This is easy to confirm on Ubuntu (14.10 in my case):
openxs@ubuntu:~$ mysql -uroot -proot -e'select version()'
Warning: Using a password on the command line interface can be insecure.
+-------------+
| version() |
+-------------+
| 5.6.23-72.1 |
+-------------+
openxs@ubuntu:~$ sudo /etc/init.d/mysql stop
[sudo] password for openxs:
* Stopping MySQL (Percona Server) mysqld [ OK ]
openxs@ubuntu:~$ /etc/init.d/mysql start
df: '/var/lib/mysql/.': Permission denied
* /etc/init.d/mysql: ERROR: The partition with /var/lib/mysql is too full!
openxs@ubuntu:~$ df -H
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/ubuntu--vg-root 7.0G 1.4G 5.3G 21% /
none 4.1k 0 4.1k 0% /sys/fs/cgroup
udev 516M 4.1k 516M 1% /dev
tmpfs 106M 353k 105M 1% /run
none 5.3M 0 5.3M 0% /run/lock
none 527M 0 527M 0% /run/shm
none 105M 0 105M 0% /run/user
/dev/sda1 247M 39M 196M 17% /boot
Nothing like this happens, say, with CentOS. The problematic code is:
61 # check for diskspace shortage
62 datadir=`mysqld_get_param datadir`
63 if LC_ALL=C BLOCKSIZE= df --portability $datadir/. | tail -n 1 | awk '{ ex it ($4>4096) }'; then
64 log_failure_msg "$0: ERROR: The partition with $datadir is too full!"
65 echo "ERROR: The partition with $datadir is too full!" | $ERR_LOGGER
66 exit 1
67 fi
68 }
This is easy to confirm on Ubuntu (14.10 in my case):
openxs@ubuntu:~$ mysql -uroot -proot -e'select version()' ubuntu- -vg-root 7.0G 1.4G 5.3G 21% /
Warning: Using a password on the command line interface can be insecure.
+-------------+
| version() |
+-------------+
| 5.6.23-72.1 |
+-------------+
openxs@ubuntu:~$ sudo /etc/init.d/mysql stop
[sudo] password for openxs:
* Stopping MySQL (Percona Server) mysqld [ OK ]
openxs@ubuntu:~$ /etc/init.d/mysql start
df: '/var/lib/mysql/.': Permission denied
* /etc/init.d/mysql: ERROR: The partition with /var/lib/mysql is too full!
openxs@ubuntu:~$ df -H
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/
none 4.1k 0 4.1k 0% /sys/fs/cgroup
udev 516M 4.1k 516M 1% /dev
tmpfs 106M 353k 105M 1% /run
none 5.3M 0 5.3M 0% /run/lock
none 527M 0 527M 0% /run/shm
none 105M 0 105M 0% /run/user
/dev/sda1 247M 39M 196M 17% /boot
Nothing like this happens, say, with CentOS. The problematic code is:
61 # check for diskspace shortage `mysqld_ get_param datadir`
62 datadir=
63 if LC_ALL=C BLOCKSIZE= df --portability $datadir/. | tail -n 1 | awk '{ ex it ($4>4096) }'; then
64 log_failure_msg "$0: ERROR: The partition with $datadir is too full!"
65 echo "ERROR: The partition with $datadir is too full!" | $ERR_LOGGER
66 exit 1
67 fi
68 }