Brief overview
The Perfstat is a NetApp storage diagnostic data Command-line collection tool that collects detailed information about NetApp storage, including configuration information and performance data, with two versions for NetApp storage OS 2 modes:
Storage mode |
Perfstat version |
7 mode |
Perfstat 7 |
Cluster Mode |
Perfstat 8 |
Perfstat 7.x collects information from 7 mode NetApp storage, Perfstat 8.x is primarily used to collect cluster mode NetApp storage information
650) this.width=650; "title=" lamp "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" lamp "src=" http:// S3.51cto.com/wyfs02/m02/8f/76/wkiom1jfhweqnycoaaahh4b_tzg193.png "width=" "height="/> Note:
After NetApp Data Ontap 8.3, 7 mode exits the historical stage.
As a command-line tool, Perfstat is still more complex to use, and for ease of use, NetApp has made the tool a GUI shell that is easy for users to use. The following is an example of Windows to illustrate how perfstat is used.
Download run
This GUI tool can be run on Windows and Linux and can be downloaded from the NetApp website (requires a NetApp account)
After the download there will be a. zip compression package, such as Gperfstat_windows.zip. Unzip a directory, note that the file path does not contain spaces, or the program does not run correctly .
Double-click Run filegPerfstat.exe。
Data collection
After the program runs, the following interface appears, if your perfstat version is not up-to-date, you can enter the NetApp website account Update software, otherwise "skip" skip the page
650) this.width=650; "alt=" user-added image "src=" https://kb.netapp.com/support/servlet/rtaImage?eid= Ka31a00000012fd&feoid=00n1a000006nnav&refid=0em1a0000006blk "/>
Select "Perfstat V8" to enter the collection parameters setting interface
650) this.width=650; "alt=" user-added image "src=" https://kb.netapp.com/support/servlet/rtaImage?eid= KA31A00000012FD&FEOID=00N1A000006NNAV&REFID=0EM1A0000006BLP "/>
Select "Cluster-mode" if collection Cluster mode is stored in "select Mode"
The cluster management address is stored under Cluster mode, and the nodes participating in the cluster have separate addresses that can optionally be collected from the cluster management address or node address. For example, select collect information for the entire cluster, disaster controller ip/name (s) fills in the storage cluster management address, and if it is a dual live cluster, tick "Metrocluster". Here you can fill in multiple cluster management addresses, separated by commas. If you fill in the hostname, make sure that DNS resolution is normal.
Complete the cluster management user name and password at Username/password. Once completed, the backend invokes OpenSSH to create the SSH key pair and upload it to the storage
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m00/8f/76/wkiom1jfhwmjro4eaajma3yklhc367.png "width=" 580 "height=" 709 "/>
Complete the cluster management user name and password at Username/password. Once filled out, the backend invokes OpenSSH to create an SSH key pair and upload it to the store, which establishes the local host and the stored SSH trust.
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m02/8f/76/wkiom1jfhwmqddhsaaa_oljiv4i790.png "width=" 397 "height=" 104 "/>
In Perfstat time, you can define when the information is collected, and by default select "6min (3x2)" Interation:3 time:2min. That is, the Perfstat command iterates through the collection 3 times, each running 2 minutes.
If you want to customize the time interval you can select "Customize perfstat ..." in the drop-down menu and select the number of runs and intervals, such as running 20 times for 2 minutes at a time. The execution time for this name will exceed: 20x2=40min
In addition, by specifying the value of runs through the "Incremental" field, you can set the number of times the Perfstat command executes, and if the value is set to 10, the entire information collection time: 40min*10=400min.
If you want to run at a specified time or run periodically, in the "Execute" field, schedule means run at a specified time, now means run immediately, once means run once at a specified time, daily means running every day at a specified time.
When the "Preset file name" drop-down menu is selected, the default is generally reserved. Select "Critical_preset" unless the perfstat takes too long to execute. Under this configuration, Perfstat will reduce the range of information collected to complete the collection as soon as possible.
If you need to pass the Perfstat command to a parameter that cannot be defined by another GUI during the collection process, add the parameter to the "Generate CMD" field.
650) this.width=650; "alt=" user-added image "src=" https://kb.netapp.com/support/servlet/rtaImage?eid= Ka31a00000012fd&feoid=00n1a000006nnav&refid=0em1a0000006bmo "height="/>
Click on the "Generate CMD" button and the above configured parameters will generate the command line.
Click "Start" button to run, the following dialog box prompts the command has started to run.
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m01/8f/74/wkiol1jfhwqdsje6aaa2lwe4mjy938.png "width=" 426 "height=" 145 "/>
Click "OK" to close the prompt window and wait for the operation to complete. A prompt window appears when you are finished running.
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m00/8f/76/wkiom1jfhwrwp6uwaaa3e-h0vmy246.png "width=" 429 "height=" 146 "/>
The running results are saved under the installation directory "Cli_logs".
If you need to terminate the program halfway, in the active process area, select the process number under the processes ID, and then click Abort. The corresponding run result output file can also be opened by the "Logs" button.
650) this.width=650; "alt=" user-added image "src=" https://kb.netapp.com/support/servlet/rtaImage?eid= Ka31a00000012fd&feoid=00n1a000006nnav&refid=0em1a0000006bmy "/>
Data upload
The data collected can be uploaded to NetApp's website https://latx.netapp.com, which requires an account to log in to the site.
The site can analyze and produce charts for uploading data. These charts make it easy to analyze the health of existing storage.
Upload steps:
1. After logging in to Https://latx.netapp.com, see the following Web page
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m00/8f/75/wkiol1jfq9mho30faach9jdpyag198.png "width=" 696 "height=" 145 "/>
2. Click "New perfstat" and the upload window appears.
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0 px;padding-top:0px;padding-left:0px;margin:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image " Src= "Http://s3.51cto.com/wyfs02/M01/8F/78/wKiom1jfQ9rALUW0AAC9hHIxRkA165.png" width= "698" height= "342"/>
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; float:none;padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" Http://s3.51cto.com/wyfs02/M00/8F/78/wKiom1jfQ9uwiVI_AABHOe0po8I419.png "width=" 551 "height=" 139 "/>
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m02/8f/78/wkiom1jfq9ys4zv0aagj0akztgy819.png "width=" 1001 "height=" 365 "/>
Data analysis
Use the LATX chart to analyze the running state of the store.
1.IO delay
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m00/8f/75/wkiol1jfq-hg-eouaapbstvfjt4901.png "width=" 773 "height=" 503 "/>
2. Disk Utilization
Overall disk utilization can be seen under the summary menu
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m01/8f/78/wkiom1jfq-lgjkvtaacotz0x-ae858.png "Width=" 701 "height=" 94 "/>
For example, the current disk utilization is 100%, which indicates that the disk is very busy and the storage responsiveness is definitely poor. So the delay is relatively high.
However, in this view, it is not possible to accurately reflect the utilization of the disk, this 100% is only the most efficient portion of the disk data. is not the average of all disks. This is because the graph we see is generated based on the results of the "sysstat" command output. The disk utilization for this command output is the highest value in all disks. Assuming this is the case, the storage node has 30 disks, only one disk is very busy, the usage reaches 100%, and the other 29 blocks are only 10%. So, the disk utilization of this command output is also 100%.
For accurate disk utilization, in the Summary view, tap disk utilization
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m02/8f/78/wkiom1jfq-pi6rjiaabcytbsimg749.png "width=" 589 "height=" 177 "/>
Then, the exact disk utilization view, which attempts to generate based on the output of the command "Statit", can be used for each disk and is labeled in shades of color depending on how busy the disk is, and the more busy the disk, the darker the color of the logo.
Under the Configure drop-down menu, select Show Grid Lines to move the slider to the far right at the grid Szie. For example, each row of a small square represents a disk, the mouse is placed on a small square, the lower right corner of the status bar will show the current 0b.01.20 disk utilization of 98%
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m00/8f/78/wkiom1jfq-tcq9ewaahxeiva4ui070.png "width=" "height=" 808 "/>
Utilization of 3.CPU
CPU utilization can also be viewed under the summary view, as shown in the data collection, where the maximum CPU utilization is 60% and the average utilization is 15%. Similar to disk utilization, because the CPU has multiple cores, the maximum CPU utilization shown here is also the utilization of the highest-utilization cores.
650) this.width=650; "title=" image "style=" border-right-width:0px;background-image:none;border-bottom-width:0px; padding-top:0px;padding-left:0px;padding-right:0px;border-top-width:0px; "border=" 0 "alt=" image "src=" http:// S3.51cto.com/wyfs02/m01/8f/78/wkiom1jfq-aglxruaaqqqe2anii294.png "width=" 1125 "height=" 346 "/>
This article is from the "Chuan Liu Information" blog, please be sure to keep this source http://tech4fei.blog.51cto.com/12552348/1912374
Collect NetApp Storage diagnostic information using the Perfstat tool