CN106020727B - A kind of method of Smart Rack performance tuning - Google Patents
A kind of method of Smart Rack performance tuning Download PDFInfo
- Publication number
- CN106020727B CN106020727B CN201610351749.XA CN201610351749A CN106020727B CN 106020727 B CN106020727 B CN 106020727B CN 201610351749 A CN201610351749 A CN 201610351749A CN 106020727 B CN106020727 B CN 106020727B
- Authority
- CN
- China
- Prior art keywords
- disk
- smart rack
- performance
- raid
- subregion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 16
- 230000004044 response Effects 0.000 claims abstract description 13
- 238000005457 optimization Methods 0.000 claims abstract description 10
- 238000005192 partition Methods 0.000 claims description 10
- 101100297422 Schizosaccharomyces pombe (strain 972 / ATCC 24843) phd1 gene Proteins 0.000 claims description 4
- 238000003491 array Methods 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 claims description 2
- 230000002708 enhancing effect Effects 0.000 claims 1
- 230000001737 promoting effect Effects 0.000 abstract description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000011800 void material Substances 0.000 description 1
- 238000013316 zoning Methods 0.000 description 1
Classifications
-
- 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
- G06F3/061—Improving I/O performance
- G06F3/0613—Improving I/O performance in relation to throughput
-
- 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
- G06F3/062—Securing storage systems
- G06F3/0622—Securing storage systems in relation to access
-
- 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
- G06F3/0638—Organizing or formatting or addressing of data
- G06F3/0644—Management of space entities, e.g. partitions, extents, pools
-
- 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
- G06F3/0671—In-line storage system
- G06F3/0683—Plurality of storage devices
- G06F3/0689—Disk arrays, e.g. RAID, JBOD
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Techniques For Improving Reliability Of Storages (AREA)
Abstract
The invention discloses a kind of method of Smart Rack performance tuning, this method is carried out by three aspects: 1) tuning Smart Rack performance by optimization disk；Disk optimization subregion: access frequency is high, on system performance influence relatively large subregion be placed in disk by outer portion, the catalogue that content is often changed is placed on individual subregion, using several main catalogues of Smart Rack system as individual file system, respectively distributes an area；2) Smart Rack performance is tuned using software PLC；By compiling support of the new kernel addition to RAID, compiles and raidtools software package is installed；3) magnetic disc i/o is tuned using elvtune to tune Smart Rack performance；By the scheduling operation of magnetic disc i/o, to control response time and the handling capacity of magnetic disc i/o.The present invention promotes the handling capacity of Smart Rack by promoting disk read-write speed, pass through the scheduling operation of magnetic disc i/o, reach the disk service life and handling capacity is coordinated, while the Information Security of Smart Rack can be greatly promoted by establishing RAID mechanism, and its cost is almost nil.
Description
Technical field
The present invention relates to the field Smart Rack, specifically a kind of method of Smart Rack performance tuning.
Background technique
Smart Rack is unified to be powered, radiate to the server node that it is carried using standard cabinet as designing unit
And management, support calculating and the memory node of variform, the storage and processing towards mass data are particularly suitable for cloud meter
The applications such as calculation, big data processing.
The performance of Smart Rack to carry out it is comprehensive consider, often solve a performance bottleneck, it is possible that separately
One performance bottleneck, the final purpose for thus surveying the tuning of Smart Rack performance is: making the items of system in a certain range
Resource, which uses, to be tended to be reasonable and keeps exactly being exactly that system resource has reached one when certain balance, i.e. system operational excellence
When a equilibrium state.And in the measurement of entire index, excessive use of any one resource can all destroy this equilibrium-like
State responds slow or load too high so as to cause system.
Therefore Smart Rack system performance tuning system is a complicated and time-consuming process, needs to comprehensively consider locating for system
Environment, the loaded business of system and service feature.Respectively in application program, operating system, server hardware, network environment etc.
Aspect is searched and is positioned, and it is two aspects of application program and operating system that it is maximum, which to influence performance, because of the two aspects
The problem is not noticeable, and concealment is very strong.
The characteristics of internet big data is that data access request is more frequent, and individual data amount is according to business difference, number
It is serious according to big fractionlet, and its performance of disk is tested, I/O throughput has harsher requirement to Smart Rack, passes through
The performance of Smart Rack is tuned, it can be with better services in big data business.
Summary of the invention
Technical assignment of the invention is to provide a kind of method of Smart Rack performance tuning.
Technical assignment of the invention realizes that this method is carried out by three aspects in the following manner:
1) Smart Rack performance is tuned by optimization disk；
Disk optimization subregion: access frequency is high, relatively large subregion is influenced on system performance and is placed in leaning on for disk
Outer portion, the catalogue that content is often changed are placed on individual subregion, and several main catalogues of Smart Rack system are made
For individual file system, an area is respectively distributed；
2) Smart Rack performance is tuned using software PLC；
By compiling support of the new kernel addition to RAID, compiles and raidtools software package is installed；
3) magnetic disc i/o is tuned using elvtune to tune Smart Rack performance；
By the scheduling operation of magnetic disc i/o, to control response time and the handling capacity of magnetic disc i/o.
When the described optimization disk, if system has muti-piece hard disk, multiple exchange partitions are used；One is divided on each disk
A exchange partition, so that Smart Rack system can be used concurrently.
The rank of the RAID is divided into RAID0, RAID1, RAID3；
RAID0 is data segmentation, and data block is alternately write disk, obtains best readwrite performance, but do not provide number
According to redundancy；
RAID1 is disk mirroring, all write in disk 1 are also write in disk 2, can be read from any one disk
It takes；
RAID3 also specifies a disc driver to store parity information other than data are divided.
A kind of method of Smart Rack performance tuning of the invention is compared to the prior art, fast by promoting disk read-write
It spends to promote the handling capacity of Smart Rack, sets up targetedly zoning schemes for different business mode, or pass through disk
The scheduling operation of I/O reaches the disk service life and handling capacity is coordinated, lead to simultaneously to control response time and the handling capacity of magnetic disc i/o
It crosses and establishes RAID mechanism and can greatly promote the Information Security of Smart Rack, and its cost is almost nil.
Specific embodiment
Embodiment 1:
The method of Smart Rack performance tuning is carried out by three aspects:
1) Smart Rack performance is tuned by optimization disk；
Disk optimization subregion: access frequency is high, relatively large subregion is influenced on system performance and is placed in leaning on for disk
Outer portion, the catalogue that content is often changed are placed on individual subregion, and several main catalogues of Smart Rack system are made
For individual file system, an area is respectively distributed；
If system has muti-piece hard disk, it may be considered that use multiple exchange partitions.An exchange point is divided on each disk
Area.By the way that pri option is arranged in/etc/fstab file, it can make multiple exchange partitions that there is same priority.Smart
Rack system can concurrently use them.The performance of exchange partition can be improved in this way.Certainly, if Installed System Memory is enough
Greatly, so much memory and system much has more than is needed, what void that, which also just takes less than, is deposited.It can consider to remove when subregion
Exchange partition.But as a Smart Rack, even if memory is sufficiently large, exchange partition still should be set.
2) Smart Rack performance is tuned using software PLC；
RAID(redundant arrays of inexpensive disks) it is to enhance disk performance and reliability by distributing data across on multiple disks
Technology.RAID is realized by software.RAID's has many ranks, and different ranks has different requirements, phase to hardware
It can should also obtain different Performance And Reliabilities.RAID0 is data segmentation, and data block is alternately write disk, it is available most
Good readwrite performance, but data redundancy is not provided.RAID1 is disk mirroring, all write in disk 1 also write disk 2
In, it can be read from any one disk.RAID3 also specifies a disc driver to store other than data are divided
Parity information.
By compiling support of the new kernel addition to RAID, compiles and raidtools software package is installed；raidtools
It is the user class tool that can initialize, start, stop and control RAID array.Setting subregion is at least needed using RAID0
Two subregions, they are located at different disks.Two subregions for being preferably RAID0 have identical size.When creating the division, it will divide
Area's type is appointed as " FD ".So that interior nuclear energy identifies them as RAID subregion, in this way, these subregions can draw every time
It is detected and starts automatically when leading.If without marking RAID subregion in this way, after each guidance, it is necessary to first
First input " raidstart -- all " could carry RAID array.
Editor/etc/raidtab file, creation/etc/raidtab file is to describe the configuration information of RAID array.It is interior
Core is detected automatically in guidance using these information and starts RAID array, it is therefore necessary to each RAID array of creation into
This configuration of row.In raidtab file, " raiddev " entry indicates the RAID array to be created；"nr-raid-
Number of disks of the disks " in specified arrays；" persistent-superblock " is set as 1, tells how is raid tool
When create the volume；The block size that " chunk-size " specifies RAID0 to use with K byte；Finally, what is specified is that composition should
Equipment, that is, subregion of array.
Mkraid and creation file system use order " # mkraid/dev/md0 " order initialization/dev/md0
Subregion, while md0 RAID0 array will be started.File system is wanted to obtain next, can create on md0.
3) magnetic disc i/o is tuned using elvtune to tune Smart Rack performance；
By the scheduling operation of magnetic disc i/o, to control response time and the handling capacity of magnetic disc i/o.By adjusting I/O request
Maximum latency in the queue can tune between response time and handling capacity.If it is required that the less response time,
So handling capacity will reduce, conversely, then available biggish handling capacity of longer response time.Tool can be used "/
Sbin/elvtune " changes maximum response time value.Application method is as follows:
Check current setting,
# /sbin/elvtune /dev/hda1
Wherein-r parameter is directed to read operation, and-w parameter is directed to write operation.
Pass through the average information of ordering " iostat-d-x/dev/hda1 " to obtain (including average request size and average
Queue length) monitor the effect of above I/O configuration, and configuration is adjusted, to obtain optimal performance.In general, for reading
It writes frequently, but the Smart Rack that the data volume operated is less, and higher to requirement of real-time, then parameter can be turned down.
If instead for reading and writing infrequently, but requiring the Smart Rack of biggish handling capacity, parameter can be tuned up, to obtain
Obtain biggish handling capacity.
The technical personnel in the technical field can readily realize the present invention with the above specific embodiments,.But it answers
Work as understanding, the present invention is not limited to above-mentioned several specific embodiments.On the basis of the disclosed embodiments, the technology
The technical staff in field can arbitrarily combine different technical features, to realize different technical solutions.
Claims (2)
1. a kind of method of Smart Rack performance tuning, which is characterized in that this method is carried out by three aspects:
1) Smart Rack performance is tuned by optimization disk；
Disk optimization subregion: access frequency is high, on system performance influence relatively large subregion be placed in disk by outside
Point, the catalogue that content is often changed is placed on individual subregion, using several main catalogues of Smart Rack system as single
Only file system respectively distributes an area；
By the way that pri option is arranged in/etc/fstab file, make multiple exchange partitions that there is same priority；Smart
Rack system can be used concurrently；
2) Smart Rack performance is tuned using software PLC；
RAID is the technology for enhancing disk performance and reliability by distributing data across on multiple disks；By software come real
Existing RAID；RAID's has many ranks, and different ranks has different requirements to hardware, accordingly can also obtain different performances
And reliability；RAID0 is data segmentation, and data block is alternately write disk, obtains best readwrite performance, but do not provide number
According to redundancy；RAID1 is disk mirroring, all write in disk 1 are also write in disk 2, can be read from any one disk
It takes；RAID3 also specifies a disc driver to store parity information other than data are divided；
By compiling support of the new kernel addition to RAID, compiles and raidtools software package is installed；Raidtools is can
With initialization, starting, stopping and the user class tool for controlling RAID array；Setting subregion at least needs two using RAID0
Subregion, two subregions are located at different disks；When creating the division, divisional type is appointed as " FD "；So that interior nuclear energy by they
It is identified as RAID subregion, these subregions are detected automatically and start in each guidance；If without marking in this way
RAID subregion, then after each guidance, it is necessary to which input " raidstart -- all " could carry RAID array first；
Editor/etc/raidtab file, creation/etc/raidtab file is to describe the configuration information of RAID array；Kernel makes
RAID array is detected and started automatically in guidance with these information, and this configuration is carried out to each RAID array of creation；
In raidtab file, " raiddev " entry indicates the RAID array to be created；" nr-raid-disks " specified arrays
In number of disks；The block size that " chunk-size " specifies RAID0 to use with K byte；Finally, what is specified is composition
Equipment, that is, subregion of the array；
Mkraid and creation file system are using order " # mkraid/dev/md0 " order initialization/dev/md0 subregion, together
When will start md0 RAID0 array；File system is wanted to obtain next, can create on md0；
3) magnetic disc i/o is tuned using elvtune to tune Smart Rack performance；
By the scheduling operation of magnetic disc i/o, to control response time and the handling capacity of magnetic disc i/o；It requests by adjusting I/O in team
Maximum latency in column tunes between response time and handling capacity；If it is required that the less response time, handles up
Amount will reduce, conversely, then handling capacity will increase the longer response time；Change maximum using tool "/sbin/elvtune "
Response time value；Application method is as follows:
Check current setting,
# /sbin/elvtune /dev/hda1
Wherein-r parameter is directed to read operation, and-w parameter is directed to write operation；
The effect of the above I/O configuration is monitored by the average information for ordering " iostat-d-x/dev/hda1 " to obtain, and
Adjustment configuration, to obtain optimal performance；The less Smart Rack of data volume frequent for reading and writing, but operating, and to reality
When property is more demanding, then parameter can be turned down；If instead for reading and writing infrequently, but requiring biggish handling capacity
Smart Rack, parameter can be tuned up, to obtain biggish handling capacity.
2. a kind of method of Smart Rack performance tuning according to claim 1, which is characterized in that the optimization magnetic
When disk, if system has muti-piece hard disk, multiple exchange partitions are used；An exchange partition is divided on each disk, so that Smart
Rack system can be used concurrently.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610351749.XA CN106020727B (en) | 2016-05-25 | 2016-05-25 | A kind of method of Smart Rack performance tuning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610351749.XA CN106020727B (en) | 2016-05-25 | 2016-05-25 | A kind of method of Smart Rack performance tuning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106020727A CN106020727A (en) | 2016-10-12 |
| CN106020727B true CN106020727B (en) | 2019-03-15 |
Family
ID=57094900
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610351749.XA Active CN106020727B (en) | 2016-05-25 | 2016-05-25 | A kind of method of Smart Rack performance tuning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106020727B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103106050A (en) * | 2013-02-22 | 2013-05-15 | 浪潮电子信息产业股份有限公司 | Method for achieving layered storage and copy of data of storage system |
| CN103246478A (en) * | 2012-02-08 | 2013-08-14 | 北京同有飞骥科技股份有限公司 | Disk array system supporting grouping-free overall situation hot standby disks based on flexible redundant array of independent disks (RAID) |
| US9195401B2 (en) * | 2014-02-18 | 2015-11-24 | University Of Florida Research Foundation, Inc. | Method and apparatus for virtual machine live storage migration in heterogeneous storage environment |
| CN105549910A (en) * | 2015-12-14 | 2016-05-04 | 浪潮(北京)电子信息产业有限公司 | IO scheduling method and apparatus |
-
2016
- 2016-05-25 CN CN201610351749.XA patent/CN106020727B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103246478A (en) * | 2012-02-08 | 2013-08-14 | 北京同有飞骥科技股份有限公司 | Disk array system supporting grouping-free overall situation hot standby disks based on flexible redundant array of independent disks (RAID) |
| CN103106050A (en) * | 2013-02-22 | 2013-05-15 | 浪潮电子信息产业股份有限公司 | Method for achieving layered storage and copy of data of storage system |
| US9195401B2 (en) * | 2014-02-18 | 2015-11-24 | University Of Florida Research Foundation, Inc. | Method and apparatus for virtual machine live storage migration in heterogeneous storage environment |
| CN105549910A (en) * | 2015-12-14 | 2016-05-04 | 浪潮(北京)电子信息产业有限公司 | IO scheduling method and apparatus |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106020727A (en) | 2016-10-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7478177B2 (en) | System and method for automatic reassignment of shared storage on blade replacement | |
| EP2766803B1 (en) | A method and system for consolidating a plurality of heterogeneous storage systems in a data center | |
| US7953929B1 (en) | Expanding the storage capacity of a virtualized data storage system | |
| JP4856515B2 (en) | Methods, systems, and products for rebalancing striped disk data (striped disk data rebalancing) | |
| KR100490723B1 (en) | Apparatus and method for file-level striping | |
| US9612758B1 (en) | Performing a pre-warm-up procedure via intelligently forecasting as to when a host computer will access certain host data | |
| US20080120462A1 (en) | System And Method For Flexible Physical-Logical Mapping Raid Arrays | |
| CN101510145B (en) | Storage system management method and apparatus | |
| JP2004013547A (en) | Data allocation method and information processing system | |
| US20140229695A1 (en) | Systems and methods for backup in scale-out storage clusters | |
| US20170199694A1 (en) | Systems and methods for dynamic storage allocation among storage servers | |
| US9747040B1 (en) | Method and system for machine learning for write command selection based on technology feedback | |
| CN104050248A (en) | File storage system and storage method | |
| US8156281B1 (en) | Data storage system and method using storage profiles to define and modify storage pools | |
| US20120303893A1 (en) | Writing of data of a first block size in a raid array that stores and mirrors data in a second block size | |
| US20150081969A1 (en) | Storage apparatus and control method thereof, and recording medium | |
| JP5884602B2 (en) | Storage control device and storage system | |
| US9069471B2 (en) | Passing hint of page allocation of thin provisioning with multiple virtual volumes fit to parallel data access | |
| US10242053B2 (en) | Computer and data read method | |
| CN101997919A (en) | Storage resource management method and device | |
| CN107329798A (en) | Method, device and the virtualization system of data duplication | |
| CN106020727B (en) | A kind of method of Smart Rack performance tuning | |
| US9870152B2 (en) | Management system and management method for managing data units constituting schemas of a database | |
| KR101754713B1 (en) | Asymmetric distributed file system, apparatus and method for distribution of computation | |
| US10768844B2 (en) | Internal striping inside a single 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 |