F42 RP4 Failed to start mariadb.service - MariaDB 10.11 database server. aarch64

Ask Fedora
1

Goodevening,

Running mariadb for years without any issues.

But today the mariadb.service failed to start.

What could be wrong?

            Stack trace of thread 4814:
                                                      #0  0x0000ffffa0088e68 __syscall_cancel_arch (libc.so.6 + 0x98e68)
                                                      #1  0x0000ffffa007b83c __internal_syscall_cancel (libc.so.6 + 0x8b83c)
                                                      #2  0x0000ffffa007bc94 __GI___futex_abstimed_wait_cancelable64 (libc.so.6 + 0x8bc94)
                                                      #3  0x0000ffffa007e928 pthread_cond_wait@@GLIBC_2.17 (libc.so.6 + 0x8e928)
                                                      #4  0x0000aaaae7d02608 _Z11mysqld_mainiPPc (/usr/libexec/mariadbd + 0x1e2608)
                                                      #5  0x0000ffffa001621c __libc_start_call_main (libc.so.6 + 0x2621c)
                                                      #6  0x0000ffffa00162fc __libc_start_main@@GLIBC_2.34 (libc.so.6 + 0x262fc)
                                                      #7  0x0000aaaae7cde4b0 _start (/usr/libexec/mariadbd + 0x1be4b0)

Yours kindly.

Omer Netherlands

2

I observed this too on multiple F42 machines after the recent mariadb dnf updates a week or so ago.

Workarounds:

(1) As root: dnf reinstall mariadb-server ; systemctl stop mariadb-server ; systemctl start mariadb-server

This allows mariadb server to start up, but sometimes it fails again if the server is rebooted.

(2) I ended up upgrading a couple of machines to F43, and the problem seems to have gone away.

Good luck!

3

Iā€™m seeing mariadb startup failures on two different Fedora 42 systems now.

After updating to mariadb Ver 15.1 Distrib 10.11.15-MariaDB mariadb fails to start at reboot.

The issue appears to be that some process is deleting the pidfile directory (/run/mariadb) at reboot.

despite re-adding /run/mariadb and setting ownership to mysql:root, the directory is removed after every reboot (itā€™s unaffected by stopping mariadb.service), causing mariadb startup to fail.

/var/log/mariadb/mariadb.log reports:

2026-01-17  3:53:12 0 [Note] Starting MariaDB 10.11.15-MariaDB source revision cb0d6dd835023a7162ace471cd047161f205dd58 server_uid IFK/IK
+ljKqnJAsyYqNn+oGTS2U= as process 5658
2026-01-17  3:53:12 0 [Note] InnoDB: Compressed tables use zlib 1.3.1.zlib-ng
2026-01-17  3:53:12 0 [Note] InnoDB: Number of transaction pools: 1
2026-01-17  3:53:12 0 [Note] InnoDB: Using crc32 + pclmulqdq instructions
2026-01-17  3:53:12 0 [Note] InnoDB: Using Linux native AIO
2026-01-17  3:53:12 0 [Note] InnoDB: innodb_buffer_pool_size_max=128m, innodb_buffer_pool_size=128m
2026-01-17  3:53:12 0 [Note] InnoDB: Initialized memory pressure event listener
2026-01-17  3:53:12 0 [Note] InnoDB: Completed initialization of buffer pool
2026-01-17  3:53:12 0 [Note] InnoDB: File system buffers for log disabled (block size=512 bytes)
2026-01-17  3:53:12 0 [Note] InnoDB: End of log at LSN=111411123468
2026-01-17  3:53:12 0 [Note] InnoDB: 128 rollback segments are active.
2026-01-17  3:53:12 0 [Note] InnoDB: Removed temporary tablespace data file: "./ibtmp1"
2026-01-17  3:53:12 0 [Note] InnoDB: Setting file './ibtmp1' size to 12.000MiB. Physically writing the file full; Please wait ...
2026-01-17  3:53:12 0 [Note] InnoDB: File './ibtmp1' size is now 12.000MiB.
2026-01-17  3:53:12 0 [Note] InnoDB: log sequence number 111411123468; transaction id 225706055
2026-01-17  3:53:12 0 [Note] Plugin 'FEEDBACK' is disabled.
2026-01-17  3:53:12 0 [Note] InnoDB: Loading buffer pool(s) from /var/lib/mysql/ib_buffer_pool
2026-01-17  3:53:12 0 [Note] Server socket created on IP: '0.0.0.0', port: '3306'.
2026-01-17  3:53:12 0 [Note] Server socket created on IP: '::', port: '3306'.
2026-01-17  3:53:12 0 [ERROR] mariadbd: Can't create/write to file '/run/mariadb/mariadb.pid' (Errcode: 2 "No such file or directory")
2026-01-17  3:53:12 0 [ERROR] Can't start server: can't create PID file: No such file or directory
terminate called without an active exception
260117  3:53:12 [ERROR] /usr/libexec/mariadbd got signal 6 ;
Sorry, we probably made a mistake, and this is a bug.

I created a systemd-tmpfiles configuration file for mariadb at /usr/lib/tmpfiles.d/mariadb.conf:

d /run/mariadb 0755 mysql mysql

Then reloaded the systemd-tmpfiles configuration with:

systemctl restart systemd-tmpfiles-clean.service

and rebooted to test the change. /run/mariadb with appropriate ownership and permissions is now created at boot time and mariadb now starts up without issue.

4

I also see a similar stack trace. However, I see the following 30-40 lines before the stack trace:

2026-01-17 3:59:28 0 [ERROR] mariadbd: Canā€™t create/write to file ā€˜/run/mariadb/mariadb.pidā€™ (Errcode: 2 ā€œNo such file or directoryā€)
2026-01-17 3:59:28 0 [ERROR] Canā€™t start server: canā€™t create PID file: No such file or directory

If you see the same message in your /var/log/mariadb/mariadb.log, please see my other comment on this thread for more details.

5

Thank you very much, i make a daily backup of the database which is empty after de update:

backup file 13 january full

after dnf update the mariadb backup file 14 empty:

sudo dnf history list --contains-pkgs=mariadb
ID Command line Date and time Action(s) Altered
71 dnf -y update 2026-01-13 20:43:01 96

So it looks like there is bug in this version:

mariadb aarch64 3:10.11.15-1.fc42

#sudo dnf downgrade mariadb
Downgrading mariadb-3:10.11.11-1.fc42.aarch64

After downgrading mariadb.service starts without any problem.

How can I help fixing this bug?

Yours kindly Omer.

6
1 Like
7

Thanks for the reference @vgaetera.

It seems that this regression is fixed with adding a mariadb.conf to a tmpfiles.d directory.
However, the suggested fix (comments 8 and 10 in the bug report you linked seems like overkill.

Just creating /usr/lib/tmpfiles.d/mariadb.conf with appropriate info for the pidfile directory does the trick.

Presumably the next update will resolve this and overwrite the created file. Unless Iā€™m missing something. Am I?