Issue with mariadb update

Support/Development for MySQL, MariaDB, and other database systems
slashwalker
New Forum User
New Forum User
Posts: 3
Joined: Mon Oct 14, 2013 3:57 am
Location: Germany

Issue with mariadb update

Unread post by slashwalker »

Hello,
a while ago I setup a new Server with Centos 6.4 and mariadb 5.5.29
Now I tried to update to mariadb 5.5.32
But /etc/init.d/mysqld restart always fails with:
131014 09:49:56 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
131014 09:49:56 mysqld_safe WSREP: Running position recovery with --log_error= --pid-file=/var/lib/mysql/myserver.de-recover.pid
131014 9:49:56 [ERROR] WSREP: Query cache is not supported (size=268435456 type=1)
131014 09:50:09 mysqld_safe WSREP: Failed to recover position:''

I have no Cluster or something. Selinux is on, which worked fine all the time.

I even tried to dump all databases and completly remove mariadb and do a fresh install. Same error.
slashwalker
New Forum User
New Forum User
Posts: 3
Joined: Mon Oct 14, 2013 3:57 am
Location: Germany

Re: Issue with mariadb update

Unread post by slashwalker »

I now cloned the machine, so I can test without worrying about downtimes.
Turns out that SELinux is the issue. If I run setenforce 0, I can do /etc/init.d/mysqld start
mariadb 5.5.29 works with SELinux, so I wonder why 5.5.32 doesn't?

I ran:
sealert -a /var/log/audit/audit.log > /root/whatswrong

And did:
grep chown /var/log/audit/audit.log | audit2allow -M chown
semodule -i chown.pp
grep mysqld /var/log/audit/audit.log | audit2allow -M mysqld
semodule -i mysqld.pp
grep rm /var/log/audit/audit.log | audit2allow -M rm
semodule -i rm.pp
grep mktemp /var/log/audit/audit.log | audit2allow -M mktemp
semodule -i mktemp.pp

This works for now on the testing machine. But I'm not sure if this is to permissive?
prupert
Forum Regular
Forum Regular
Posts: 573
Joined: Tue Aug 01, 2006 2:45 pm
Location: Netherlands

Re: Issue with mariadb update

Unread post by prupert »

Are you sure you haven't installed the cluster version?
# rpm -qa | grep -i galera

And haven't configured any cluster settings?
# grep -rin wsrep /etc/my*

If both commands yield no output, you have not.
Lemonbit Internet Dedicated Server Management
slashwalker
New Forum User
New Forum User
Posts: 3
Joined: Mon Oct 14, 2013 3:57 am
Location: Germany

Re: Issue with mariadb update

Unread post by slashwalker »

prupert wrote:Are you sure you haven't installed the cluster version?
# rpm -qa | grep -i galera

And haven't configured any cluster settings?
# grep -rin wsrep /etc/my*

If both commands yield no output, you have not.
Both show no output.
rpm -qa | grep maria shows:

mariadb-libs-5.5.32-12.el6.art.x86_64
mariadb-5.5.32-12.el6.art.x86_64
mariadb-server-5.5.32-12.el6.art.x86_64

And as said, 5.5.29-11 worked with SELinux (enforcing, targeted) without changing any policy.

Here my my.cnf:

Code: Select all

[client]
default-character-set= utf8
[mysql]
default-character-set= utf8

[mysqld]
user=mysql
collation-server = utf8_unicode_ci
init-connect='SET NAMES utf8'
character-set-server = utf8
datadir=/var/lib/mysql
socket=/var/lib/mysql/mysql.sock
tmpdir=/tmp
query_cache_type=1
query_cache_limit=10M
query_cache_min_res_unit=1K
query_cache_size=256M
table_cache=500
max_connections=300
performance_schema=off
symbolic-links=0
max_heap_table_size=512M
tmp_table_size=512M
skip-name-resolve
thread_cache_size=10

[mysqld_safe]
log-error=/var/log/mysqld.log
pid-file=/var/run/mysqld/mysqld.pid

Confs in my.cnf.d are untouched.
Post Reply