Unreasonable memory configuration can result in frequent system crashes, which are difficult to reproduce, complicating troubleshooting.
The built-in System O&M plugin in FineDataLink provides a Health Inspection function that helps you check whether the system environments are reasonably configured and assists you in optimizing the configuration to better prevent system crashes.
It supports both manual and periodic automated checks of the system environment configuration.
It enables proactive notification of unreasonable configuration.
It provides actionable optimization suggestions.
This document introduces the inspection logic and items of Health Inspection.
The System O&M plugin of versions before V1.3.1 only supports environment configuration inspections.
Starting from V1.3.1, the System O&M plugin supports the following inspection items: system configuration, system operation status, port availability, network connectivity, service connectivity, business configuration, environment configuration, disk condition, information of non-containerized deployment, and virtual machine configuration. The following table describes specific content.
For details, see the "System Inspection" section of O&M Monitoring Guide.
System Inspection
The inspection result only includes exception information and provides no configuration recommendations.
The inspection items include operating system information, glibc/GCC version, hostname, time zone, CPU clock speed, core quantity, memory size, used memory size, available memory size, swap, inspection time, disk space of the root directory, and free disk space of the root directory.
System Operation Status
Plugin information: plugin name, version, and whether the plugin is enabled or not
Registration authorization information: whether the project is registered or not
Container version
Abnormal project folder size: the size of each folder in the WEB-INF folder
Temporary output directory: the path and free space in the directory
Project startup path: directory path
Whether there is a web.xml file in the WEB-INF folder
Project JAR package information: versions of various modules in the project and JAR package dates
Port Inspection
Downtime inspection port: port used by the auto downtime handling tool
Cluster port: ports of nodes in the cluster
WebSocket port: The WebSocket port is mainly used for token refresh, notifying user session termination, platform messaging, memory/CPU display, displaying online user count in platform logs, and synchronizing the data connection edit status.
Network Inspection
The system performs ping tests to confirm connectivity between the file server, the external database, the state server, and the application and to detect packet loss.
Service Connectivity Inspection
The system checks whether the project is connected to the Redis component, the SFTP component, and the FineDB database.
Business Inspection
finedb: The system inspects the database type of FineDB. You are advised to configure an external database.
Encoding/character set of the external database: the database encoding/character set of the external FineDB database.
Log level: You are advised to click the xxx icon under System Management > Intelligent O&M > Platform Log and set System Log Level to ERROR.
Log cleanup: You are advised to click the xxx icon under System Management > Intelligent O&M > Platform Log, enable Auto Cleanup, and set Clearance Frequency to Every 3 Months Auto Clear Once.
Maximum disk space occupied by automatic backup records: You are advised to click the xxx icon under System Management > Intelligent O&M > Backup & Restoration, and set the maximum disk space occupied by automatic backup records.
Maximum number of automatic backup records: You are advised to click the xxx icon under System Management > Intelligent O&M > Backup & Restoration, and set the maximum number of automatic backup records to less than 5 records.
Second-level cache enabling status: You are advised to configure the second-level cache for the system.
Load alert enabling: You are advised to enable Intelligent Alert under System Management > Intelligent O&M > Load Management > Load Surveillance.
Cluster parameter configuration: You are advised to configure Internal Forwarding Parameter under System Management > Intelligent O&M > Cluster Configuration > Parameter Configuration.
Maximum cell quantity during quick analysis: It is recommended that the value of DistributedOptimizationConfig.spiderConfig.spider_fast_compute_limit_unit should not exceed 10000000.
Maximum row quantity during quick analysis: It is recommended that the value of DistributedOptimizationConfig.spiderConfig.spider_fast_compute_limit_row should not exceed 10000000.
Maximum updated cell quantity during quick analysis: It is recommended that the value of DistributedOptimizationConfig.spiderConfig.spider_update_fast_compute_limit_cell should not exceed 10000.
Maximum memory size during quick analysis: You are advised to set the parameter value of DistributedOptimizationConfig.spiderConfig.spider_fast_compute_limit_memory to less than 500000000.
Maximum memory size of spark calculation result: You are advised to set the parameter value of DistributedOptimizationConfig.spiderConfig.spark_driver_maxResultSize to less than 10000000.
Environment Inspection
vm.max_map_count: You are advised to set the value of vm.max_map_count to 262144.
Maximum opened files: It is recommended that the value of open_files be at least 65536.
overcommit_memory: You are advised to set the value of vm.overcommit_memory to 0.
overcommit_ratio: You are advised to set the value of vm.overcommit_ratio to 50.
glibc memory leak configuration: You are advised to configure export MALLOC_ARENA_MAX=1.
Disk Inspection
The inspection items include disk space, used disk space, free disk space, sequential read, sequential write, random read, small file creation, and small file deletion.
Non-Containerized Deployment Information
The inspection items include the process user and the permission on the server's main folder.
Java VM Inspection
Physical memory: It outputs information only.
On-heap memory: It outputs information only.
Off-heap memory occupied by FineIO reading: You are advised to set the off-heap memory occupied by FineIO reading to 2 GB for the JVM.
Off-heap memory occupied by FineIO writing: You are advised to set the off-heap memory occupied by FineIO writing to 2 GB for the JVM.
Headless mode: You are advised to configure -Djava.awt.headless=true to enable the headless mode.
Dump export: You are advised to configure HeapDumpOnOutOfMemoryError and HeapDumpPath to ensure a normal dump export.
DisableExplicitGC Parameter: You are advised not to configure DisableExplicitGC Parameter, which may disable System.gc(), reducing system stability.
JDK version: You are advised to use JDK8 of V1.8.0_181 or later versions.
Stack space: It is recommended that the stack space per thread be at most 1024 KB. (Namely, the -Xss parameter value should be at most 1024).
Garbage collector type: You are advised to set the garbage collector type to ParallelScavenge for the JVM.
Ratio of old generation memory to young generation memory: You are advised to set the ratio of old generation to young generation memory in a heap to 2 (namely, -XX:NewRatio=2).
recompilationCutoff-related parameters: You are advised to set the values of recompilationCutoff-related parameters to -1.
Debug mode: You are advised to disable the debug mode and delete the -agentlib:jdwp and -Xrunjdwp parameters.
Xms configuration: You are advised to set Xms to the same size as Xmx.
Zip-related configuration -Dsun.zip.disableMemoryMapping: You are advised to configure -Dsun.zip.disableMemoryMapping=true.
Language configuration -Duser.country: You are advised to configure -Duser.country=EN.
Language configuration -Duser.language: You are advised to configure -Duser.language=en.
Code cache configuration: You are advised to configure -XX:ReservedCodeCacheSize=250m.
Installation package deployment: The inspection result only includes information.
Stack information: The inspection result only includes information.
Abnormal items are displayed on the page after inspection.
You can use the Quick Fix function to set item values to the system-recommended values with one click.
The Quick Fix function is unavailable in the following environments.
1. Projects deployed in non-Tomcat environments, such as IBM WebSphere Application Server and Oracle WebLogic Server
2. Projects with the custom setenv file
3. Tomcat projects that are started as services
4. Tomcat projects started with an exe file in a Windows environment
5. Environments where the system physical memory is less than or equal to 16 GB
6. Cluster projects
If the current project environment meets the requirements for Quick Fix and there are inspection items that need one-click configuration, Quick Fix will be displayed in the top left corner of the Health Inspection page.
Click Quick Fix, select the inspection items that need configuration, and click Apply Configuration to configure the values of the abnormal items with one click.
1. Inspection items supporting Quick Fix include on-heap memory, off-heap NIO memory, off-heap FineIO memory, garbage collector type, ratio of old-generation memory to young-generation memory, stack space, and Xms configuration.
2. If you have no file read and write permission or are unable to modify the configuration files, a prompt will pop up: "Failed to apply configurations as no read/write permission was granted to the configuration file." Click OK to return to the Health Inspection page.
After successful configuration, a prompt will pop up: “The configuration is applied successfully and takes effect after server restart.” The configuration of all exceptional items is corrected after a restart.
An inspection report is generated after each health inspection. You can download the report to local or preview it online, as shown in the following figure.
Inspection Report supports the viewing of abnormal inspection items or all abnormal items. The following figure shows an example.
The system configuration is obtained every hour after a system restart.
When conducting the first health inspection, you need to click Start Inspection. The system automatically performs the inspection, obtains the most recent system configuration, and updates the inspection results.
When conducting the health inspection again, you need to click Re-diagnosis. The system automatically performs the inspection, obtains the latest system configuration, and updates the inspection results.
You (the admin) can enable System Auto Inspection and click Save to save the settings on the Health Inspection Handling tab page.
After the function is enabled, the system automatically inspects whether any unreasonable configuration exists at 11 AM every Sunday. If so, you can be notified through three forms, namely SMS Reminder, Platform Message, and Email Notification.
1. SMS Reminder: You need to enable the Use SMS Platform function. For details, see SMS.
2. Email Notification: You need to configure the mailbox server. For details, see Mailbox.
3. Platform Message: WebSocket port needs to be configured and opened at the load balancing level. For details, see Websocket Port is occupied.
O&M Interface Key is the interface key bound to the current project in FineOps. For details, see Function Introduction.
滑鼠選中內容,快速回饋問題
滑鼠選中存在疑惑的內容,即可快速回饋問題,我們將會跟進處理。
不再提示
10s後關閉
Submitted successfully
Network busy