Hi
Just in the process of attempting an upgrade from percona 5.5.35 server to percona cluster 5.5.34 and it keeps failing on boot.
I’m just creating a single node at the moment, with the view of creating 2nd nodes once initial cluster is running.
I’m running the init script to start the server and i get the following error in the error log:
140110 11:34:16 mysqld_safe mysqld from pid file /var/run/mysqld/mysqld.pid ended
140110 11:43:36 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
140110 11:43:36 mysqld_safe Skipping wsrep-recover for empty datadir: /var/lib/mysql
140110 11:43:36 mysqld_safe Assigning 00000000-0000-0000-0000-000000000000:-1 to wsrep_start_position
140110 11:43:36 [ERROR] WSREP: Query cache is not supported (size=16777216 type=0)
140110 11:43:36 [ERROR] Aborting
I’ve got the following config in my.cnf:
[mysqld]
basedir=/usr/local/mysql
datadir=/usr/localmysql/data
pid-file=/var/run/mysqld/myqsqld.pid
log-error=/var/log/mysql/mysqld.err
plugin-load=thread_pool.so
thread_pool_size=32
innodb_buffer_pool_size=216G
innodb_log_file_size=512M
innodb_log_buffer_size=16M
innodb_flush_log_at_trx_commit=2
innodb_flush_method=O_DIRECT
innodb_file_per_table
transaction-isolation=READ-COMMITTED
#Added for cluster
#innodb_data_home_dir=/usr/local/mysql/data
#innodb_log_group_home_dir=/usr/local/mysql/data
max_connections=500
log-bin=mysql.bin
#Changed from MIXED for cluster
binlog-format=ROW
server-id=1
#ignore-builtin-innodb
#plugin-load=innodb=ha_innodb_plugin.so
default-storage-engine=InnoDB
thread_cache_size=24
tmp_table_size=32M
max_heap_table_size=32M
#slow_query_log_file=/storage/logs/mysql/slow-query/sql1.log
#slow_query_log=1
long_query_time=3
#log-slow-admin-statements
#log-queries-not-using-indexes
expire_logs_days=7
net_read_timeout=5000
net_write_timeout=5000
max_allowed_packet=128M
group_concat_max_len=32M
binlog_cache_size=1M
table_open_cache=4096
table_definition_cache=4096
join_buffer_size=1M
innodb_lock_wait_timeout=700
skip-name-resolve
#query_cache_size=0
bind_address=0.0.0.0
thread_cache_size=50
#Cluster Variables
wsrep_provider=/usr/local/mysql/lib/libgalera_smm.so
wsrep_cluster_address=gcomm://
innodb_locks_unsafe_for_binlog=1
innodb_autoinc_lock_mode=2
wsrep_node_address=192.168.4.101
wsrep_sst_method=xtrabackup
wsrep_cluster_name=test_cluster
wsrep_sst_auth=“sstuser:***********”
As you can see form conf i’ve tried commenting out existing query cache size.
From the error its almost like elements of my conf are being skipped as its trying to pick up the data dir from /var/lib/mysql (tried creating this dir and simlinking it over) to correct datadir path but to no avail.
I’ve tried adding the basedir and datadir paths directly to mysqld_safe and /etc/init.d.mysql.server but this doesn’t work either.
I’m positive i’ve got a conf issue here but i’m a bit stumped as to exactly what i need to do to rectify it.
Thanks in advance
Matt Smith
