Website & aaPanel Unreachable Until VPS Reboot

aaP_jackievon51

Hi aaPanel team,

I’m having a recurring issue and would appreciate your assistance.

Environment:
VPS provider: Hostinger
VPS specs: 4 vCPU, 16 GB RAM
Panel: aaPanel (latest version)
Website: WordPress
DNS: Cloudflare (proxied mode enabled)

Issue Description:
From time to time, my WordPress website suddenly becomes unreachable. When the issue happens:

Cloudflare shows:
“Web server is down” Host Error;
Browser → Working;
Cloudflare → Working;

aaPanel cannot be accessed either, giving:
connection was unexpectedly closed. ERR_CONNECTION_CLOSED

However, in Hostinger’s dashboard the VPS appears to be running normally (no CPU spikes, no overload alerts, etc.).

The issue can ONLY be resolved by rebooting the VPS.

This problem repeats randomly, which makes the server unreliable.

What could cause both aaPanel and all sites to stop responding simultaneously?

Any recommended steps to diagnose or monitor this kind of intermittent failure?

Any help or suggestions would be greatly appreciated.
Thank you!

aaPanel_Kern

Please check if nginx/apache is running?

aaP_jackievon51

aaPanel_Kern aaPanel is inaccessible, so I cannot check the status of Nginx/Apache.

aaPanel_Kern

Hello,
It is recommended that you check whether the server is running normally? Check whether the server load is normal? Is it possible to use ssh or the server provider's VNC?

aaP_jackievon51

aaPanel_Kern The VPS status is normal, and SSH access works without any issues.

aaPanel_Kern

Hello, is the problem solved?
If it has not been resolved, you can use the ssh tool to execute the command to check it:
curl -k https://www.aapanel.com/script/aapanel_check.sh | bash

aaP_jackievon51

aaPanel_Kern Hello, this issue occurs almost every 1–2 months (the VPS itself is running normally, but both the control panel and the website become inaccessible). Are there any internal logs in the control panel that could help identify what is causing this? Thank you.

aaP_jackievon51

aaPanel_Kern

Hello, recently my VPS has again encountered an issue where the control panel and website were inaccessible, while the VPS itself was still accessible via SSH. After running this script, the website has returned to normal. I’ve attached a screenshot taken after running the script—could you please help me identify what might have caused this issue? Thank you.

aaPanel_Kern

Hello,

It is recommended that you check whether the process was closed by the system?

dmesg|grep -iE "kill|httpd|apache|http|mysql|oom"
grep -i 'killed' /var/log/messages
journalctl | grep -i 'killed'

aaP_jackievon51

aaPanel_Kern

Hello

Is the panel currently running normally?
Can fix it by executing this command to update and repair it to the latest version?

mv /www/server/panel/pyenv /www/server/panel/pyenv_bak
rm -f /tmp/update_to7.pl && curl -k https://node.aapanel.com/install/update_panel_en.sh | bash

aaP_jackievon51

aaPanel_Kern

Hello,

thank you for your reply and help.
After running the script:curl -k https://www.aapanel.com/script/aapanel_check.sh | bash
the control panel became accessible again.

I’m a bit confused about what exactly causes the panel to become inaccessible from time to time. I would like to resolve this issue completely. Do you have any better suggestions or solutions?

Thank you.

aaPanel_Kern

Hello,

Please check if the service is running
It is usually caused by the panel service not running.

aaP_jackievon51

aaPanel_Kern
Hi,
Recently, I found that both my website and aaPanel became inaccessible at the same time, while the VPS itself was still accessible normally via SSH.

After checking, I found the following issues:
My server encountered OOM (Out of Memory) issues. After logging into the aapanel, I noticed that PHP was killed by the system.

Here are the related system logs:

[2385679.736402] oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0,global_oom,task_memcg=/system.slice/mysqld.service,task=mysqld,pid=673788,uid=1002
[2385679.739086] Out of memory: Killed process 673788 (mysqld) total-vm:3481776kB, anon-rss:1081196kB, file-rss:0kB, shmem-rss:0kB, UID:1002 pgtables:2664kB oom_score_adj:0
[2385681.793562] oom_reaper: reaped process 673788 (mysqld), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB
[2397270.571286] oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0,global_oom,task_memcg=/system.slice/mysqld.service,task=mysqld,pid=720473,uid=1002
[2397270.573806] Out of memory: Killed process 720473 (mysqld) total-vm:3470512kB, anon-rss:855880kB, file-rss:0kB, shmem-rss:0kB, UID:1002 pgtables:2216kB oom_score_adj:0
[2397309.764953] oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0,global_oom,task_memcg=/system.slice/btpanel.service,task=php-fpm,pid=720669,uid=1001
[2397309.767429] Out of memory: Killed process 720669 (php-fpm) total-vm:476064kB, anon-rss:131968kB, file-rss:0kB, shmem-rss:203760kB, UID:1001 pgtables:912kB oom_score_adj:0
May 13 22:20:46 xxx kernel: Out of memory: Killed process 673788 (mysqld) total-vm:3481776kB, anon-rss:1081196kB
May 14 01:33:54 xxx kernel: Out of memory: Killed process 720473 (mysqld) total-vm:3470512kB, anon-rss:855880kB
May 14 01:34:32 xxx kernel: Out of memory: Killed process 720669 (php-fpm) total-vm:476064kB, anon-rss:131968kB, shmem-rss:203760kB

My PHP-FPM config is currently:
pm = dynamic
pm.max_children = 200
pm.start_servers = 15
pm.min_spare_servers = 15
pm.max_spare_servers = 50

PHP settings:
memory_limit = 512M

VPS environment:
16GB RAM
WordPress + WooCommerce
aaPanel
MySQL + PHP 8.4

Question:
Could pm.max_children = 200 be the main reason for the OOM issue? For a 16GB RAM server running WordPress + WooCommerce, what would be a more reasonable PHP-FPM configuration (especially pm.max_children)? Also, should I lower memory_limit from 512M?

Thanks in advance.

aaPanel_Kern

Hello,

Does the website use caching plugins?
It is recommended that you use 8G memory configuration for PHP.
Does the operating system enable SWAP?

aaP_jackievon51

aaPanel_Kern Thank you for your suggestion. I will increase the PHP memory allocation to 8GB and continue monitoring the server.

In addition, my website is using a caching plugin.

I still have one question: why did aaPanel also become inaccessible when PHP was killed?

I have already enabled automatic PHP monitoring in aaPanel, so whenever PHP stops, the system should automatically restart PHP.

However, in this case, not only was PHP stopped, but aaPanel itself also became unavailable and I could not access the backend.

aaPanel_Kern

It is recommended that you check whether the panel service is running normally? Has it also been shut down by the system?

aaP_jackievon51