What issue or error are you experiencing?
The issue is a plain error page
Service Unavailable
The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later.
This is not the colorized Unavailable page when a site is not started.
After much much much debugging, I managed to get PHP-FPM to reveal the error;
[31-Jan-2025 13:17:18] WARNING: [pool www] child 78201 said into stderr: "php-fpm(78201,0x2043b8240) malloc: *** error for object 0x1ffffffff: pointer being freed was not allocated"
[31-Jan-2025 13:17:18] WARNING: [pool www] child 78201 said into stderr: "php-fpm(78201,0x2043b8240) malloc: *** set a breakpoint in malloc_error_break to debug"
[31-Jan-2025 13:17:18] WARNING: [pool www] child 78201 exited on signal 6 (SIGABRT) after 42.227657 seconds from start
[31-Jan-2025 13:17:18] NOTICE: [pool www] child 78238 started
This happens when using dns_get_record or checkdnsrr.
While one might think this is a DNS resolver issue;
- problem seems not to apply with the wp-cli / php cli shell.
- gethostbyname works just fine.
What steps can be taken to replicate the issue? Feel free to include screenshots, videos, etc
<?php dns_get_record('acato.nl', DNS_A);
call test.php via the webbrowser
System Details
Local version: Version 9.1.1+6738
OS: macOS Sequoia 15.2
Local Logs
There is nothing in the logs that is useful here. It is spammed with
{"class":"Process","level":"warn","message":"nginx: [warn] the \"listen ... http2\" directive is deprecated, use the \"http2\" directive instead in %%router.runPath%%/nginx/conf/server-block-ssl.conf:1","process":"nginx","thread":"main","timestamp":"2025-01-30T15:07:13.016Z"}
and that should be resolved, but is completely unrelated and out of scope for this issue.
The log needed is the php-fpm log and I did not find that anywhere on the system
By finding the process (ps -A | grep SITEID), killing it and replacing it with a console non-daemonized instance
kill PROCESSID ; /Users/username/Library/Application\ Support/Local/lightning-services/php-8.1.29+0/bin/darwin-arm64/sbin/php-fpm -F -R -p /Users/username/Library/Application\ Support/Local/run/SITEID/conf/php/ -y /Users/username/Library/Application\ Support/Local/run/SITEID/conf/php/php-fpm.conf
I managed to get the error as listed above.
It is a hassle to get all this, but I can list that the following PHP-FPM versions ALL crash;
- 8.0.30
- 8.1.29
- 8.2.23
And now LocalWP won’t finish the “Starting up Site Services” so I want to do a system reboot, and therefore I have to cut this investigation short.
After a reboot;
- 5.6.39
- 7.4.30
- 8.3.11
All crash the same.
And for all versions listed; the CLI version does NOT crash, works perfectly.
Additional information discovered; when disabling IPv6 on my WiFi – setting it to link local only – The problem goes away, even if the site still runs on ::1 and I haven’t restarted the site …
