A week ago I wrote a post about promoting a mysql slave as master. Today I will write topology refactoring using classic replication commands. Context: Initial state: dbfoo1(M) – dbfoo2(MS) – dbrfoo3(S) Desired state: dbfoo1(M) – dbfoo2(S) \ dbfoo3(S) Let’s get to work: stop slave on dbfoo3 MariaDB [(none)]> stop slave; check the “freezing point” … Continue reading Topology refactoring using classic MySQL replication commands – version 1 →
I assume you you have a high availability (master – slave topology) setup. I will be brief and technical on this one :) Context: dbfoo1 is master and dbfoo2 is slave – dbfoo1(M) – dbfoo2(S) dbfoo2 has to be promoted to master – dbfoo2(M) – dbfoo1(S) note that in some cases the slave requires a … Continue reading mysql: Promoting a slave to master →
It happened many times to need a quick access to the remote console of a server like HP’s ILO and not be able to open a web browser just to access it. So let access it via SSH. Here goes. 2. Enter your ILO admin account and password. After that you will see the ILO prompt. hpILO-> 3. … Continue reading Access HP’s ILO remote console via SSH →
It happened today… why? Well, I was testing parallel replication (slave_parallel_threads) and after setting the number of threads to zero and when started the slave I got this wonderful error message: MariaDB [(none)]> show slave status\G *************************** 1. row *************************** Slave_IO_State: Waiting for master to send event Master_Host: master.example.com Master_User: user Master_Port: 3306 Connect_Retry: 60 … Continue reading Relay log read failure: Could not parse relay log event entry →
Sometimes connecting to the collectd unixsocket is not possible giving a connection refused. This can be fixed manually by removing manually the socket file and then restart collectd collectd does not create a new socket file when there is already such a file. But the old file, which is not connected to any process certainly … Continue reading How to fix “ERROR: Connecting to daemon at /var/run/collectd-unixsock failed” →
Why?! To avoid a lengthy warmup period after restarting the server, particularly for instances with large InnoDB buffer pools. Things to be added to /etc/mysq/my.cnf: # Shortened warm-up times with a preloaded InnoDB buffer pool innodb_buffer_pool_dump_at_shutdown=ON innodb_buffer_pool_load_at_startup=ON Basically when MySQL receives the TERM signal will dump the InnoDB buffer pool /data/db/mysql/ib_buffer_pool and it will load it … Continue reading mysql: preload InnoDB buffer pool →
I can't memorize much, so I take notes!