[BUG] php-fpm segfaults caused by php8

Is there an existing issue for this?

I have searched the existing issues

Current Behavior

#512
Since the above issue was closed and can't reply i have to open a new issue.
What I did, removed all mods, fresh instance.
Install/update app, php-fpm segfaults. Recovery by pkill php-fpm.
It has 12GB of RAM, so is not a memory issue.
Because of php8 there are alot of errors in the logs:

Implicit nullable parameters are deprecated - PHP 8.4 now deprecates implicit nullable parameter declarations and requires explicit nullable types
Stricter type checking causing fatal errors instead of warnings
Changes to null handling in string operations
Many deprecations from PHP 7.4-8.1 became breaking changes by PHP 8.2+.

I did not have the chance to build a new docker image with an older version of php (not that good with docker), maybe somebody can try it out?

Expected Behavior

Not to crash nextcloud/php-fpm.

Steps To Reproduce

Environment

- OS: debian 12
- How docker service was installed: sh

CPU architecture

arm64

Docker creation

docker compose

Container logs

This issue has been automatically marked as stale because it has not had recent activity. This might be due to missing feedback from OP. It will be closed if no further activity occurs. Thank you for your contributions.

not stale

Could you complete all the parts of the issue request please with the information requested? Not just leaving them blank / -

here in this closed ticket are all the details #512

Install/update app, php-fpm segfaults

I don't know what you mean by this.

Your other issue was tainted with other stuff. That info doesn't apply here.

Nextcloud has been on php8 for a long time. No one else has reported anything like this. This is most likely a you issue.

You need to provide us with all the info requested, perhaps more if you customized other stuff yourself. We need to be able to reproduce the issue ourselves in order to diagnose and fix. Otherwise we cannot help you.

This issue has been automatically marked as stale because it has not had recent activity. This might be due to missing feedback from OP. It will be closed if no further activity occurs. Thank you for your contributions.

I can confirm php-fpm crash issue on arm image (Raspberry Pi). When attempting to reinstall any of the NextCloud applications, such as Office or Contacts, the server starts returning a 503 error. To reproduce the problem, open Apps, delete one of the installed applications, then click Download and Enable for the same application.

Here are PHP-FPM logs:

[31-Oct-2025 08:24:24] NOTICE: fpm is running, pid 426
[31-Oct-2025 08:24:24] NOTICE: ready to handle connections
[31-Oct-2025 08:28:48] WARNING: [pool www] child 479 exited on signal 11 (SIGSEGV - core dumped) after 118.997951 seconds from start
[31-Oct-2025 08:28:48] NOTICE: [pool www] child 484 started
[31-Oct-2025 08:28:54] WARNING: [pool www] child 480 exited on signal 11 (SIGSEGV - core dumped) after 122.344011 seconds from start
[31-Oct-2025 08:28:54] NOTICE: [pool www] child 486 started
[31-Oct-2025 08:31:27] WARNING: [pool www] child 484 exited on signal 11 (SIGSEGV - core dumped) after 159.143359 seconds from start
[31-Oct-2025 08:31:27] NOTICE: [pool www] child 489 started
[31-Oct-2025 08:31:29] WARNING: [pool www] child 483 exited on signal 11 (SIGSEGV - core dumped) after 174.746978 seconds from start
[31-Oct-2025 08:31:29] NOTICE: [pool www] child 490 started
[31-Oct-2025 08:31:29] WARNING: [pool www] child 485 exited on signal 11 (SIGSEGV - core dumped) after 161.080357 seconds from start
[31-Oct-2025 08:31:29] NOTICE: [pool www] child 491 started
[31-Oct-2025 08:31:34] WARNING: [pool www] child 475 exited on signal 11 (SIGSEGV - core dumped) after 392.790328 seconds from start
[31-Oct-2025 08:31:34] NOTICE: [pool www] child 492 started
[31-Oct-2025 08:32:32] WARNING: [pool www] child 490 exited on signal 11 (SIGSEGV - core dumped) after 62.858789 seconds from start
[31-Oct-2025 08:32:32] NOTICE: [pool www] child 493 started
[31-Oct-2025 08:34:09] NOTICE: Terminating ...
[31-Oct-2025 08:34:09] NOTICE: exiting, bye-bye!

And Nginx log:

2025/10/31 08:28:48 [error] 456#456: *2 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: <host>, server: _, request: "GET /apps/files/ HTTP/2.0", upstream: "fastcgi://127.0.0.1:9000", host: "<host>"
2025/10/31 08:28:54 [error] 456#456: *2 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: <host>, server: _, request: "GET /settings/apps/installed HTTP/2.0", upstream: "fastcgi://127.0.0.1:9000", host: "<host>"
2025/10/31 08:29:55 [error] 456#456: *2 upstream timed out (110: Operation timed out) while reading response header from upstream, client: <host>, server: _, request: "GET /favicon.ico HTTP/2.0", upstream: "fastcgi://127.0.0.1:9000", host: "<host>"
2025/10/31 08:32:32 [error] 457#457: *163 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: <host> server: _, request: "PROPFIND /remote.php/dav/files/admin/ HTTP/1.1", upstream: "fastcgi://127.0.0.1:9000", host: "<host>"

Docker Compose config is almost default:

---
services:
  nextcloud:
    image: lscr.io/linuxserver/nextcloud:32.0.1
    container_name: nextcloud
    environment:
      - PUID=1000
      - PGID=1000
      - TZ=Etc/UTC
    volumes:
      - ./data/nextcloud/config:/config
      - ./data/data:/data
    ports:
      - 443:443
    restart: unless-stopped

The issue only related to ARM image, on x86 this image works fine.

This issue has been automatically marked as stale because it has not had recent activity. This might be due to missing feedback from OP. It will be closed if no further activity occurs. Thank you for your contributions.

@j0nnymoe

I never had issues while the image was based on php 8.3, since 8.4 all sort of issues appeared.
@j0nnymoe @aptalca maybe a rollback to php 8.3 could be possible?

@Seriyyy95

@Seriyyy95 disable JIT, in php-local.ini

opcache.jit=0
opcache.jit_buffer_size=0

Restart the nextcloud container.

It should be also noted that JIT in general works on x86 only, not on ARM.

@VeloAddict I have installed php84-dbg but still didn't get working mappings. Here is what I managed to get with gdb:

(gdb) bt full
#0  0x0000aaaab5b017dc in ?? ()
No symbol table info available.
#1  0x0000aaaab5a020b0 in ?? ()
No symbol table info available.
#2  0x0000ffffc70ec348 in ?? ()
No symbol table info available.
Backtrace stopped: previous frame inner to this frame (corrupt stack?)
(gdb) info proc mappings
Mapped address spaces:

          Start Addr           End Addr       Size     Offset objfile
      0xaaaaada00000     0xaaaab5a00000  0x8000000        0x0 /dev/zero (deleted)
      0xaaaab5a00000     0xaaaabda00000  0x8000000  0x8000000 /dev/zero (deleted)
      0xaaaabdb20000     0xaaaabdf58000   0x438000        0x0 /usr/sbin/php-fpm84
      0xaaaabdf6c000     0xaaaabdff0000    0x84000   0x43c000 /usr/sbin/php-fpm84
      0xaaaabdff0000     0xaaaabdff2000     0x2000   0x4c0000 /usr/sbin/php-fpm84
      0xffff89554000     0xffff89556000     0x2000        0x0 /dev/zero (deleted)
      0xffff895f6000     0xffff8b5f6000  0x2000000        0x0 /dev/zero (deleted)
      0xffff8b984000     0xffff8b9a3000    0x1f000        0x0 /usr/lib/libffi.so.8.1.4
      0xffff8b9a3000     0xffff8b9a4000     0x1000     0xf000 /usr/lib/libffi.so.8.1.4
      0xffff8b9a4000     0xffff8b9a5000     0x1000    0x10000 /usr/lib/libffi.so.8.1.4

However, disabling JIT helped. After I disabled it, php-fpm doesn't crash anymore. Thanks.

This issue has been automatically marked as stale because it has not had recent activity. This might be due to missing feedback from OP. It will be closed if no further activity occurs. Thank you for your contributions.

the issue will not go away unless JIT is disabled for the arm64 build

This issue has been automatically marked as stale because it has not had recent activity. This might be due to missing feedback from OP. It will be closed if no further activity occurs. Thank you for your contributions.

This issue is locked due to inactivity

Navigation Menu

Repository navigation

[BUG] php-fpm segfaults caused by php8 #514

Is there an existing issue for this?

Current Behavior

Expected Behavior

Steps To Reproduce

Environment

CPU architecture

Docker creation

Container logs

11 remaining items

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

[BUG] php-fpm segfaults caused by php8 #514

Description

Is there an existing issue for this?

Current Behavior

Expected Behavior

Steps To Reproduce

Environment

CPU architecture

Docker creation

Container logs

Activity

LinuxServer-CI commented on Sep 9, 2025

unknown-error-maker commented on Sep 9, 2025

j0nnymoe commented on Sep 9, 2025

unknown-error-maker commented on Sep 14, 2025

aptalca commented on Sep 14, 2025

LinuxServer-CI commented on Oct 15, 2025

Seriyyy95 commented on Oct 31, 2025

LinuxServer-CI commented on Dec 1, 2025

11 remaining items

unknown-error-maker commented on Feb 7, 2026

unknown-error-maker commented on Feb 7, 2026

unknown-error-maker commented on Feb 8, 2026

Seriyyy95 commented on Feb 9, 2026

LinuxServer-CI commented on Mar 12, 2026

unknown-error-maker commented on May 4, 2026

LinuxServer-CI commented on Jun 3, 2026

LinuxServer-CI commented on Sep 2, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions