Issue with MariaDB "Got error: 2002: "Can't connect to local server through socket '/tmp/mysql.sock' (2)" when trying to connect"

[YousufSSyed]

I'm unable to startup Mariadb, spent a while troubleshooting, and haven't found any solutions. At this point, I just want it working so that I can export my database and switch to running Mariadb on Lima, because Mariadb on Homebrew is simply that unreliable. I ran all the commands outputted by brew doctor, ran brew reinstall mariadb, and a few other things a while ago that I forgot, but I still get the same error. For each things I tried, I ran brew services restart mariadb. I'm on macOS 12.4.

I'm going to be very patient with fixing if I have to do lots of back and forth with troubleshooting.

Edit: I'm using an M1 Mac.

[SMillerDev]

What does brew services say about mariadb? Is there actually a /tmp/mysql.sock file to use?

[SMillerDev]

What does ls -al /opt/homebrew/etc/my.cnf* say?

[YousufSSyed]

-rw-r--r--@ 1 yousuf  admin  204 Jul  6 04:28 /opt/homebrew/etc/my.cnf
-rwxrwxrwx  1 yousuf  admin  206 Jul  6 06:17 /opt/homebrew/etc/my.cnf.default
-rwxrwxrwx  1 yousuf  admin  113 Jul  6 06:17 /opt/homebrew/etc/my.cnf.default.default

/opt/homebrew/etc/my.cnf.d:
total 0
drwxrwxrwx   3 yousuf  admin    96 Apr 10 09:02 .
drwxrwxrwx  35 yousuf  admin  1120 Jul  6 04:27 ..
-rwxrwxrwx   1 yousuf  admin     0 Jul  6 06:17 .homebrew_dont_prune_me

[SMillerDev]

And what is the content of /opt/homebrew/etc/my.cnf?

[carlocab]

Can't read dir of '/opt/homebrew/etc/my.cnf.'

That extra . looks suspicious

[gromgit]

-rw-r--r--@ 1 yousuf  admin  204 Jul  6 04:28 /opt/homebrew/etc/my.cnf

The @ sign after the permissions says this file has extended attributes, one or more of which may be preventing you from opening it.

What's the output of ls -ld@ /opt/homebrew/etc/my.cnf?

[YousufSSyed]

@carlocab Yeah I thought the . might be noteworthy.

@gromgit ls -ld@ /opt/homebrew/etc/my.cnf outputs this:

-rw-r--r--@ 1 yousuf  admin  204 Jul  6 04:28 /opt/homebrew/etc/my.cnf
	com.apple.LaunchServices.OpenWith	150 
	com.apple.lastuseddate#PS	 16

(I deleted the comment since the section was getting overflown).

[gromgit]

This doesn't look like a proper command looking at the output

It is a proper command, I was really looking for this output:

Default options are read from the following files in the given order:
/opt/homebrew/etc/my.cnf ~/.my.cnf 

to verify that the command wasn't looking in some weird place like, well, /opt/homebrew/etc/my.cnf. (with the trailing period).

Is there a ~/.my.cnf on your system? If so, what is its contents?

[YousufSSyed]

/opt/homebrew/etc/my.cnf. outputs No such file or directory. Same thing with ~/.my.cnf

[SMillerDev]

How about a /etc/my.cnf?

[YousufSSyed]

It's either an alias or symlink to /opt/homebrew/etc/my.cnf (I'm pretty sure it's a symlink).

[YousufSSyed]

Ok I just figured out it's a symlink.

[gromgit]

What are the contents of /opt/homebrew/var/mysql/MacBook-Pro.local.err?

[gromgit]

What does lsof /tmp/mysql.sock output?

[maazmmd]

➜ lsof /tmp/mysql.sock
COMMAND    PID    USER   FD   TYPE             DEVICE SIZE/OFF NODE NAME
mariadbd 66984 mohmaaz   25u  unix 0xedbfea7a63d0ad4e      0t0      /tmp/mysql.sock

[gromgit]

There's your stray mariadbd process. ps -ef | grep 66984 to see what exactly was run, that might also give a clue as to why it didn't appear in any of the ps/pgrep outputs.

[maazmmd]

I restarted my mac, as I was getting another issue with port already in use. Here's the output of stray mariadbd process

➜ ps -ef | grep 1341
1013988708  1341  1110   0 11:59am ??         0:00.83 /opt/homebrew/opt/mariadb/bin/mariadbd --basedir=/opt/homebrew/opt/mariadb --datadir=/opt/homebrew/var/mysql --plugin-dir=/opt/homebrew/opt/mariadb/lib/plugin --log-error=/opt/homebrew/var/mysql/capel0084052675..err --pid-file=capel0084052675..pid --socket=/tmp/mysql.sock
1013988708  2358  2357   0 12:01pm ttys003    0:00.01 q _ inline-shell-completion --buffer ps -ef | grep 1341
1013988708  2361  1769   0 12:01pm ttys003    0:00.00 grep --color=auto --exclude-dir=.bzr --exclude-dir=CVS --exclude-dir=.git --exclude-dir=.hg --exclude-dir=.svn --exclude-dir=.idea --exclude-dir=.tox 1341

[maazmmd]

Fixed the issue ➜ gave permissions to /tmp and .err and .pid files as mariadb was unable to read/write files and adding tmpdir = /tmp in my.cnf. Thanks for your help @gromgit

# add tmp directory in my.cnf
tmpdir          = /tmp

# Grant permissions
sudo chmod 0777 /tmp
sudo chmod 0777 /opt/homebrew/var/mysql/{your_mac_profile}..err
sudo chmod 0777 /opt/homebrew/var/mysql/{your_mac_profile}..pid
chmod 0777 /opt/homebrew/var/{your_mac_profile}..err

Reference: https://stackoverflow.com/questions/8633373/mysql-cant-create-write-to-file-error13

[draven3166]

There are many cases for Mariadb or any other service to stop working properly on the computer. My case MacOs Sequoia
causes

• Sometimes a memory error and/or ram crash.
• OS error

They can be the specific cause in this type of problems and for that a generic question in the communities to it does not always lead you to the answer/solution valid since the causes are several/different.
In my case I have followed the guidelines above but it did not solve the problem. And one of those was to execute the process of direct call to mariadb and its handler that uses the socket by default, this when executed tells you which threads and dependencies it has, it draws you technically where everything happens, besides showing you how it works.
# mysqld
preview it is important to see
# mysqld --verbose --help | grep “socket”
To locate exactly where your configurations are in the system and what values they are taking.
My specific case was a memory error and a break when initializing services and dependencies which broke everything and had no service and could not schedule a three hours!!!

It is solved by making a backup of the tc.log file (same path in #coment) delete the last one, this case a read and follow this thread and not have the file mysql.sock in place. And run.
# mysqld
so that the instances in mysql.servers regenerate the paths and processes to regenerate the path and be able to start the service with the correct configuration without breaking or forcing any process to act arbitrarily because that is the workflow where there is a problem you have to investigate to solve it and start everything again well.
Do yourself a favor as a developer read the warnings that the prompt returns and the errors that cause them, it will help you a lot.

Note

Dont forget restart service to include last configuration in your pc.
Happy code.