Unresponsive Sites Troubleshooting

alrnetwork

alrnetwork

Verified User
Joined
Feb 12, 2021
Messages
162
Location
Europe UTC+2
Hi all,

I am hoping that someone here with a bit more knowledge of how Apache / PHP handles requests, could help me figure out, why randomly the web server seems to stop responding to traffic / requests.

I can't seem to find a specific trigger which starts this issue off, but I'll find that as I'm working on a WordPress website, suddenly the web browser will just continuously load and not move on to the page or complete the action that I have requested.

I have UptimeRobot checking my sites and it seems as though all sites on that server will go "down", which can be further tested by trying to reach a site.

Interestingly, sites which have caching, sometimes load, but if I try to hit a page where it needs to query the database directly, such as wp-admin, then the site will show the same continuous loading as others.

Double-interestingly, the DirectAdmin control panel loads fine, which suggests a few things to me:

- Server load is okay as DA pages are responsive and quick like normal
- Possibly working because it's on a different port number, and not 80/443?

SSH into terminal and using "htop" shows no excessive CPU / Memory consumption, often having over 70% free resources

I haven't changed any Apache configurations, not comfortable doing that anyway, but I do normally change certain PHP variables:
  • max_execution_time
  • max_input_time
  • memory_limit
  • post_max_size
  • upload_max_filesize
  • max_file_uploads
It seems that by restarting httpd via Terminal or via DirectAdmin, solves the issue and sites return to normal functioning behaviour, and it's unlikely that I will see this issue again for a day or two.

Any ideas?
 
It might be an odd php-fpm issue like I was experiencing on our previous server. However, could be anything.

You could be start by finding the cause of this. So when this is happening, check php-fpm log, apache log, apache error log, domain log and domain error log at that moment, and check for any lines that might point to this issue.

Without any clues from the log, it's very hard to say what could be causing this.
 
Thanks! It just happened for around 58 minutes just now when I wasn't even working on the sites, so certainly going to dig through logs and see what I can make sense of,
 
Further research on this:

Other websites on the same server for other users remain unaffected, however, all sites under the user which has the issue, are affected.

Therefore, I concentrated my attention to the PHP-FPM log, and picked up on the following:

"[17-May-2021 19:10:03] WARNING: [pool alrnetwork] server reached max_children setting (10), consider raising it"

It's clear that the solution will be to raise the max_children setting, however, I would like to know why this is required? Ideally want to avoid changing configurations if possible.
 

Version 1.57.3 | Directadmin Docs

DirectAdmin Knowledge Base
www.directadmin.com

oldest installer will remain max 100 children
since 1.57 it lower to 10

If you are larger hosting or have many user/website try increase it to fit your product like increase to 30 50 70 100

Setting should be match your hardware, if too much process can't handle by your hardware. it make negative result.
 
I tried to change the setting as per instructions here: https://help.directadmin.com/item.php?id=496

Now, none of the websites on the server work. Log shows: "
[17-May-2021 19:55:43] ALERT: [pool alrnetwork] pm.max_children must be a positive value"

The "custom" config file has been removed, but even though the server was rebooted, it's not working. All site down.
 
I can confirm that using the following (linked in your previous reply) solved the issue:

cd /usr/local/directadmin/custombuild
./build rewrite_confs
 
You must log in or register to reply here.
Back
Top