SOLVED: How to find more info on 503 errors?

[mkaaaay]

Since moving away from Apache, we're getting a bunch of 503 errors on pages outside the caching system. How do I find why this is happening?

have already set abort: no and auto fix=yes.

This is seriously having an impact on business and I'm left wondering is it even worth it if this keeps happening?

The pages in questions are all various pages behind an admin system, and include pages that have very little overhead - such as login page etc...

The site is running php7.0, and uses opcode.

 

[NiteWave]

https://docs.litespeedtech.com/extapp/php/troubleshooting/503/

 

[mkaaaay]

What does this suggest?

2020-05-28 08:19:17.020922 [NOTICE] [1643793] [ip.number.:50424#APVH_website.com] No request delivery notification has been received from LSAPI process group [-1], possible run away process.
2020-05-28 08:19:17.020944 [NOTICE] [1643793] [ip.number.:50424#APVH_website.com] Max retries has been reached, 503!
2020-05-28 08:19:17.020951 [NOTICE] [1643793] [ip.number.:50424#APVH_website.com] oops! 503 Service Unavailable
2020-05-28 08:19:17.020953 [NOTICE] [1643793] [ip.number.:50424#APVH_website.com] Content len: 179, Request line: 'POST /admin/login.php HTTP/1.1'
2020-05-28 08:19:17.020954 [INFO] [1643793] [ip.number.:50424#APVH_website.com] Cookie len: 313, __utma=260911760.1023120361.1568156047.1590548143.1590617558.408; __utmz=260911760.1568156047.1.1.utmcsr=google|utmccn=(organic)|utmcmd=organic|utmctr=(not%20provided); site_admin_cookie=site_admin_logged_in; PHPSESSID=o6bv7nonhhvba1ho22fpfn7e84; __utmb=260911760.2.10.1590617558; __utmc=260911760; __utmt=1
2020-05-28 08:19:17.020957 [INFO] [1643793] [ip.number.:50424#APVH_website.com] abort request..., code: 4
2020-05-28 08:19:17.020990 [INFO] [1643793] [ip.number.:50424#APVH_website.com] File not found [/home/website/subdomains/here/httpdocs/503.shtml]

 

[mkaaaay]

not sure why it's aborting a request if we have the flag to prevent that already set?

 

[Pong]

Just need to follow the 503 doc and the solution should be one of them. 503 is php crash error.

 

[mkaaaay]

What doesn't make sense though, is that it;s only happening in a 30 minute window, on a specific page. All other times of the day, the page is perfectly fine.

 

[mkaaaay]

OK, so i worked out, the cause was from a stock update script that takes about 45 mins to get and sync data fro an external db. The sync runs via cron.

<IfModule lsapi_module>
   php_flag display_errors Off
   php_value max_execution_time 90
   php_value max_input_time 120
   php_value max_input_vars 3000
   php_value memory_limit -1
   php_value post_max_size 512M
   php_value session.gc_maxlifetime 14400
   php_value session.save_path "/var/cpanel/php/sessions/ea-php70"
   php_value upload_max_filesize 64M
   php_flag zlib.output_compression On
</IfModule>

Attached the is the custom php.ini, so why are we getting the 503 still? Is there another place we need to make a change to execution time or something?

** this was never an issue under apache.

 

[serpent_driver]

Maybe it's a stupid question, but have you really installed lsapi?

 

[Pong]

Do you mean you only get 503 error when a stock update script run around 30 mins on a specific page, none of other situations?

 

[mkaaaay]

lsapi is running, yes (we had actually switched to lsapi before moving to LSWS).

And yes, the 503 only seems to happen when this script is running in the background.

Is there a way to whitelist a path, or tell LSWS to ignore limits on a dir or something?

 

[serpent_driver]

Is there a way to whitelist a path, or tell LSWS to ignore limits on a dir or something?

With LSWS something works different as with Apache and with limits and timeouts. Check "tuning" settings in LSWS WebAdmin. For example if you want to increase max_execution_time you have to change Connection Timeout in WebAdmin. For LSPHP you find more information at: https://docs.litespeedtech.com/extapp/php/configuration/control/

 

[mkaaaay]

Still no luck really. I added .htacess and "whitelisted" (no abort/timeout) all files in the folder that has the script that runs. The script runs fine.

Issue is, almost as soon as I run it, either as a wget or as a url, it almost immediately throws a 503 on any non-cached file on the site (and continues to run successfully for 40 mins or so) .

 

[Pong]

Could you log a ticket with us and provide tmp root ssh login to your server, and detailed wget and curl commands for us to check the errors?

 

[mkaaaay]

Could you log a ticket with us and provide tmp root ssh login to your server, and detailed wget and curl commands for us to check the errors?

Thank you yes.

 

[mkaaaay]

Issue appears to be resolved now, thanks to the help from admin.
For anyone having a similar issue, this is what I did.

1. lsws web admin->Server->General->PHP suEXEC Max Conn
   changed from default 10 > 40 (our server has 8 cores)
   restart lsws
   
   
2. pkill lsphp (just to make sure restart applied changes)

 

[serpent_driver]

Thank you, good to know