Core files constantly being generated and filling partition

[lostsoul77]

[lostsoul77]

@laf

select * from config where config_name=‘poller_modules.mib’;

I ran the command with correct syntax and deleted from the database; i’ll monitor and see if the errors and core files stop generating

[lostsoul77]

@laf

delete from config where config_id

Errors were still generating; ran database command again using discovery_modules.cisco-sla instead of poller_modules.cisco-sla; and showed enabled in the database; I deleted and will monitor

[lostsoul77]

so core files are still generating after removing cisco-sla from database. I look at the time of the last core file which was 12:57 pm today and did a grep on the librenms.log file for that time and getting YamlDiscovery warnings as shown in the screenshot:

[lostsoul77]

All…I’m at a loss. The files are still generating and filling the filesystem every 2-3 days with these core files. I don’t know much about mysql unfortunately and looking for additional guidance.

[lostsoul77]

running 25.7.0 Release (#18066):
one of the core files generated at 9:15 and 9:30 this morning; when i search librenms.log, there are no warnings or errors. When I search poller_wrapper.log for those times I see these warnings:

discovery-wrapper.log shows these errors:
2025-08-04 07:33:27,303 :: WARNING :: worker Thread-20 finished device 1777 in 900 seconds with exit code -254
2025-08-04 07:33:27,303 :: ERROR :: discovery-wrapper checked 415 devices in 3625 seconds with 16 workers with 220 errors

[murrant]

I’ve never seen php core dump from LibreNMS. It could be an issue with your underlying PHP install.

1. you could try running discovery by hand and see if it core dumps PHP.
2. perhaps you have the cisco-sla module defined in config.php (FYI, I doubt this is a related issue)

[lostsoul77]

here’s an example of the files…fills the filesystem about every 3 days and getting frustrating as I don’t know how to prevent this. I have to login and do an rm -rf core.* to get the space back.

[lostsoul77]

So I found a command to read the core files @laf @murrant @rhinoau

gdb ‘/path/to/executable /path/to/core’

after running it against several of the files, it appears to be an issue with discovery.php -h new …see screenshot. Hopefully someone can chime in and provide more guidance now that I’ve found the cause of the core dump files

[lostsoul77]

Also, i found that the crontab is referencing the discovery.php -h new command (the command that is generating the core dumps); could this be the issue?

[laf]

I mean discovery.php -h new is run on 1000s of installs perfectly fine.

If it was me, I’d remove all traces of php and re-install it. Ensuring you remove config files and see if that helps.

[murrant]

This is PHP itself crashing. You have an issue with the installed version of PHP on your system. Nothing LibreNMS can do can fix that. You’ll need to update or repair your PHP.