Laravel Octane + FrankenPHP segfault (exit status 139) on PHP 8.3 ZTS with native extensions

[Chanceium]

Description

Issue

The Solidtime container crashes with a segmentation fault (exit status 139) shortly after startup when running with CONTAINER_MODE: http.

Environment

• Image: solidtime/solidtime:latest
• PHP: 8.3.23 (ZTS)
• Server: FrankenPHP with Laravel Octane
• Container: Docker
• Orchestration: Docker Compose

Steps to Reproduce

1. Start container with docker-compose up
2. Observe server starts and caches are built successfully
3. Server crashes with supervisor restart loop

Expected Behavior

Server should stay running and accept requests without segfaults.

Actual Behavior

2026-03-18 00:45:03,641 WARN exited: octane_0 (exit status 139; not expected)
2026-03-18 00:45:04,644 INFO spawned: 'octane_0' with pid 53
2026-03-18 00:45:05,646 INFO success: octane_0 entered RUNNING state

Exit status 139 = SIGSEGV (segmentation fault at C level)

Root Cause Analysis

• PHP 8.3.23 built with ZTS (Thread Safe)
• FrankenPHP + Octane keep workers long-lived with threading
• Loaded high-risk extensions: uv, rdkafka, vips, memcached, redis
• Issue: ZTS + FrankenPHP + these extensions = memory corruption in worker lifecycle

PHP Modules Loaded

bcmath, bz2, Core, ctype, curl, date, dom, exif, fileinfo, filter, gd, hash, 
iconv, igbinary, intl, json, ldap, libxml, mbstring, memcached, mysqlnd, 
openssl, pcntl, pcre, PDO, pdo_mysql, pdo_pgsql, pdo_sqlite, pgsql, Phar, 
posix, random, rdkafka, readline, redis, Reflection, session, SimpleXML, 
sockets, sodium, SPL, sqlite3, standard, tokenizer, uv, vips, xml, xmlreader, 
xmlwriter, Zend OPcache, zip, zlib

Suggested Fix

1. Use PHP 8.2/8.3 NON-ZTS instead of ZTS build
2. Switch from FrankenPHP to RoadRunner (more stable for Octane)
3. Reduce extension load - particularly uv, rdkafka, vips if not critical
4. Document Octane + FrankenPHP limitations in setup guide

Notes

• The Laravel application itself is sound (boots correctly)
• Issue is runtime/extension level, not application code
• RoadRunner is known to be more stable with heavy native extension loads

Steps To Reproduce

Going to /admin in the solidtime and going to Users or Organizations

Self-hosted or Cloud?

Self-Hosted

Version of solidtime: (for self-hosted)

latest

solidtime self-hosting guide: (for self-hosted)

No response

[Onatcer]

What system are you experiencing this behaviour on? I know there was a problem with OPCache enabled on ARM Systems (php/frankenphp#1031) but I'm not sure it still occurs.

Edit: ok that old issue seems to be solved as of PHP 8.3.15 (php/php-src#16727) I do need some way to reproduce the issue then. it would also help if you try to disable extensions like vips, rdkafka and uv one-by-one and check it the issue still persists.

[Chanceium]

persists.

ARM system

[Onatcer]

are you experiencing the segfault when using the main application as well or only when accessing the admin panel? If it's only the admin panel it might be related to this livewire issue livewire/livewire#9763 (the PR is merged there but people still report they are experiencing it)

[Chanceium]

are you experiencing the segfault when using the main application as well or only when accessing the admin panel? If it's only the admin panel it might be related to this livewire issue livewire/livewire#9763 (the PR is merged there but people still report they are experiencing it)

Only the admin panel

[Chanceium]

@Onatcer Disabling docker-php-ext-opcache.ini fixed the issue; I can now access admin panel without it breaking.