CentOS - Sep 2016 - Install Mariadb on Centos 7 fails

At 09:36 AM 9/23/2016, you wrote:
>On 2016-09-22 18:07, david wrote:
> > Folks
> >
> > I'm having problems installing Mariadb in Centos 7.
> >
> >
> > Here's what I did:
> >
> > 0)  Become root
> > 1)  Make sure that mariadb and mysql are NOT installed.
> > 2)  Delete residual libs
> >          rm -rf /var/lib/mysq*
> > 3)  yum install -y mariadb-server
> >     installed mariadb-5.5.50-1.el7_2.x86_64 from the Centos
'updates'
> > repository
> > 4)  systemctl start mariadb
> > 5)  systemctl status mariadb
> >
> > All works well until step 5, where I get the result:
> >
> >
> >
> > # systemctl status mariadb
> > ? mariadb.service - MariaDB database server
> >    Loaded: loaded (/usr/lib/systemd/system/mariadb.service; enabled;
vendor
> > preset: disabled)
> >    Active: failed (Result: exit-code) since Thu 2016-09-22 17:46:38
PDT;
> > 4min 3s ago
> >  Main PID: 16261 (code=exited, status=1/FAILURE)
> >
> > Sep 22 17:46:37 goat.xxx.com mysqld_safe[16261]: /usr/bin/mysqld_safe:
line
> > 139: /var/log/mysqld.log: Permission denied
> > Sep 22 17:46:37 goat.xxx.com mysqld_safe[16261]: /usr/bin/mysqld_safe:
line
> > 183: /var/log/mysqld.log: Permission denied
> > Sep 22 17:46:37 goat.xxx.com mysqld_safe[16261]: touch: cannot touch
> > `/var/log/mysqld.log': Permission denied
> > Sep 22 17:46:37 goat.xxx.com mysqld_safe[16261]: chown: cannot access
> > `/var/log/mysqld.log': No such file or directory
> > Sep 22 17:46:37 goat.xxx.com mysqld_safe[16261]: chmod: cannot access
> > `/var/log/mysqld.log': No such file or directory
> > Sep 22 17:46:37 goat.xxx.com systemd[1]: mariadb.service: main process
> > exited, code=exited, status=1/FAILURE
> > Sep 22 17:46:38 goat.xxx.com systemd[1]: mariadb.service: control
process
> > exited, code=exited status=1
> > Sep 22 17:46:38 goat.xxx.com systemd[1]: Failed to start MariaDB
database
> > server.
> > Sep 22 17:46:38 goat.xxx.com systemd[1]: Unit mariadb.service 
> entered failed
> > state.
> > Sep 22 17:46:38 goat.xxx.com systemd[1]: mariadb.service failed.
> > Hint: Some lines were ellipsized, use -l to show in full.
> > [root at goat ~]#
> >
> > My environment is a Centos 7 installation running as a VM inside 
> Vmware, and
> > is not a production system.  It is updated nightly.
> >
> > When I examined the files alluded to,
> >
> >   /var/log/mysqld.log
> >  does not exist
> > but
> >   /var/log/mariadb/mariadb.log
> >  exists and is empty.
>
>Hi David,
>
>What do these commands show?
>
>     ls -lF /var/log/mysqld.log
>     ls -lZ /var/log/mysqld.log
>     getenforce
>
>Best regards,
>--
>Charles Polisher

Charles
As the original note pointed out, /var/log/mysqld.log does not exist, 
as shown by these two commands:

[root at goat ~]# ls -1F /var/log/mysqld.log
ls: cannot access /var/log/mysqld.log: No such file or directory
[root at goat ~]# ls -1Z /var/log/mysqld.log
ls: cannot access /var/log/mysqld.log: No such file or directory
[root at goat ~]#

HOWEVER, /var/log/mariadb/mariadb.log does exist and is empty
[root at goat ~]# ls -l /var/log/mariadb/*
-rw-r----- 1 mysql mysql 0 Sep 23 10:08 /var/log/mariadb/mariadb.log

David

PS:  I forgot to say that SELinux is disabled:

[root at goat ~]# getenforce
Disabled
[root at goat ~]#

On Sep 23, 2016, at 11:20 AM, david <david at daku.org>
wrote:
> 
> As the original note pointed out, /var/log/mysqld.log does not exist, as
shown by these two commands:
> 
> [root at goat ~]# ls -1F /var/log/mysqld.log
> ls: cannot access /var/log/mysqld.log: No such file or directory
> [root at goat ~]# ls -1Z /var/log/mysqld.log
> ls: cannot access /var/log/mysqld.log: No such file or directory
> [root at goat ~]#
In that case, the error means that user mysql doesn?t have write access to
/var/log, which is as it should be, since my EL7 box says:

    $ ls -ld /var/log
    drwxr-xr-x. 21 root root 4096 Sep 18 03:28 /var/log/

You need to configure MariaDB to write its logs to a directory it does have
write access to, by default /var/log/mariadb.  MariaDB is configured this way
here:

    [mysqld_safe]
    log-error=/var/log/mariadb/mariadb.log

If your /etc/my.cnf says something else, then the poster up-thread who guessed
that you have an old my.cnf file nailed it.  Unless you have precious
configuration info in it, I?d replace it with /etc/my.cnf.rpmnew if it exists,
or just remove it and reinstall the RPM.

(Incidentally, you mistranscribed the commands we gave you.  It?s an ?l?, not a
?1? in ls -lF and ls -lZ, though it doesn?t matter in this case.)