MySQL demon already running with same unix socket

Question

When trying to start the mysqld demon on a CentOS 6.5 box I get the following error:

[root@hardin ~]# service mysqld start
Another MySQL daemon already running with the same unix socket.
Starting mysqld:                                           [FAILED]

Which is odd because the service isn't started (it's configured to run on the default port) :

[root@hardin ~]# service mysqld status
mysqld is stopped

I'm not sure how to find the Unix socket that's causing the problem:

Active UNIX domain sockets (servers and established)
Proto RefCnt Flags       Type       State         I-Node Path
unix  10     [ ]         DGRAM                    12056  /dev/log
unix  2      [ ACC ]     STREAM     LISTENING     8720   @/com/ubuntu/upstart
unix  2      [ ACC ]     STREAM     LISTENING     12878  /var/run/mcelog-client
unix  2      [ ]         DGRAM                    9289   @/org/kernel/udev/udevd
unix  2      [ ]         DGRAM                    12440  @/org/freedesktop/hal/udev_event
unix  2      [ ACC ]     STREAM     LISTENING     12275  /var/run/dbus/system_bus_socket
unix  2      [ ACC ]     STREAM     LISTENING     13128  /var/run/abrt/abrt.socket
unix  2      [ ACC ]     STREAM     LISTENING     12415  @/var/run/hald/dbus-jQMFHN9EAI
unix  2      [ ACC ]     STREAM     LISTENING     12408  @/var/run/hald/dbus-YQIZ9h7s4Q
unix  2      [ ACC ]     STREAM     LISTENING     12371  /var/run/acpid.socket
unix  2      [ ]         DGRAM                    14119
unix  3      [ ]         STREAM     CONNECTED     13579  /var/run/dbus/system_bus_socket
unix  3      [ ]         STREAM     CONNECTED     13578
unix  2      [ ]         DGRAM                    13521
unix  3      [ ]         STREAM     CONNECTED     13493
unix  3      [ ]         STREAM     CONNECTED     13492
unix  2      [ ]         DGRAM                    13130
unix  2      [ ]         DGRAM                    13058
unix  2      [ ]         DGRAM                    13023
unix  2      [ ]         DGRAM                    12875
unix  2      [ ]         DGRAM                    12840
unix  3      [ ]         STREAM     CONNECTED     12773  /var/run/acpid.socket
unix  3      [ ]         STREAM     CONNECTED     12772
unix  3      [ ]         STREAM     CONNECTED     12767  @/var/run/hald/dbus-YQIZ9h7s4Q
unix  3      [ ]         STREAM     CONNECTED     12766
unix  3      [ ]         STREAM     CONNECTED     12716  @/var/run/hald/dbus-YQIZ9h7s4Q
unix  3      [ ]         STREAM     CONNECTED     12634
unix  3      [ ]         STREAM     CONNECTED     12714  @/var/run/hald/dbus-YQIZ9h7s4Q
unix  3      [ ]         STREAM     CONNECTED     12534
unix  3      [ ]         STREAM     CONNECTED     12435  @/var/run/hald/dbus-jQMFHN9EAI
unix  3      [ ]         STREAM     CONNECTED     12434
unix  3      [ ]         STREAM     CONNECTED     12410  /var/run/dbus/system_bus_socket
unix  3      [ ]         STREAM     CONNECTED     12409
unix  2      [ ]         DGRAM                    12375
unix  3      [ ]         STREAM     CONNECTED     12295  /var/run/dbus/system_bus_socket
unix  3      [ ]         STREAM     CONNECTED     12294
unix  3      [ ]         STREAM     CONNECTED     12289
unix  3      [ ]         STREAM     CONNECTED     12288
unix  3      [ ]         DGRAM                    9308
unix  3      [ ]         DGRAM                    9307

I use this box as a Git server, running GitLab, and it used to work. After a cold boot it doesn't work any more because unicorn can't connect to MySQL for the reason above:

/home/git/gitlab/vendor/bundle/ruby/2.1.0/gems/mysql2-0.3.11/lib/mysql2/client.rb:44:in `connect': Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (111) (Mysql2::Error)

I'm wondering how to troubleshoot this and find what service is running on the MySQL port.

Can you check : is mysql listening `netstat -anp | grep 3306` ? Socket exists `ls -l /var/lib/mysql/mysql.sock` ? — krisFR, Feb 06 '14 at 09:30

user9517 · Accepted Answer · 2014-02-06T09:41:39.253

5

You can use

lsof -U | grep -i mysql
mysqld    28944     mysql   12u  unix 0xffff88002cacb6c0      0t0  532368 /var/lib/mysql/mysql.sock
mysqld    28944     mysql   29u  unix 0xffff88002a724380      0t0  532402 /var/lib/mysql/mysql.sock

to see which process (if any) owns the socket and take appropriate action

If you're sure that there is no other MySQL process running, you can simply delete the /var/lib/mysql/mysql.sock file. When MySQL starts, if the socket doesn't exist, it will recreate it.

edited Feb 06 '14 at 09:41

answered Feb 06 '14 at 09:34

user9517

115,471
20
215
297

2

There was no process using it, but removing the file fixed the issue. I didn't know it was possible to do that. If upon next reboot this happens again I'll investigate some more. Thanks. – Astaar Feb 06 '14 at 11:17

MySQL demon already running with same unix socket

1 Answers1