Service eligibility and regional restrictions
PuppyIP serves only compliant overseas businesses and their authorized personnel. Proxy services are not available in mainland China. The service may only be used for lawful business activities outside mainland China. Use of this service within mainland China is prohibited.
Hosting a proxy IP or server overseas does not change these restrictions. The service must not be provided to end users in mainland China through relaying, forwarding, sharing or resale. Before use, read the Terms of Service.
Key Takeaways
- An SSH drop may involve a control connection routed through a restarting proxy, or another network or service change.
- A different result after disabling the desktop proxy shows that the client path changed; it does not identify whether the client, server or a security device was responsible.
- Before restarting remote Xray or network services, prepare an approved recovery path such as a cloud console or out-of-band management.
- Compare the effective SSH configuration, TCP reachability and service logs for the same target; a TCP check does not prove SSH authentication.
Check whether SSH uses the desktop proxy
SSH may use ProxyCommand, ProxyJump, user-level configuration or desktop proxy software to change its route. If the SSH connection itself passes through local Xray, a proxy reload can affect that TCP session. Timing alone does not establish which component caused the drop.
On the affected computer, inspect the effective configuration for the SSH target, especially its hostname, port, ProxyCommand and ProxyJump. Output may contain hostnames, usernames or proxy arguments; inspect it locally and do not post the full configuration or command line in a public ticket.
Inspect the effective SSH proxy configuration in Windows OpenSSH
powershell$sshTarget = 'REPLACE_WITH_AUTHORIZED_SSH_CONFIG_ALIAS'
ssh -G $sshTarget | Select-String '^(hostname|user|port|proxycommand|proxyjump) '
Check the target TCP path and port
On a Windows client, first check TCP reachability to the SSH target port. Use the host and port from an existing authorized configuration; do not try other public ports to bypass network controls. If the TCP check fails, record the client network, VPN/proxy state, interface and time, then compare them with server-side listeners and cloud firewall settings.
A successful TCP check means only that a connection could be established at that moment. It does not validate SSH host keys, user authentication or session stability. If policy prohibits a direct path, follow that policy and use an approved proxy route; do not disable the security proxy globally for troubleshooting.
Read-only TCP check for the SSH target
powershell$sshHost = 'REPLACE_WITH_AUTHORIZED_SSH_HOST'
$sshPort = [int]'REPLACE_WITH_SSH_PORT'
Test-NetConnection -ComputerName $sshHost -Port $sshPort
Correlate Xray state events with the session time
On the server, check the actual Xray unit's active state, recent start time and journal entries for the same time range. Determine whether a restart, crash or configuration-load error coincides with the SSH drop. If logs are missing, rotated or the service runs in a container, inspect the records from its actual manager instead.
Check the SSH listener address and port, cloud security group and host firewall, but do not change the SSH port or flush all rules just because Xray restarted. If the SSH service did not restart while the client's proxy connection changed, keep tracing the control path and local proxy state.
Read the confirmed Xray unit state and logs
shUNIT='REPLACE_WITH_CONFIRMED_XRAY_UNIT'
sudo systemctl show "$UNIT" -p ActiveState -p SubState -p ExecMainStartTimestamp
sudo journalctl -u "$UNIT" --since 'REPLACE_WITH_START_TIME' --no-pager
Keep a management route independent of the failing proxy
Before restarting Xray, desktop proxy software or network services, confirm an authorized recovery path such as a cloud-provider console, out-of-band management or another operations channel that does not pass through the proxy. The current SSH session is not an adequate rollback path if SSH itself is proxied.
If policy allows a path comparison for one management target, change only that target's existing routing rule and record its original value. Do not turn off the computer's proxy globally or change the default route. If no approved direct or alternate path exists, keep the current policy and investigate through the console.
Restore only the service or path supported by evidence
If logs confirm that the Xray unit exited, its configuration passes validation and a management path is available, have the service administrator restore that unit through its deployment's manager during a maintenance window. If the service is healthy but SSH still fails, inspect the local proxy, target route, cloud security group and SSH service logs instead of repeatedly restarting Xray.
If a single-target routing rule or desktop client setting was changed for comparison, restore its original value afterward and confirm that other traffic still follows the required policy. A proxy toggle alone cannot rule out an unknown security device, carrier interruption or server restart; keep the timeline and narrow the cause with evidence.
Verify login and session stability
After restoring an authorized path, reconnect to the same SSH alias and check the host key, authentication result and server-side logs. Observe the connection through the intended management route, then confirm that Xray state, client proxy settings and routing policy remain as expected. Do not treat one TCP probe or successful login as a guarantee of long-term stability.
If SSH works only after disabling the desktop proxy, limit the conclusion to a relationship with a changed client path. Confirming the cause still requires the effective SSH configuration, Xray events and server-side connection records. Keep a redacted timeline, network location, test results and unknowns; mark unconfirmed network or security-device behavior as pending.
Sources
Frequently Asked Questions
Does SSH recovery after disabling the desktop proxy prove Xray was the root cause?
No. It shows that the client used a different network path. Check the effective SSH configuration, Xray events and server-side logs as well.
Does a successful Test-NetConnection mean SSH login works?
No. It checks TCP connection establishment, not SSH host keys, authentication or session stability.
Can I restart Xray or networking from the remote SSH session?
First confirm whether SSH depends on that proxy and prepare an authorized cloud console or independent management route. Do not reload networking remotely without a recovery path.