
Before upgrading, suspend the automatic load balancing function of the Servers-all of distributed caching cluster, in order to avoid
Owing to occurring during subsequent upgrade in cluster that automatic data migration between server causes upgrading mistake
Miss or failure, on the other hand, after the Servers-all of distributed caching cluster empties and upgrades,
Recover the automatic load balancing function of the Servers-all of distributed caching cluster, it is achieved load balancing,
To balance data access pressure follow-up on each server.
Further, the application is by when there being part server to be in number automatically in distributed caching cluster
During according to transition state, other suspending this distributed caching cluster is not at automatic data migration state
The automatic load balancing function of server, and complete at the described server being in automatic data migration state
After becoming automatic data migration, suspend the automatic load balancing merit of this server completing automatic data migration
Can, thus the server realizing being in automatic data migration is gently cut off certainly after completing Autonomic Migration Framework
Dynamic load equalization function, it is to avoid be in automatic data migration state and suspended suddenly automatic load balancing merit
Can and lose the situation of data.
Further, the current total surplus storage of other server in current distributed caching cluster
When amount can not accommodate all data on the server that described internal memory quota is maximum, the application is by working as
After front distributed caching cluster increases new server, the clothes maximum by internal memory quota in described sequence
All Data Migrations on business device, on the server newly increased, realize emptying and rising of server smoothly
Level.
Further, respective according to other server in current distributed caching cluster of the application
Residue amount of storage, all data server on maximum by internal memory quota in described sequence are with close to flat
Equal mode will migrate on other server each of this distributed caching cluster, thus average out as far as possible
The access pressure of data on follow-up each server.
Further, the application is by owning on the server that internal memory quota in described sequence is maximum
Data move to other server each of this distributed caching cluster successively according to the order of described sequence
On, thus only have in ensureing a certain moment between two-server and carrying out Data Migration, as far as possible
Make the machine quantity being in transition state minimum, it is to avoid to cause impact on a large scale, even if migrate simultaneously
Make mistakes, also can scope control on 2 machines, even and if migrate make mistakes, it is also possible to according to work as
Before make mistakes server and described sequence learns that the data on which server have migrated successfully, which
The most unsuccessful, carry out migration from the server made mistakes adjust as long as follow-up, migrated and successfully may be used
Considering.
Accompanying drawing explanation
The detailed description that non-limiting example is made made with reference to the following drawings by reading, this Shen
Other features, objects and advantages please will become more apparent upon:
Fig. 1 illustrates the flow process of the upgrade method of the distributed caching cluster according to one aspect of the application
Figure；
Fig. 2 illustrates the number of the most front server that sorts in the distributed caching cluster of the application one embodiment
According to emptying schematic diagram；
Fig. 3 shows that in the distributed caching cluster of another embodiment of the application, the data of other server empty
Schematic diagram；
Fig. 4 illustrates the upgrade method of the distributed caching cluster according to one preferred embodiment of the application
Flow chart；
Fig. 5 illustrates the upgrade method of the distributed caching cluster according to one preferred embodiment of the application
Flow chart；
Fig. 6 illustrates the upgrade method of the distributed caching cluster according to one preferred embodiment of the application
Flow chart；
Fig. 7 illustrates the updating apparatus according to the application a kind of distributed caching cluster in terms of another
Schematic diagram；
Fig. 8 illustrates the updating apparatus planting distributed caching cluster according to one preferred embodiment of the application
Schematic diagram；
Fig. 9 illustrates that the upgrading planting distributed caching cluster according to another preferred embodiment of the application sets
Standby schematic diagram.
In accompanying drawing, same or analogous reference represents same or analogous parts.
Detailed description of the invention
Below in conjunction with the accompanying drawings the present invention is described in further detail.
In one typical configuration of the application, terminal, the equipment of service network and trusted party all include
One or more processors (CPU), input/output interface, network interface and internal memory.
Internal memory potentially includes the volatile memory in computer-readable medium, random access memory
(RAM) and/or the form such as Nonvolatile memory, such as read only memory (ROM) or flash memory (flash
RAM).Internal memory is the example of computer-readable medium.
Computer-readable medium includes that removable media permanent and non-permanent, removable and non-is permissible
Information storage is realized by any method or technology.Information can be computer-readable instruction, data knot
Structure, the module of program or other data.The example of the storage medium of computer includes, but are not limited to phase
Become internal memory (PRAM), static RAM (SRAM), dynamic random access memory
(DRAM), other kinds of random access memory (RAM), read only memory (ROM), electricity
Erasable Programmable Read Only Memory EPROM (EEPROM), fast flash memory bank or other memory techniques, read-only
Compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage,
Magnetic cassette tape, magnetic disk storage or other magnetic storage apparatus or any other non-transmission medium,
Can be used for the information that storage can be accessed by a computing device.According to defining herein, computer-readable
Medium does not include non-temporary computer readable media (transitorymedia), as modulation data signal and
Carrier wave.
As it is shown in figure 1, the application provides the upgrade method of a kind of distributed caching cluster, including to often
Individual distributed caching cluster is handled as follows:
Step S2, the size according to the internal memory quota of the Servers-all of distributed caching cluster will be all
Server carries out descending sequence, here, the internal memory quota of every station server is used for limiting this service
Memory data output on device；
Step S3, all Data Migrations server on maximum by internal memory quota in described sequence are to being somebody's turn to do
On other server of distributed caching cluster, server maximum for internal memory quota in described sequence is made
For initial server current to be upgraded；Here, in the most described sequence of server of internal memory quota maximum
Coming the most front server, such as the server M1 in Fig. 2, the server in Fig. 2 is according to internal memory quota
Descending it is ordered as M1, M2, M3, M4 ..., the most first all by server M1
Data Migration is to other server M2, M3, M4 of this distributed caching cluster ... on, to empty
Described server M1, it is simple in subsequent step, the service routine on the server M1 emptied is carried out
Upgrading, now, will empty the server M1 after data as initial server current to be upgraded；
Step S4, upgrades the service routine on described current server to be upgraded, by described row
Sequence is tightly inferior on the Data Migration on current server to be upgraded to current server to be upgraded, by described
Sequence tight inferior to current server to be upgraded as new server current to be upgraded, repeat this step S4
Until the service routine on last server completes upgrading in described sequence；Here, often obtain one
Platform empty after i.e. no data server machine after, it is possible to the clothes on the server of this no data
Business program is upgraded, and after i.e. server M1 empties and upgrades, needs successively to other server
M2, M3, M4 ... empty and upgrade, as it is shown on figure 3, this step uses is similar to step
The mode of upgrading after S3 server M1 is emptied, the most respectively to other server M2, M3,
M4 ... first carry out emptying then upgrading, such as, during upgrade server M2, first by server
Data Migration on M2 is on server M1, and due to server M1 be empty and its internal memory quota is big
In the internal memory quota of server M2, so by all data migration server M1 of server M2
Time upper, the memory size of server M1 is large enough to hold all data of server M2, do not have by
On server M1, memory size does not accommodates not the situation of the part data of server M2 and sends out
Raw, after all data on server M2 empty, server M2 can be upgraded, in like manner,
As long as follow-up server M3, M4 according to as shown in Figure 3 ... order i.e. according to internal memory quota by greatly to
Little order is to M3, M4 ... empties one by one and upgrades, wherein, as it is shown on figure 3,
By clothes bigger for the Data Migration on the server M3 on clothes to internal memory quota before server M3 upgrading
On business device M2, then server M3 is upgraded, by the server on clothes before server M4 upgrading
Data Migration on M4 is on the bigger server M3 of internal memory quota, then to server M4 liter
Level, the rest may be inferred.Service on upper last the minimum server of sequence in distributed caching cluster
After program completes upgrading, owing to not having described sequence to be tightly inferior to its server after this server, because of
This without carry out described sequence tightly inferior to the Data Migration on current server to be upgraded to currently treating
On upgrade server, and without carrying out described sequence being tightly inferior to current server to be upgraded as newly
Server current to be upgraded, after last server updating minimum in sequence is complete be one temporarily
Without data cached server.Here, make the machine quantity being in transition state minimum as far as possible,
Avoid causing impact on a large scale.Make mistakes even if migrate simultaneously, also can be scope control at 2 machines
On.The present embodiment can realize the server of distributed caching cluster upgrading when, can carry out steadily
Transition upgrading, escalation process can not be lost data, not interrupt outside service.
As shown in Figure 4, in upgrade method one preferred embodiment of the distributed caching cluster of the application,
Also include step S1 before described step S2, after step S4, also include step S5, this enforcement
The upgrade method of distributed caching cluster, is handled as follows including to each distributed caching cluster:
Step S1, suspends the automatic load balancing function of the Servers-all of distributed caching cluster,
This, described automatic load balancing function is for realizing the Servers-all in a distributed caching cluster
Between automatic data migration, each server in subsequent distribution formula cache cluster rises here
Before Ji, need to suspend the automatic load balancing function of the Servers-all of distributed caching cluster, with
Exempt from the automatic data migration owing to occurring in cluster between server during subsequent upgrade and cause upgrading
Mistake or failure, it addition, all can revise the right of the server that data are distributed after Data Migration every time
According to table, it is simple to after Data Migration, when there is data access according to the synopsis constantly revised remain to from
This data are found on the corresponding server after migration；
Step S2, the size according to the internal memory quota of the Servers-all of distributed caching cluster will be all
Server carries out descending sequence, here, the internal memory quota of every station server is used for limiting this service
Memory data output on device；
Step S3, all Data Migrations server on maximum by internal memory quota in described sequence are to being somebody's turn to do
On other server of distributed caching cluster, server maximum for internal memory quota in described sequence is made
For initial server current to be upgraded；Here, in the most described sequence of server of internal memory quota maximum
Coming the most front server, such as the server M1 in Fig. 2, the server in Fig. 2 is according to internal memory quota
Descending it is ordered as M1, M2, M3, M4 ..., the most first all by server M1
Data Migration is to other server M2, M3, M4 of this distributed caching cluster ... on, to empty
Described server M1, it is simple in subsequent step, the service routine on the server M1 emptied is carried out
Upgrading, now, will empty the server M1 after data as initial server current to be upgraded；
Step S4, upgrades the service routine on described current server to be upgraded, by described row
Sequence is tightly inferior on the Data Migration on current server to be upgraded to current server to be upgraded, by described
Sequence tight inferior to current server to be upgraded as new server current to be upgraded, repeat this step S4
Until the service routine on last server completes upgrading in described sequence；Here, often obtain one
Platform empty after i.e. no data server machine after, it is possible to the clothes on the server of this no data
Business program is upgraded, and after i.e. server M1 empties and upgrades, needs successively to other server
M2, M3, M4 ... empty and upgrade, as it is shown on figure 3, this step uses is similar to step
The mode of upgrading after S3 server M1 is emptied, the most respectively to other server M2, M3,
M4 ... first carry out emptying then upgrading, such as, during upgrade server M2, first by server
Data Migration on M2 is on server M1, and due to server M1 be empty and its internal memory quota is big
In the internal memory quota of server M2, so by all data migration server M1 of server M2
Time upper, the memory size of server M1 is large enough to hold all data of server M2, do not have by
On server M1, memory size does not accommodates not the situation of the part data of server M2 and sends out
Raw, after all data on server M2 empty, server M2 can be upgraded, in like manner,
As long as follow-up server M3, M4 according to as shown in Figure 3 ... order i.e. according to internal memory quota by greatly to
Little order is to M3, M4 ... empties one by one and upgrades, wherein, as it is shown on figure 3,
By clothes bigger for the Data Migration on the server M3 on clothes to internal memory quota before server M3 upgrading
On business device M2, then server M3 is upgraded, by the server on clothes before server M4 upgrading
Data Migration on M4 is on the bigger server M3 of internal memory quota, then to server M4 liter
Level, the rest may be inferred.Service on upper last the minimum server of sequence in distributed caching cluster
After program completes upgrading, owing to not having described sequence to be tightly inferior to its server after this server, because of
This without carry out described sequence tightly inferior to the Data Migration on current server to be upgraded to currently treating
On upgrade server, and without carrying out described sequence being tightly inferior to current server to be upgraded as newly
Server current to be upgraded, after last server updating minimum in sequence is complete be one temporarily
Without data cached server.Here, make the machine quantity being in transition state minimum as far as possible,
Avoid causing impact on a large scale.Make mistakes even if migrate simultaneously, also can be scope control at 2 machines
On；
Step S5, recovers the automatic load balancing function of the Servers-all of distributed caching cluster, from
And be on empty server by the Data Migration on other server to current data, i.e. move to internal memory
On the upper minimum server of quota sequence, it is achieved load balancing, to balance number follow-up on each server
According to accessing pressure.On the one hand the present embodiment empties at the Servers-all of distributed caching cluster and upgrades
Before, suspend the automatic load balancing function of the Servers-all of distributed caching cluster, in order to avoid due to
Automatic data migration between server occurs during subsequent upgrade in cluster and cause staging error or
Failure, on the other hand, after the Servers-all of distributed caching cluster empties and upgrades, recovers
The automatic load balancing function of the Servers-all of distributed caching cluster, it is achieved load balancing, with flat
Weigh data access pressure follow-up on each server.
As it is shown in figure 5, in upgrade method one preferred embodiment of the distributed caching cluster of the application,
The step that the service routine on described current server to be upgraded carries out upgrading is included in step S4:
Step S41, is updated to new service journey by the service routine on described current server to be upgraded
Sequence；
Step S42, restarts the service routine after updating, and waits the service after described renewal
Program is restarted normally.Here, wait that the service routine after described renewal just continues rear after restarting normally
Next server continuous is upgraded, thus before ensureing next server updating, and this server can normal work
Make, i.e. can normally receive the data come from next server migration and service is normally provided.
In upgrade method one preferred embodiment of the distributed caching cluster of the application, step S1, temporarily
The automatic load balancing function of the Servers-all stopping distributed caching cluster includes:
When the Servers-all of distributed caching cluster is not at automatic data migration state, suspend institute
There is the automatic load balancing function of service；
When there being part server to be in automatic data migration state in distributed caching cluster, suspending should
The automatic load of other server being not at automatic data migration state of distributed caching cluster is equal
Weighing apparatus function, and after the described server being in automatic data migration state completes automatic data migration,
Suspend the automatic load balancing function of this server completing automatic data migration, thus realize being in certainly
The server of dynamic Data Migration gently cuts off automatic load balancing function after completing Autonomic Migration Framework, keeps away
Exempt to be in automatic data migration state to be suspended suddenly automatic load balancing function and lose the feelings of data
Condition.
As shown in Figure 6, in upgrade method one preferred embodiment of the distributed caching cluster of the application,
Step S3, all Data Migrations server on maximum by internal memory quota in described sequence are to this distribution
On other server of formula cache cluster, using server maximum for internal memory quota in described sequence as just
The server current to be upgraded begun, including:
Step S31, it is judged that the current total surplus of other server in current distributed caching cluster is deposited
Whether reserves can accommodate all data on the server that described internal memory quota is maximum, here, permissible
Believed by the monitoring of the current total surplus amount of storage of other server in current distributed caching cluster
Breath judges that the Servers-all in current distributed caching cluster is the most all nearly at full data mode,
If it is, that this situation is not suitable for Upgrading Now, need current distributed caching cluster increases newly
Empty server；
If so, step S32, all data server on maximum by internal memory quota in described sequence
Step S34 is forwarded to after moving to respectively on other server each of this distributed caching cluster；
If it is not, step S33, after increasing new server in current distributed caching cluster, by institute
State all Data Migrations on the server that internal memory quota in sequence is maximum to the server newly increased,
The size of the internal memory quota of the current Servers-all according to distributed caching cluster is by Servers-all
Forwarding step S34 to after re-starting descending sequence, wherein, the internal memory of described new server is joined
Volume is less than the internal memory quota of server maximum in described sequence；Here, due to current distributed caching
Cluster adds new server, thus follow-up other server is upgraded before need currently
Servers-all in distributed caching cluster re-starts sequence by the size of internal memory quota, in order to after
Data on continuous each server sequentially migrate；
Step S34, rises server maximum for internal memory quota in described sequence as initial currently waiting
Level server, here, owing to the internal memory quota of described new server is less than maximum in described sequence
The internal memory quota of server, so the server that in before ranking replacement and after updating, internal memory quota is maximum begins
It is same server eventually, such as the server M1 in Fig. 2 and Fig. 3.The present embodiment is in current distribution
Servers-all in formula cache cluster is all nearly at full data mode, it is impossible to carry out the clear of server
Empty with when upgrading, by increasing new server in current distributed caching cluster, realize clothes smoothly
Business the emptying and upgrade of device.
In upgrade method one preferred embodiment of the distributed caching cluster of the application, step S31 it
Before also include:
Obtain the currently stored amount of the Servers-all of current distributed caching cluster；Based on this, follow-up
Step S31 can be according to the clothes maximum except internal memory quota in described sequence of current distributed caching cluster
The current residual amount of storage of other server each outside business device determines the current of other server described
Total surplus amount of storage.
In upgrade method one preferred embodiment of the distributed caching cluster of the application, obtain current point
After the step of the currently stored amount of the Servers-all of cloth cache cluster, also include:
Internal memory quota and the currently stored amount of the Servers-all according to current distributed caching cluster are estimated
The upgrading of the Servers-all counting current distributed caching cluster expends the time, thus more reasonably arranges
With the upgrading mode of Servers-all and the time adjusting current distributed caching cluster.
In upgrade method one preferred embodiment of the distributed caching cluster of the application, step S32,
It is distributed that all data server on maximum by internal memory quota in described sequence move to this respectively
Step on other server each of cache cluster includes:
According to the respective residue amount of storage of other server in current distributed caching cluster, by institute
State all data on the server that in sequence, internal memory quota is maximum will migrate in the way of average
On other server each of this distributed caching cluster, thus average out on follow-up each server as far as possible
The access pressure of data.
In upgrade method one preferred embodiment of the distributed caching cluster of the application, step S32,
It is distributed that all data server on maximum by internal memory quota in described sequence move to this respectively
Step on other server each of cache cluster includes,
All data server on maximum by internal memory quota in described sequence are according to described sequence
Order moves on other server each of this distributed caching cluster successively, i.e. as in figure 2 it is shown, first
After the part Data Migration on server M1 to server M2, then on just server M1
Part Data Migration is to server M3, and the rest may be inferred, until by last the remaining on server M1
Remainder evidence moves to sort on last server, thus only two services in ensureing a certain moment
Carrying out Data Migration between device, making the machine quantity being in transition state minimum as far as possible, it is to avoid
Cause on a large scale impact, make mistakes even if migrate simultaneously, also can scope control on 2 machines,
Make mistakes even if migrating, it is also possible to make mistakes server and described sequence learns which takes according to current
Data on business device have migrated successfully, and which is the most unsuccessful, enters from the server made mistakes as long as follow-up
Row migration adjusts, and has migrated and successfully can consider, such as, if as in figure 2 it is shown,
Server M1 makes mistakes migrating to server M4 when, then migrated successfully
Server M2 and M3 then need not consider, as long as subsequent migration is from the beginning of server M4, it is considered to
Server M4 and sequence server behind.
As it is shown in fig. 7, according to the another side of the application, the application also provides for a kind of distributed caching collection
The updating apparatus 100 of group, including:
First device 1, big for the internal memory quota according to the Servers-all of distributed caching cluster
Little Servers-all is carried out descending sequence；Here, the internal memory quota of every station server is used for limiting
Memory data output on this server fixed；
Second device 2, all data on the server that internal memory quota in described sequence is maximum
Move on other server of this distributed caching cluster, by internal memory quota maximum in described sequence
Server is as initial server current to be upgraded；Here, the server i.e. institute that internal memory quota is maximum
Stating and come the most front server in sequence, such as the server M1 in Fig. 2, the server in Fig. 2 depends on
Internal memory quota is descending is ordered as M1, M2, M3, M4 ..., the most first by server M1
On all Data Migrations to other server M2, M3, M4 of this distributed caching cluster ... on,
To empty described server M1, it is simple to the service journey on the server M1 emptied in subsequent step
Sequence is upgraded, and now, will empty the server M1 after data as initial clothes current to be upgraded
Business device；
3rd device 3, upgrades the service routine on described current server to be upgraded, by institute
Stating sequence tight inferior on the Data Migration on current server to be upgraded to current server to be upgraded, inciting somebody to action
Described sequence tightly inferior to current server to be upgraded as new server current to be upgraded, repeat
3rd device 3, until the service routine on last server completes upgrading in described sequence.?
This, often obtain one empty after i.e. no data server machine after, it is possible to this no data
Service routine on server is upgraded, and after i.e. server M1 empties and upgrades, needs successively
To other server M2, M3, M4 ... empty and upgrade, as it is shown on figure 3, the 3rd device
3 use the mode of upgrading after the emptying server M1 of the second device 2 that is similar to, the most right
Other server M2, M3, M4 ... first carry out emptying then upgrading, such as, upgrade service
During device M2, first by the Data Migration on server M2 to server M1, due to server M1
The internal memory quota of server M2 it is more than for empty and its internal memory quota, so owning server M2
Time on data migration server M1, the memory size of server M1 is large enough to hold server M2's
All data, do not have owing on server M1, memory size does not accommodates server M2 not
Part data situation occur, after all data on server M2 empty, can be to server
M2 upgrades, in like manner, if follow-up server M3, M4 according to as shown in Figure 3 ... suitable
Sequence i.e. according to the descending order of internal memory quota to M3, M4 ... empty one by one and upgrade,
Wherein, the Data Migration on the front server M3 that will take as it is shown on figure 3, server M3 upgrades
To the server M2 that internal memory quota is bigger, then server M3 is upgraded, server M4 liter
By on the Data Migration on the server M4 on clothes to the bigger server M3 of internal memory quota before level,
Then upgrading server M4, the rest may be inferred.In distributed caching cluster, the upper minimum of sequence is last
After service routine on one server completes upgrading, owing to not having described sequence tight after this server
Inferior to its server, therefore without carrying out described sequence tightly inferior on current server to be upgraded
Data Migration is on current server to be upgraded, and without carrying out described sequence tightly inferior to currently treating
Upgrade server is as new server current to be upgraded, last server liter minimum in sequence
Level complete after be one temporarily without data cached server.Here, make as far as possible to be in transition state
Machine quantity minimum, it is to avoid cause on a large scale impact.Make mistakes even if migrate simultaneously, also can be model
System of containing is on 2 machines.The present embodiment can realize the server of distributed caching cluster in upgrading
Time, stable transition upgrading can be carried out, escalation process can not be lost data, do not interrupt outside
Service.
In updating apparatus one preferred embodiment of the distributed caching cluster of the application, described 3rd dress
Put 3, for the service routine on described current server to be upgraded is updated to new service routine,
Service routine after updating is restarted, and waits that the service routine after described renewal is restarted normally.
Here, wait that the service routine after described renewal just continues to enter next server follow-up after restarting normally
Row upgrading, thus before ensureing next server updating, this server can normally work, i.e. can be normal
Receive the data come from next server migration and service is normally provided.
As shown in Figure 8, in updating apparatus one preferred embodiment of the distributed caching cluster of the application,
Described equipment 100 also includes the 4th device 4, for suspending the Servers-all of distributed caching cluster
Automatic load balancing function；Here, described automatic load balancing function be used for realizing one distributed
The automatic data migration between Servers-all in cache cluster, here at subsequent distribution formula cache set
Before each server in Qun is upgraded, need to suspend the Servers-all of distributed caching cluster
Automatic load balancing function, in order to avoid due to occur during subsequent upgrade in cluster between server from
Move Data Migration and cause staging error or failure, it addition, number all can be revised after each Data Migration
, when there is data access according to not in the synopsis according to the server being distributed, it is simple to after Data Migration
The synopsis of disconnected amendment remains to find this data the server after corresponding migration；
Described equipment 100 also includes the 5th device 5, for recovering all clothes of distributed caching cluster
The automatic load balancing function of business device, thus by the Data Migration on other server to current data be
On empty server, i.e. move on the upper minimum server of internal memory quota sequence, it is achieved load balancing,
To balance data access pressure follow-up on each server.The present embodiment is on the one hand at distributed caching collection
Group Servers-all empty and upgrade before, suspend distributed caching cluster Servers-all from
Dynamic load equalization function, in order to avoid owing to automatically counting between server in cluster occurring during subsequent upgrade
Staging error or failure is caused, on the other hand, in all services of distributed caching cluster according to migration
After device empties and upgrades, recover the automatic load balancing merit of the Servers-all of distributed caching cluster
Can, it is achieved load balancing, to balance data access pressure follow-up on each server.
In updating apparatus one preferred embodiment of the distributed caching cluster of the application, described 4th dress
Put 4, be used for when the Servers-all of distributed caching cluster is not at automatic data migration state,
Suspend the automatic load balancing function of all services；
When there being part server to be in automatic data migration state in distributed caching cluster, suspending should
The automatic load of other server being not at automatic data migration state of distributed caching cluster is equal
Weighing apparatus function, and after the described server being in automatic data migration state completes automatic data migration,
Suspend the automatic load balancing function of this server completing automatic data migration, thus realize being in certainly
The server of dynamic Data Migration gently cuts off automatic load balancing function after completing Autonomic Migration Framework, keeps away
Exempt to be in automatic data migration state to be suspended suddenly automatic load balancing function and lose the feelings of data
Condition.
As it is shown in figure 9, in updating apparatus one preferred embodiment of the distributed caching cluster of the application,
Described second device 2 includes:
2nd 1 unit 21, for judging the current of other server in current distributed caching cluster
Whether total surplus amount of storage can accommodate all data on the server that described internal memory quota is maximum, if
It is to forward the two or two unit 22 to, if it is not, then forward the two or three unit 23 to；Here, can pass through
The monitoring information of the current total surplus amount of storage of other server in current distributed caching cluster is sentenced
Servers-all in disconnected current distributed caching cluster is the most all nearly at full data mode, if
It is that this situation is not suitable for Upgrading Now, needs current distributed caching cluster increases new sky
Server；
Two or two unit 22, all numbers on the server that internal memory quota in described sequence is maximum
The two or four unit 24 is forwarded to according to after moving to respectively on other server each of this distributed caching cluster；
Two or three unit 23, after increasing new server in current distributed caching cluster, will
The all Data Migrations on server that in described sequence, internal memory quota is maximum are to the server newly increased
On, according to the size of the internal memory quota of the current Servers-all of distributed caching cluster by all services
Device forwards the two or four unit 24, wherein, described new server to after re-starting descending sequence
Internal memory quota is less than the internal memory quota of server maximum in described sequence；Here, due to current distribution
Formula cache cluster adds new server, thus follow-up other server is upgraded before need
Servers-all in current distributed caching cluster is re-started sequence by the size of internal memory quota,
So that the data on follow-up each server sequentially migrate；
Two or four unit 24, for the server that internal memory quota in described sequence is maximum as initial
Current server to be upgraded.Here, owing to the internal memory quota of described new server is less than described sequence
The internal memory quota of the server of middle maximum, so internal memory quota is maximum in before ranking replacement and after updating
Server is same server all the time, such as the server M1 in Fig. 2 and Fig. 3.The present embodiment exists
Servers-all in current distributed caching cluster is all nearly at full data mode, it is impossible to take
When emptying and upgrade of device of business, by increasing new server in current distributed caching cluster, suitable
Profit realizes emptying and upgrading of server.
In updating apparatus one preferred embodiment of the distributed caching cluster of the application, described equipment is also
Including the 6th device 6, for obtaining the currently stored of the Servers-all of current distributed caching cluster
Amount.Based on this, follow-up 2nd 1 unit 21 can according to current distributed caching cluster except described
The current residual amount of storage of other server each outside the server that in sequence, internal memory quota is maximum is true
The current total surplus amount of storage of fixed other server described.
In updating apparatus one preferred embodiment of the distributed caching cluster of the application, described equipment is also
Including the 7th device 7, for the internal memory quota of the Servers-all according to current distributed caching cluster
Upgrading with the Servers-all that currently stored amount estimates current distributed caching cluster expends the time, from
And more reasonably arrange and adjust current distributed caching cluster Servers-all upgrading mode and
Time.
In updating apparatus one preferred embodiment of the distributed caching cluster of the application, the described 2nd 2
Unit 22, for the respective residue storage according to other server in current distributed caching cluster
Amount, by all data on server maximum for internal memory quota in described sequence by close in the way of average
Will migrate on other server each of this distributed caching cluster, thus average out follow-up each clothes as far as possible
The access pressure of data on business device.
In updating apparatus one preferred embodiment of the distributed caching cluster of the application, the described 2nd 2
Unit 22, all data on the server that internal memory quota in described sequence is maximum are according to described
The order of sequence moves on other server each of this distributed caching cluster, i.e. such as Fig. 2 institute successively
Show, first by after the part Data Migration on server M1 to server M2, more just server
Part Data Migration on M1 is to server M3, and the rest may be inferred, until by server M1
After remaining data move to sequence on last server, thus ensure in a certain moment only two
Carrying out Data Migration between station server, making to be in the machine quantity of transition state as far as possible
Few, it is to avoid to cause impact on a large scale, make mistakes even if migrate simultaneously, also can be scope control at 2
On machine, even and if migrate and make mistakes, it is also possible to according to the current server and described sort of makeing mistakes
Know that the data on which server have migrated successfully, which is the most unsuccessful, if follow-up from makeing mistakes
Server carries out migration and adjusts, and has migrated and successfully can consider, such as, such as Fig. 2 institute
Show, if server M1 makes mistakes migrating to server M4 when, then the most
Migrate successful server M2 and M3 then need not consider, as long as subsequent migration is opened from server M4
Begin, it is considered to server M4 and sequence server behind.
In sum, the application is by by all data on server maximum for internal memory quota in sequence
Move on other server of this distributed caching cluster, by internal memory quota maximum in described sequence
Server is as initial server current to be upgraded, by the service on described current server to be upgraded
Program is upgraded, and the Data Migration described sequence being tightly inferior on current server to be upgraded is to current
On server to be upgraded, described sequence is tightly risen as new currently waiting inferior to current server to be upgraded
Level server, can realize the server of distributed caching cluster upgrading when, can carry out smoothly
Transition is upgraded, and can not lose data, do not interrupts outside service in escalation process.
Further, the application emptied by the one hand Servers-all at distributed caching cluster and
Before upgrading, suspend the automatic load balancing function of the Servers-all of distributed caching cluster, in order to avoid
Owing to occurring during subsequent upgrade in cluster that automatic data migration between server causes upgrading mistake
Miss or failure, on the other hand, after the Servers-all of distributed caching cluster empties and upgrades,
Recover the automatic load balancing function of the Servers-all of distributed caching cluster, it is achieved load balancing,
To balance data access pressure follow-up on each server.
Further, the application is by when there being part server to be in number automatically in distributed caching cluster
During according to transition state, other suspending this distributed caching cluster is not at automatic data migration state
The automatic load balancing function of server, and complete at the described server being in automatic data migration state
After becoming automatic data migration, suspend the automatic load balancing merit of this server completing automatic data migration
Can, thus the server realizing being in automatic data migration is gently cut off certainly after completing Autonomic Migration Framework
Dynamic load equalization function, it is to avoid be in automatic data migration state and suspended suddenly automatic load balancing merit
Can and lose the situation of data.
Further, the current total surplus storage of other server in current distributed caching cluster
When amount can not accommodate all data on the server that described internal memory quota is maximum, the application is by working as
After front distributed caching cluster increases new server, the clothes maximum by internal memory quota in described sequence
All Data Migrations on business device, on the server newly increased, realize emptying and rising of server smoothly
Level.
Further, respective according to other server in current distributed caching cluster of the application
Residue amount of storage, all data server on maximum by internal memory quota in described sequence are with close to flat
Equal mode will migrate on other server each of this distributed caching cluster, thus average out as far as possible
The access pressure of data on follow-up each server.
Further, the application is by owning on the server that internal memory quota in described sequence is maximum
Data move to other server each of this distributed caching cluster successively according to the order of described sequence
On, thus only have in ensureing a certain moment between two-server and carrying out Data Migration, as far as possible
Make the machine quantity being in transition state minimum, it is to avoid to cause impact on a large scale, even if migrate simultaneously
Make mistakes, also can scope control on 2 machines, even and if migrate make mistakes, it is also possible to according to work as
Before make mistakes server and described sequence learns that the data on which server have migrated successfully, which
The most unsuccessful, carry out migration from the server made mistakes adjust as long as follow-up, migrated and successfully may be used
Considering.
Obviously, those skilled in the art the application can be carried out various change and modification without deviating from
Spirit and scope.So, if these amendments of the application and modification belong to the application power
Profit requires and within the scope of equivalent technologies, then the application is also intended to comprise these changes and modification exists
In.
It should be noted that the present invention can be carried out in the assembly of hardware at software and/or software, example
As, special IC (ASIC), general purpose computer or any other similar hardware can be used to set
For realizing.In one embodiment, the software program of the present invention can perform to realize by processor
Steps described above or function.Similarly, the software program (including the data structure being correlated with) of the present invention
Can be stored in computer readable recording medium storing program for performing, such as, RAM memory, magnetically or optically driver or
Floppy disc and similar devices.It addition, some steps of the present invention or function can employ hardware to realize, example
As, perform the circuit of each step or function as coordinating with processor.
It addition, the part of the present invention can be applied to computer program, such as computer program
Instruction, when it is computer-executed, by the operation of this computer, can call or provide basis
The method of the present invention and/or technical scheme.And call the programmed instruction of the method for the present invention, may be deposited
Store up fixing or movably in record medium, and/or by broadcast or other signal bearing medias
Data stream and be transmitted, and/or be stored in the computer equipment that runs according to described programmed instruction
In working storage.Here, include a device according to one embodiment of present invention, this device bag
Include the memorizer for storing computer program instructions and for performing the processor of programmed instruction, its
In, when this computer program instructions is performed by this processor, trigger this plant running based on aforementioned
The method of multiple embodiments and/or technical scheme according to the present invention.
It is obvious to a person skilled in the art that the invention is not restricted to the thin of above-mentioned one exemplary embodiment
Joint, and without departing from the spirit or essential characteristics of the present invention, it is possible to concrete with other
Form realizes the present invention.Therefore, no matter from the point of view of which point, embodiment all should be regarded as exemplary
, and be nonrestrictive, the scope of the present invention is limited by claims rather than described above
It is fixed, it is intended that all changes fallen in the implication of equivalency and scope of claim are included
In the present invention.Any reference in claim should not be considered as limit involved right want
Ask.Furthermore, it is to be understood that " an including " word is not excluded for other unit or step, odd number is not excluded for plural number.Dress
Multiple unit or the device of putting statement in claim can also be passed through software by a unit or device
Or hardware realizes.The first, the second word such as grade is used for representing title, and is not offered as any specific
Order.
Claims (18)
1. a upgrade method for distributed caching cluster, wherein, including:
The size of the internal memory quota of the Servers-all according to distributed caching cluster is by Servers-all
Carry out descending sequence；
All Data Migrations server on maximum by internal memory quota in described sequence are distributed to this
On other server of cache cluster, using server maximum for internal memory quota in described sequence as initially
Server current to be upgraded；
Service routine on described current server to be upgraded is upgraded, described sequence is tightly inferior to
Data Migration on current server to be upgraded is on current server to be upgraded, by the most secondary for described sequence
In current server to be upgraded as new server current to be upgraded, repeat this step until described row
In sequence, the service routine on last server completes upgrading.
The most the method for claim 1, wherein by the clothes on described current server to be upgraded
Business program is upgraded, including:
Service routine on described current server to be upgraded is updated to new service routine；
Service routine after updating is restarted, and waits that the service routine after described renewal is just restarted
Often.
3. method as claimed in claim 1 or 2, wherein, owning according to distributed caching cluster
Before Servers-all is carried out descending sequence by the size of the internal memory quota of server, also include:
Suspend the automatic load balancing function of the Servers-all of distributed caching cluster；
After in described sequence, the service routine on last server completes upgrading, also include:
Recover the automatic load balancing function of the Servers-all of distributed caching cluster.
4. method as claimed in claim 3, wherein, suspends all services of distributed caching cluster
The automatic load balancing function of device, including:
When the Servers-all of distributed caching cluster is not at automatic data migration state, suspend institute
There is the automatic load balancing function of service；
When there being part server to be in automatic data migration state in distributed caching cluster, suspending should
The automatic load of other server being not at automatic data migration state of distributed caching cluster is equal
Weighing apparatus function, and after the described server being in automatic data migration state completes automatic data migration,
Suspend the automatic load balancing function of this server completing automatic data migration.
5. the method as described in any one of Claims 1-4, wherein, joins internal memory in described sequence
The all Data Migrations on server that volume is maximum on other server of this distributed caching cluster,
Including:
Whether judge the current total surplus amount of storage of other server in current distributed caching cluster
The all data on the server that described internal memory quota is maximum can be accommodated；
If so, all data on server maximum for internal memory quota in described sequence are moved to respectively
On other server each of this distributed caching cluster；
If it is not, after increasing new server in current distributed caching cluster, in described sequence
Deposit all Data Migrations on the server that quota is maximum to the server newly increased, according to distributed
The size of the internal memory quota of the current Servers-all of cache cluster Servers-all is re-started by
Big to little sequence, wherein, the internal memory quota of described new server is less than clothes maximum in described sequence
The internal memory quota of business device.
6. method as claimed in claim 5, wherein, it is judged that currently its in distributed caching cluster
Whether the current total surplus amount of storage of its server can accommodate the server that described internal memory quota is maximum
On all data before, also include:
Obtain the currently stored amount of the Servers-all of current distributed caching cluster.
7. method as claimed in claim 6, wherein, obtains all of current distributed caching cluster
After the currently stored amount of server, also include:
Internal memory quota and the currently stored amount of the Servers-all according to current distributed caching cluster are estimated
The upgrading of the Servers-all counting current distributed caching cluster expends the time.
8. the method as described in any one of claim 5 to 7, wherein, joins internal memory in described sequence
The all data on server that volume is maximum move to other clothes each of this distributed caching cluster respectively
Include on business device:
According to the respective residue amount of storage of other server in current distributed caching cluster, by institute
State all data on the server that in sequence, internal memory quota is maximum will migrate in the way of average
On other server each of this distributed caching cluster.
9. the method as described in any one of claim 5 to 8, wherein, joins internal memory in described sequence
The all data on server that volume is maximum move to other clothes each of this distributed caching cluster respectively
Include on business device:
All data server on maximum by internal memory quota in described sequence are according to described sequence
Order moves on other server each of this distributed caching cluster successively.
10. a updating apparatus for distributed caching cluster, wherein, including:
First device, for the size of the internal memory quota of the Servers-all according to distributed caching cluster
Servers-all is carried out descending sequence；
Second device, all data on the server that internal memory quota in described sequence is maximum are moved
Move on on other server of this distributed caching cluster, the clothes maximum by internal memory quota in described sequence
Business device is as initial server current to be upgraded；
3rd device, for the service routine on described current server to be upgraded is upgraded, will
Described sequence tightly inferior on the Data Migration on current server to be upgraded to current server to be upgraded,
Using described sequence tightly inferior to current server to be upgraded as new server current to be upgraded, repeat to hold
This device of row, until the service routine on last server completes upgrading in described sequence.
11. equipment as claimed in claim 10, wherein, described 3rd device, for working as described
Service routine on front server to be upgraded is updated to new service routine, to the service routine after updating
Restart, and wait that the service routine after described renewal is restarted normally.
12. equipment as described in claim 9 or 10, wherein, described equipment also includes the 4th device,
For suspending the automatic load balancing function of the Servers-all of distributed caching cluster；
Described equipment also includes the 5th device, for recovering the Servers-all of distributed caching cluster
Automatic load balancing function.
13. equipment as claimed in claim 12, wherein, described 4th device, for when distributed
When the Servers-all of cache cluster is not at automatic data migration state, suspend the automatic of all services
Load-balancing function；
When there being part server to be in automatic data migration state in distributed caching cluster, suspending should
The automatic load of other server being not at automatic data migration state of distributed caching cluster is equal
Weighing apparatus function, and after the described server being in automatic data migration state completes automatic data migration,
Suspend the automatic load balancing function of this server completing automatic data migration.
14. equipment as described in any one of claim 10 to 13, wherein, described second device includes:
Unit the 2nd 1, for judging the current total of other server in current distributed caching cluster
Whether residue amount of storage can accommodate all data on the server that described internal memory quota is maximum, if
It is to forward Unit the two or two to, if it is not, then forward Unit the two or three to；
Unit the two or two, all data on the server that internal memory quota in described sequence is maximum
Unit the two or four is forwarded to after moving to respectively on other server each of this distributed caching cluster；
Unit the two or three, after increasing new server in current distributed caching cluster, by institute
State all Data Migrations on the server that internal memory quota in sequence is maximum to the server newly increased,
The size of the internal memory quota of the current Servers-all according to distributed caching cluster is by Servers-all
Unit the two or four, wherein, the internal memory of described new server is forwarded to after re-starting descending sequence
Quota is less than the internal memory quota of server maximum in described sequence；
Unit the two or four, works as initial for the server that internal memory quota in described sequence is maximum
Front server to be upgraded.
15. equipment as claimed in claim 14, wherein, described equipment also includes the 6th device, uses
Currently stored amount in the Servers-all obtaining current distributed caching cluster.
16. equipment as claimed in claim 15, wherein, described equipment also includes the 7th device, uses
Internal memory quota and currently stored amount in the Servers-all according to current distributed caching cluster are estimated
The upgrading of the Servers-all of current distributed caching cluster expends the time.
17. equipment as described in any one of claim 14 to 16, wherein, described Unit the two or two,
For the respective residue amount of storage according to other server in current distributed caching cluster, by institute
State all data on the server that in sequence, internal memory quota is maximum will migrate in the way of average
On other server each of this distributed caching cluster.
18. equipment as described in any one of claim 14 to 17, wherein, described Unit the two or two,
All data on the server that internal memory quota in described sequence is maximum are according to described sequence
Order moves on other server each of this distributed caching cluster successively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510216024.5A CN106201561B (en) | 2015-04-30 | 2015-04-30 | The upgrade method and equipment of distributed caching cluster |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510216024.5A CN106201561B (en) | 2015-04-30 | 2015-04-30 | The upgrade method and equipment of distributed caching cluster |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106201561A true CN106201561A (en) | 2016-12-07 |
| CN106201561B CN106201561B (en) | 2019-08-23 |
Family
ID=57458469
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510216024.5A Active CN106201561B (en) | 2015-04-30 | 2015-04-30 | The upgrade method and equipment of distributed caching cluster |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106201561B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107071069A (en) * | 2017-06-14 | 2017-08-18 | 北京凤凰理理它信息技术有限公司 | The method for cutting data traffic transparent to user, device, storage medium and electronic equipment |
| CN109213507A (en) * | 2018-08-27 | 2019-01-15 | 郑州云海信息技术有限公司 | A kind of upgrade method and server |
| WO2019137539A1 (en) * | 2018-01-15 | 2019-07-18 | 华为技术有限公司 | Software upgrade method and system |
| CN110505269A (en) * | 2019-06-21 | 2019-11-26 | 广州虎牙科技有限公司 | Transaction processing system, method for processing business and server |
| CN110519363A (en) * | 2019-08-23 | 2019-11-29 | 苏州浪潮智能科技有限公司 | A kind of memory device set group and its management method |
| CN110651256A (en) * | 2017-06-09 | 2020-01-03 | 思杰系统有限公司 | System and method for preventing service interruption during software update |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739263A (en) * | 2008-11-11 | 2010-06-16 | 英业达股份有限公司 | Method and device for realizing operating system update in multi-computer cluster system |
| CN102591668A (en) * | 2011-01-05 | 2012-07-18 | 阿里巴巴集团控股有限公司 | Device, method and system for updating elastic computing cloud system |
| JP2012177990A (en) * | 2011-02-25 | 2012-09-13 | Nippon Telegr & Teleph Corp <Ntt> | Service provision system and operation method of the same |
| CN103516731A (en) * | 2012-06-15 | 2014-01-15 | 华为技术有限公司 | Cache server service method, cache server, and system |
| CN104426976A (en) * | 2013-09-03 | 2015-03-18 | 方正宽带网络服务股份有限公司 | Online updating device and online updating method for virtual server cluster |
-
2015
- 2015-04-30 CN CN201510216024.5A patent/CN106201561B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739263A (en) * | 2008-11-11 | 2010-06-16 | 英业达股份有限公司 | Method and device for realizing operating system update in multi-computer cluster system |
| CN102591668A (en) * | 2011-01-05 | 2012-07-18 | 阿里巴巴集团控股有限公司 | Device, method and system for updating elastic computing cloud system |
| JP2012177990A (en) * | 2011-02-25 | 2012-09-13 | Nippon Telegr & Teleph Corp <Ntt> | Service provision system and operation method of the same |
| CN103516731A (en) * | 2012-06-15 | 2014-01-15 | 华为技术有限公司 | Cache server service method, cache server, and system |
| CN104426976A (en) * | 2013-09-03 | 2015-03-18 | 方正宽带网络服务股份有限公司 | Online updating device and online updating method for virtual server cluster |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110651256A (en) * | 2017-06-09 | 2020-01-03 | 思杰系统有限公司 | System and method for preventing service interruption during software update |
| CN107071069A (en) * | 2017-06-14 | 2017-08-18 | 北京凤凰理理它信息技术有限公司 | The method for cutting data traffic transparent to user, device, storage medium and electronic equipment |
| CN107071069B (en) * | 2017-06-14 | 2020-08-25 | 北京凤凰理理它信息技术有限公司 | Method and device for transparently switching data traffic to user, storage medium and electronic equipment |
| WO2019137539A1 (en) * | 2018-01-15 | 2019-07-18 | 华为技术有限公司 | Software upgrade method and system |
| US11381453B2 (en) | 2018-01-15 | 2022-07-05 | Huawei Technologies Co., Ltd. | Method and system for service software upgrade in an evolved packet core (EPC) system |
| CN109213507A (en) * | 2018-08-27 | 2019-01-15 | 郑州云海信息技术有限公司 | A kind of upgrade method and server |
| CN110505269A (en) * | 2019-06-21 | 2019-11-26 | 广州虎牙科技有限公司 | Transaction processing system, method for processing business and server |
| CN110519363A (en) * | 2019-08-23 | 2019-11-29 | 苏州浪潮智能科技有限公司 | A kind of memory device set group and its management method |
| CN110519363B (en) * | 2019-08-23 | 2022-03-22 | 苏州浪潮智能科技有限公司 | Storage device cluster and management method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106201561B (en) | 2019-08-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106201561A (en) | The upgrade method of distributed caching cluster and equipment | |
| CN108683516B (en) | Application instance upgrading method, device and system | |
| EP2288975B1 (en) | Method for optimizing cleaning of maps in flashcopy cascades containing incremental maps | |
| EP3186760B1 (en) | Dynamic load-based merging | |
| CN106339386B (en) | Database flexible scheduling method and device | |
| CN110609743A (en) | Method, electronic device and computer program product for configuring resources | |
| CN104539708A (en) | Capacity reduction method, device and system for cloud platform resources | |
| TW201832072A (en) | Resource reclamation method and apparatus | |
| CN110347651A (en) | Method of data synchronization, device, equipment and storage medium based on cloud storage | |
| CN106302632A (en) | The method for down loading of a kind of foundation image and management node | |
| CN103246549B (en) | A kind of method and system of data conversion storage | |
| CN103631894A (en) | Dynamic copy management method based on HDFS | |
| CN108491163A (en) | A kind of big data processing method, device and storage medium | |
| CN109510852A (en) | The method and device of gray scale publication | |
| CN113867957A (en) | Method and device for realizing elastic expansion of number of cross-cluster containers | |
| CN105933154A (en) | Management method of cloud calculation resources | |
| CN108875035A (en) | The date storage method and relevant device of distributed file system | |
| CN108390914A (en) | A kind of service update method and device, system | |
| EP4170491A1 (en) | Resource scheduling method and apparatus, electronic device, and computer-readable storage medium | |
| CN108234231A (en) | The method, apparatus and electronic equipment of elastic telescopic arrangement in cloud computing environment | |
| CN112860412B (en) | Service data processing method and device, electronic equipment and storage medium | |
| CN104517067A (en) | Method, device and system for data access | |
| CN109788013B (en) | Method, device and equipment for distributing operation resources in distributed system | |
| CN104239520B (en) | A kind of HDFS data block Placement Strategies based on historical information | |
| CN109522177A (en) | A kind of task daily record processing system, method and device |
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 |