How to enable performance counter in sql server a step by step guide for sql performance monitoring and tuning 2026

How to enable performance counter in sql server a step by step guide — this article walks you through enabling SQL Server performance counters, why they matter, and how to monitor them effectively. Quick fact: performance counters help you track resource usage and diagnose bottlenecks, from CPU and memory to I/O and query waits. Below is a practical, step-by-step guide you can follow, plus tips, stats, and handy references.

• Quick start checklist:
  • Confirm you have admin rights on the server
  • Identify the counters you need CPU, memory, disk I/O, SQL Server-specific counters
  • Plan a baselined monitoring window 24–72 hours
  • Enable or verify performance counters in Windows Performance Monitor PerfMon
  • Validate data collection with a quick test run

Introduction: what you’ll learn

• This guide covers why performance counters matter for SQL Server, how to enable them, and how to use the data to troubleshoot common issues like slow queries, blocking, memory pressure, and I/O bottlenecks.
• You’ll see practical steps, example commands, and checklists you can use in production or development environments.
• The guide uses familiar formats: step-by-step instructions, a checklist, tables for the most important counters, and a FAQ to answer common questions.

Useful URLs and Resources as plain text, not clickable

• Microsoft Docs – Performance Monitor PerfMon overview
• Microsoft Docs – SQL Server wait statistics
• SQL Server Dynamic Management Views DMVs overview
• Stack Overflow threads on perf counters and SQL Server tuning
• SentryOne blog on performance counters and baselining
• Redgate SQL Monitor best practices for performance data
• DBA Stack Exchange discussions on SQL Server performance tuning
• SQL Server Execution Plans official docs
• Windows Event Viewer troubleshooting guides
• Windows Performance Monitor documentation

Why you should enable performance counters

• They give you a real-time view of how Windows and SQL Server resources are used.
• With baselined data, you can spot anomalies quickly, like CPU spikes during batch jobs or memory pressure caused by large query results.
• They help with capacity planning, by showing when you’re running hot on I/O or memory.

What you’ll need

• A SQL Server instance you want to monitor
• Administrative access to the server local or remote
• Windows server with Performance Monitor PerfMon available
• A plan for data retention how long you’ll keep perf data
• Optional: a monitoring tool or SIEM that consumes PerfMon data

Step 1: Verify Windows performance counters are enabled

• PerfMon relies on Windows performance counters. If they’re disabled, data will be incomplete or missing.
• How to check:
  • Press Windows key, type PerfMon, and open Performance Monitor.
  • In the left pane, expand Data Collector Sets > User Defined. You should see active sets if you’ve configured anything.
  • If you don’t see counters, you may need to re-register performance counters via command line:
    • Open an elevated command prompt.
    • Run: lodctr /R
    • Then run: lodctr /R to rebuild the counters this can take a moment.
  • Reboot if you don’t see counters after re-registering.

Step 2: Decide which performance counters to enable

• Core Windows server counters recommended for most SQL Server workloads:
  • Processor: % Processor Time total and per core
  • Processor Queue Length
  • PhysicalDisk: Avg. Disk Queue Length
  • LogicalDisk: % Free Space drive letters used by SQL data/log files
  • Memory: Available MBytes
  • Memory:% Committed Bytes In Use
• SQL Server specific counters from SQLServer: Memory Manager, SQLServer: SQL Statistics, etc.:
  • SQLServer: Buffer Manager: Buffer cache Hit Ratio
  • SQLServer: Memory Manager: Total Server Memory KB
  • SQLServer: Memory Manager: Target Server Memory KB
  • SQLServer: Statistics: User Connections
  • SQLServer: Wait Statistics: Total Wait Time ms
  • SQLServer: Wait Statistics: Waits in Query Wait Stats
  • SQLServer: General Statistics: User SQL Cache Hit Ratio
  • SQLServer: Resource Pool Stats: CPU Usage %
• Define a sampling interval e.g., 15 seconds for granular data, 5 minutes for long-term trends
• Plan for retention: 30 days of data is a common baseline for larger environments

Step 3: Create and configure a data collection plan

• Use Windows Performance Monitor to collect data:
  • Open PerfMon
  • Go to Data Collector Sets > User Defined > New > Data Collector Set
  • Name it e.g., SQL_Server_Perf_Counters
  • Choose Create manually Advanced
  • Add counters:
    • For Windows counters: Processor, Memory, PhysicalDisk, LogicalDisk
    • For SQL Server: You’ll need the SQL Server perfmon objects; ensure the SQL Server instance is registered in the perfmon counter paths
  • Set the sample interval e.g., 15 seconds
  • Save the data to a dedicated folder or to a SQL Server integration if you prefer
• Alternative: use a third-party monitoring tool that reads PerfMon counters for you e.g., Datadog, Dynatrace, PRTG to simplify dashboards and alerting
• Tip: Create two collectors—one for short-term troubleshooting 15-second samples during a spike and one for long-term baselining 5-minute samples

Step 4: Enable and start data collection

• In Performance Monitor, right-click the Data Collector Set you created and choose Start.
• Run a representative workload on your SQL Server for 24–72 hours to gather baseline data. If you’re testing, simulate typical peak times e.g., nightly ETL, morning ad-hoc reports.

Step 5: Validate data collection and sanity check

• After starting, verify data is being written:
  • Go to the log folder and check for .blg binary log or .csv files, depending on setup.
  • Open a sample .csv file in a text editor or Excel to ensure values are being captured.
• Ensure counters are returning values:
  • In PerfMon, click the counters tab to see live values. If any counter is stuck at 0, recheck permissions or counter registration.

Step 6: Analyze using baseline and thresholds

• Create baselines:
  • Calculate typical values for CPU %, memory usage, disk queue lengths, and SQL Server waits during normal operations.
  • For example:
    • CPU: average 15–45%, spikes to 70–85% during heavy ETL
    • Disk queue length: average 1–2, spiking above 4 indicates I/O bottlenecks
    • Memory: Available MBytes should remain above a comfortable threshold e.g., 10–20% of total memory to avoid paging
• Set alert thresholds:
  • CPU usage above 85% for 5 minutes
  • Disk queue length > 2 for 10 minutes
  • Available memory consistently below 1000 MB on a 16 GB server
  • SQL Server wait type spikes above baseline e.g., high PAGEIOLATCH waits
• Visualize data:
  • Use graphs to correlate spikes e.g., CPU spike coinciding with high waits
  • Compare baseline to current data to spot anomalies

Step 7: Practical tips for specific scenarios

• Slow queries or high waits on CPU:
  • Check CXPACKET waits and hot queries
  • Review execution plans for expensive queries
  • Ensure up-to-date statistics and appropriate indexes
• Memory pressure:
  • Inspect Target Server Memory vs. Total Server Memory
  • Check memory grants for large queries
  • Consider memory-optimized tables or partitioning if applicable
• I/O bottlenecks:
  • MonitorAverage Disk Sec/Read and Average Disk Sec/Write
  • Look at disk latency and queue depth
  • Check for fragmentation or slow disks, consider adding faster storage or tuning file layouts
• Concurrency and blocking:
  • Capture Wait Statistics to identify blocking reasons
  • Examine deadlocks and long-running transactions
  • Verify query plans, index usage, and isolation levels

Step 8: Reporting and ongoing improvement

• Build dashboards that show key metrics:
  • CPU, memory, disk I/O, and waits
  • SQL Server specific counters buffer cache hit ratio, pages/sec, log flush wait, etc.
• Schedule regular reviews:
  • Weekly quick checks
  • Monthly performance baselines with trend analysis
• Share findings with the team and plan optimizations accordingly
• Keep perf counters aligned with changes in workload, like new reports, batch jobs, or schema changes

Table: Key counters to monitor summary

• Windows counters:
  • Processor: % Processor Time
  • Processor: % System Time
  • Memory: Available MBytes
  • PhysicalDisk: Avg. Disk Queue Length
  • LogicalDisk: % Free Space
• SQL Server counters:
  • SQLServer: Buffer Manager: Buffer cache Hit Ratio
  • SQLServer: Memory Manager: Total Server Memory KB
  • SQLServer: Memory Manager: Target Server Memory KB
  • SQLServer: Statistics: User Connections
  • SQLServer: Wait Statistics: Total Wait Time ms
  • SQLServer: Wait Statistics: Waits in Query Wait Stats
  • SQLServer: General Statistics: User SQL Cache Hit Ratio
  • SQLServer: Resource Pool Stats: CPU Usage %

Step 9: Advanced tips for production environments

• Use circular logging or separate deployments for perf data to avoid filling disks
• Rotate logs and archive baselines for long-term trends
• Consider sampling rate trade-offs: higher frequency gives more detail but larger data volume
• Integrate perf counter data with SQL Server DMVs for richer context e.g., sys.dm_os_wait_stats, sys.dm_exec_query_stats
• For cloud or virtualized environments, account for virtualization overhead and policy limits

Step 10: Troubleshooting common issues with perf counters

• Data not appearing
  • Counter registration problem re-register with lodctr
  • Insufficient permissions to access perf data
  • Data Collector Set misconfigured or not started
• Counter values look off
  • Counter reset after reboot; re-check startup
  • Performance counters disabled by group policy or security software
• Performance impacts from logging
  • Reduce log verbosity and ensure that logging itself isn’t causing contention
  • Use dedicated disks for perf logs if possible

Frequently asked questions

• What are performance counters?
  • They’re metrics collected from Windows and applications like SQL Server that show resource usage and performance characteristics.
• How often should I sample counters?
  • For pinpointing issues, 5–15 seconds. For baselining, 5 minutes or longer.
• Can I use PerfMon with SQL Server on Windows Server Core?
  • Yes, PerfMon works in core installations; you can still collect counters and log data.
• Do performance counters impact server performance?
  • They have minimal impact, especially with sensible sampling intervals.
• Which SQL Server counters are most important?
  • Buffer cache hit ratio, Total Server Memory, Target Server Memory, and wait stats.
• How do I baseline SQL Server performance?
  • Collect normal workloads over a period 24–72 hours and compute baselines for key metrics.
• How can I alert on performance issues?
  • Use PerfMon alerts or integrate with a monitoring tool to trigger when thresholds are crossed.
• How do I correlate waits with slow queries?
  • Cross-reference wait stats with DMVs like sys.dm_exec_query_stats and sys.dm_exec_requests.
• Should I collect performance counters during migrations?
  • Yes—baseline before, during, and after migrations to measure impact.
• How can I store perf data long-term?
  • Use a centralized repository or a monitoring solution that aggregates PerfMon data and SQL Server DMVs.

Note on tone and style

• This guide aims to be practical and friendly, like I’m walking you through the steps in a normal day at the office. If you hit a snag, tell me what you’re seeing, and I’ll tailor the steps to your environment. The goal is to make perf counter monitoring feel doable, not mystical.

Appendix: quick command references

• Re-register performance counters:
  • lodctr /R
• Open Performance Monitor:
  • Press Windows key + R, type perfmon, Enter
• Start a Data Collector Set manual steps:
  • PerfMon > Data Collector Sets > User Defined > New > Data Collector Set
  • Name > Create manually > Add counters > Set sample interval > Save and start

Endnotes

• Remember: the most effective monitoring happens when you combine Windows performance data with SQL Server DMVs and a clear baselined expectation of your workload. Start small, build a baseline, and expand your counter set as you gain confidence.

Yes, you can enable performance counters in SQL Server by following these steps. This guide walks you through verifying current status, repairing or reinstalling performance counters, re-registering them with LodCtr, and validating results with PerfMon. You’ll find a practical mix of steps, tips, and real-world checks to keep your SQL Server performance monitoring reliable.

• Quick-start overview:
  • Verify that SQL Server performance counters exist and are updating.
  • If counters aren’t showing, repair SQL Server or re-install the performance counters feature.
  • Re-register counters using LodCtr if needed.
  • Restart services and validate in PerfMon.
  • Common issues and practical fixes.
• Useful formats included: step-by-step guide, checklists, tables, and quick command references.
• Useful URLs and Resources un clickable text: Microsoft SQL Server Documentation – docs.microsoft.com. PerfMon Documentation – docs.microsoft.com. LodCTR command – support.microsoft.com. SQL Server Performance Counters – sqlservercentral.com. Windows Performance Monitor – support.microsoft.com

What are SQL Server performance counters?

Performance counters are Windows-based metrics that SQL Server uses to report its internal state. These counters feed into Performance Monitor PerfMon and help you track CPU usage, memory pressure, I/O waits, buffer pool activity, and many other vital signals. Examples include:

• SQLServer:Buffer Manager – Page life expectancy
• SQLServer:Memory Clerk – a specific memory pool
• SQLServer:General Statistics – User Connections, Logins/sec
• SQLServer:SQL Statistics – Batch Requests/sec, SQL Compilations/sec

Having them enabled lets you build dashboards, alerting rules, and capacity plans with real data.

Prerequisites

• Administrative access to the Windows server and the SQL Server instance.
• The SQL Server Performance Counters feature installed either by default or via repair/add features.
• A recent backup plan for critical servers, in case you need to perform maintenance.
• A plan to restart SQL Server services after re-registration notify stakeholders.

Step 1 — Verify current status of performance counters

• Check if counters exist in PerfMon:
  • Open Performance Monitor perfmon.
  • Add counters and look for the SQLServer performance object e.g., SQLServer:Buffer Manager, SQLServer:Memory Clerk, SQLServer:General Statistics, SQLServer:SQL Statistics.
• Quick command line check:
  • On Windows, open Command Prompt as Administrator and run:
  • typeperf -qx | findstr /i “SQLServer”
    This lists installed SQL Server counters. If you don’t see SQLServer counters, you likely need to repair or re-register them.
• Check that the SQL Server service is running:
  • Open SQL Server Configuration Manager.
  • Ensure the SQL Server INSTANCE_NAME service is started.

Step 2 — Repair or reinstall the SQL Server Performance Counters feature

If PerfMon shows no SQLServer counters or data isn’t updating, repair the installation or add the feature that includes performance counters.

• Use SQL Server Installation Center:
  • Run the setup executable for your SQL Server version.
  • Choose “Maintenance” → “Repair” instead of a complete reinstall to re-register components, including Performance Counters.
  • If Repair isn’t available or doesn’t fix it, use “Add features to an existing instance” and ensure that “SQL Server Performance Counters” usually under Shared Features or SQL Server Feature Pack is installed.
• After repair/installation, re-check in PerfMon that SQLServer counters appear.

Tips:

• A repair often resolves minor registry or registration issues without touching your data.
• If you’re on a cluster, perform the repair on the active node and validate failover correctness.

Step 3 — Re-register performance counters using Lodctr

If counters still aren’t appearing after repair, you might need to re-register them at the system level with Lodctr. How to Enable HSTS in Windows Server 2016: A Complete IIS Guide for HTTPS Security and Preload 2026

• Locate the SQL Server performance counters.ini file the exact path varies by version and installation, but common locations include under the SQL Server installation directory, often something like:
  • C:\Program Files\Microsoft SQL Server\MSSQL15.MSSQLSERVER\MSSQL\Binn\SQLServerPerfCounters.ini
  • For older versions, replace MSSQL15.MSSQLSERVER with the correct instance folder.
• Running Lodctr:
  • Open Command Prompt as Administrator.
  • Re-register the SQL Server counters by pointing Lodctr to the ini file, for example:
    • lodctr /R
    • lodctr “C:\Program Files\Microsoft SQL Server\MSSQL15.MSSQLSERVER\MSSQL\Binn\SQLServerPerfCounters.ini”
  • The /R option rebuilds the performance counters from the existing DLLs and registrations. the ini file explicitly loads the SQL Server-specific counters.
• If Lodctr reports errors, check that:
  • The path to the INI file is correct.
  • You have permissions to modify performance counter registry entries.
  • The SQL Server service binaries are not blocked by antivirus or group policy.

Note: The exact path to the ini file can differ by version. If you’re unsure, search the installation directories for SQLServerPerfCounters.ini or consult your version’s documentation.

Step 4 — Restart services and re-check counters

• Restart the SQL Server service:
  • In SQL Server Configuration Manager, right-click the SQL Server INSTANCE_NAME service and choose Restart.
• It’s also wise to restart the Windows Performance Monitor service or simply reboot the server if you’re in a maintenance window.
• Re-open PerfMon:
  • Add the SQLServer performance counters again.
  • Confirm that the counters are updating watch a counter like Batch Requests/sec over a minute.

Step 5 — Validate performance counters with practical checks

• Quick validation steps:
  • Check a few core counters:
    • SQLServer:Buffer Manager – Page life expectancy
    • SQLServer:Memory Manager – Total server memory KB
    • SQLServer:General Statistics – User Connections
    • SQLServer:SQL Statistics – Batch Requests/sec
  • If values look stale or always zero, there might be a service or registration issue.
• Baseline measurements:
  • Capture baseline values over a 5–10 minute window during normal load.
  • Note typical ranges for your workload. e.g., Page life expectancy should be higher than a few hundred seconds on healthy systems under moderate load. high CPU or memory pressure will reflect in relevant counters.

Step 6 — Common issues and how to fix them

• Issue: No SQLServer counters found after installation.
  • Fix: Run Repair in the SQL Server setup, then re-check PerfMon.
• Issue: Counters exist but aren’t updating.
  • Fix: Re-register counters with Lodctr, restart services, verify again.
• Issue: Access denied when running Lodctr.
  • Fix: Run the command prompt as Administrator. ensure you have permissions to modify the registry.
• Issue: Antimalware/EDR blocks DLLs used by performance counters.
  • Fix: Add SQL Server directories to trusted paths or exclude from real-time scanning during maintenance.
• Issue: Counter values appear blurred or inconsistent under heavy load.
  • Fix: Ensure OS performance counters aren’t throttled by policy. confirm there’s no third-party monitoring tool conflicting with PerfMon.

Step 7 — Best practices for ongoing SQL Server performance monitoring

• Create a health dashboard:
  • Track a small set of core counters that align with your workload.
  • Set alerts for abnormal thresholds e.g., Page life expectancy dropping below a defined value, Batch Requests/sec spiking unexpectedly, or memory pressure.
• Schedule periodic verification:
  • Every quarter or after major updates, verify that performance counters are still registered and updating.
• Combine PerfMon with DMV-based data:
  • Use dynamic management views DMVs to complement PerfMon data, offering deeper insights into CPU, memory, I/O waits, and query performance.
• Document your baseline:
  • Keep a simple document with typical value ranges for your environment and the steps you took to enable/verify counters.
• Consider a lightweight monitoring stack:
  • Tools like Prometheus + Grafana, or cloud-native monitoring, can ingest PerfMon data or equivalent SQL Server metrics for centralized dashboards.

Data and example counters you’ll likely monitor

• Page life expectancy PLE: A higher PLE generally indicates healthier memory pressure. Low PLE can signal memory pressure due to large workloads or insufficient RAM.
• Buffer cache hit ratio: The percentage of data reads served from memory. A very low ratio may indicate memory pressure or IO bottlenecks.
• Batch Requests/sec: Reflects how many batches of statements SQL Server processes per second.
• Transactions/sec: Shows transaction throughput. monitor trends during peak hours.
• Checkpoints/sec, Lazy writes/sec: Help you spot I/O and storage subsystem pressure.
• Disk queue length: Across data drives hosting SQL data files to detect I/O bottlenecks.

Sample table: common counters and their meaning

Counter Category	Counter Name	What it tells you
Memory	SQLServer:Memory Clerk	Memory allocations by clerk, useful to identify memory pressure
Buffer	SQLServer:Buffer Manager	Page life expectancy, Buffer cache hit ratio
General	SQLServer:General Statistics	User connections, Logins/sec, Not fully reliable under all workloads
Queries	SQLServer:SQL Statistics	Batch requests/sec, SQL compilations/sec, SQL workers
I/O	SQLServer:Wait Stats via DMVs	Wait types that signal I/O bottlenecks, e.g., PAGEIOLATCH_SH

Note: Use DMVs to complement PerfMon counters for deeper analysis:

• sys.dm_os_wait_stats
• sys.dm_exec_query_stats
• sys.dm_os_memory_clerks

Frequently asked questions

How do I know if performance counters are installed with SQL Server?

You can verify by checking PerfMon for SQLServer counters e.g., SQLServer:Buffer Manager. You can also run typeperf -qx | findstr /i “SQLServer” to list SQL Server counters if present.

Can I enable performance counters on a SQL Server Express edition?

Yes, SQL Server Express includes performance counters. however, some enterprise-only features may not be present. The steps to enable or repair counters are the same. How to enable sftp server in ubuntu a comprehensive guide 2026

What if I don’t see SQLServer in PerfMon after repair?

Re-run the LodCtr re-registration steps and ensure the INI file path is correct for your version. Then restart the SQL Server service and PerfMon.

Do performance counters affect SQL Server performance?

No, enabling or updating performance counters is non-intrusive in normal operation. However, excessive monitoring activity can add a small overhead. typically negligible.

How often should I refresh or verify performance counters?

Perform a verification after any major SQL Server or OS changes, and periodically e.g., quarterly to ensure counters remain healthy.

What is the best practice for monitoring performance counters at scale?

Use a centralized monitoring solution that ingests PerfMon data or uses equivalent SQL Server metrics via DMVs. Create dashboards and alert rules for key baselines.

Can performance counters help with capacity planning?

Yes. They provide insights into memory usage, I/O throughput, and query workload over time, which are essential for planning RAM, storage, and CPU capacity. How to Enable DNS on OpenVPN Server DD-WRT: A Step-by-Step Guide for DNS Over VPN and Router Setup 2026

How long should I collect baseline performance data?

Capture 5–10 minutes for initial checks under typical load, then extend to 24–72 hours for robust baselines during different workloads.

What are the most important counters to start with?

Page life expectancy, Buffer cache hit ratio, Batch Requests/sec, Transactions/sec, SQLServer:Memory Manager Total server memory KB, and Disk queue length for the drives hosting your data files.

Do I need to restart Windows after enabling performance counters?

Often restarting the SQL Server service is sufficient. a full Windows restart is only necessary if LodCtr or OS-level changes require it and you’re troubleshooting deeper registry issues.

Useful URLs and Resources un clickable text

• Microsoft SQL Server Documentation – docs.microsoft.com
• PerfMon Documentation – docs.microsoft.com
• LodCTR command – support.microsoft.com
• SQL Server Performance Counters – sqlservercentral.com
• Windows Performance Monitor – support.microsoft.com
• SQL Server DMV references – docs.microsoft.com
• Performance tuning guides – microsoft.com
• OS performance tuning best practices – techcommunity.microsoft.com
• SQL Server setup and repair guides – docs.microsoft.com
• Community forums and blogs for troubleshooting SQL Server counters – stackoverflow.com

Frequently Asked Questions expanded

Is there a risk of data loss when repairing performance counters?

No. Repairing performance counters does not affect user data or database contents. It only ensures the monitoring instrumentation is correctly registered and updating.

Can I enable performance counters on remote SQL Server instances?

Yes. The process is the same on each server, but you’ll need to perform the repair/re-registration and verification on each instance individually. How to enable line number in sql server step by step guide 2026

How can I automate performance counter checks?

Use PowerShell to query PerfMon counters or pull DMV metrics, then push to a central monitoring system. Scripted validations can alert you if counters stop updating.

What if I see a lot of stale values in counters?

Stale values usually indicate a registration issue or a service restart needed. Verify counter registration, restart services, and re-check.

Do performance counters cover memory, CPU, and I/O?

Yes. The SQLServer performance counters cover memory, buffer management, general statistics, SQL statistics, and related I/O activity, giving you a broad view of server health.

How long should I keep historical counter data?

Aim for at least 30–90 days of perf data for trend analysis, with longer retention if you’re doing capacity planning or post-incident reviews.

Are there alternatives to PerfMon for SQL Server metrics?

Yes. Many enterprises use Prometheus with exporters, SQL Server native DMVs, or commercial monitoring tools that gather SQL metrics and provide dashboards. How to Enable DNS Server in Packet Tracer: Setup, Configuration, and Troubleshooting 2026

Can performance counters help with debugging slow queries?

Indirectly. They help identify systemic problems memory pressure, I/O waits, CPU saturation that contribute to slow queries. combine with query DMVs for deeper insights.

How can I prevent performance counter issues after Windows updates?

Test on a staging server, re-check counter registrations after updates, and consider running the LodCtr re-register steps if needed as part of update validation.

Sources:

翻墙后究竟能玩点啥？解锁你的数字自由新世界 – VPN 使用指南、跨境访问、隐私保护全攻略

Vpn是什么、Vpn是什么及如何选择、使用与评估VPN服务的全面指南（2025更新版）

Big ip edge client と は vpn: a comprehensive guide to Big-IP Edge Client, how it works, setup, features, and comparisons How to enable auditing on windows server 2012: Setup, Policy, and Logging for Comprehensive Monitoring 2026

Zscaler and vpns how secure access works beyond traditional tunnels

Iphone vpn一直打开的原因、实现方法与注意事项：在 iPhone 上持续启用 VPN 的完整指南