Check MySQL status const String & check_query = "SHOW VARIABLES WHERE " "(Variable_name = 'log_bin' AND upper(Value) = 'ON') " "OR (Variable_name = 'binlog_format' AND upper(Value) = BinlogEventPtr & receive_event, MaterializeMetadata & metadata) { if (receive_event->type() == MYSQL_WRITE_ROWS_EVENT) { … } else if (receive_event->type() == MYSQL_DELETE_ROWS_EVENT) MaterializeMetadata & metadata) { if (receive_event->type() == MYSQL_WRITE_ROWS_EVENT) { WriteRowsEvent & write_rows_event = static_cast(*receive_event);

query © 2018 Percona. 7 Clickhouse Answers • 10x+ times space reduction compared to Raw Text Log Files High Compression (column store + LZ4) • Typically 100x faster than MySQL on Single com/Altinity/clicktail Created by my friends at Altinity Replaced HoneyComb “Sender” with Clickhouse Added Audit Log Support Meet ClickTail © 2018 Percona. 14 Installing ClickTail • curl -s https://packagecloud To run as a service © 2018 Percona. 15 MySQL Logs Primer General Query Log Binary Log Slow Query Log Audit Log © 2018 Percona. 16 MySQL Audit Logs to ClickHouse © 2018 Percona. 17 When

Separate log_dir and datadir. All storage types benefit from this. For both Percona and MySQL Server, it means setting up the parameters from Appendix A marked with either or <log storage> tmpdir /tmp /<log storage>/mysql_log lc‐messages‐dir /usr/share/mysql explicit_defaults_for_timestamp innodb_log_group_home_dir /<log storage>/mysql_log Best Practices for innodb_undo_directory /<log storage>/mysql_log innodb_buffer_pool_size 3GB 12GB innodb_thread_concurrency 0 innodb_temp_data_file_path '../../../<log storage>/mysql_log/ibtmp1:72 M:autoextend'

finalizes the installation for products that do not require configuration. It enables you to copy the log to a clipboard and to start certain applications, such as MySQL Workbench and MySQL Shell. Click Finish defining custom file paths for the error log, general log, slow query log (including the configuration of seconds it requires to execute a query), and the binary log. During the configuration process, click in this group should be limited and managed. Windows requires a newly added member to first log out and then log in again to join a local group. • Full access to all users (NOT RECOMMENDED). This option

Traffic Management Scaling with Large Number of Connections Routing Traffic to Right “Shard” Read-Write Splitting Load Management Avoiding “Dead” Nodes © 2018 Percona. 33 Options in MySQL © 2018 Group Replication © 2018 Percona. 40 MySQL Group Replication (New in 5.7) “Group of Peers” Write-Anywhere or Dedicated Writer Asynchronous Replication with Flow Control Conflicts Prevented through 2018 Percona. 42 Percona XtraDB Cluster Topology © 2018 Percona. 43 PXC/Galera Properties Write to Any Node Certification Based Replication Virtually Synchronous; Can ensure no stale reads

d/authkeys auth 1 1 crc  ha.cf 的配置 master 的 ha.cf 的配置 vim /etc/ha.d/ha.cf logfile /var/log/ha-log logfacility local0 keepalive 2 deadtime 30 warntime 10 initdead 60 udpport 694 ucast hacluster /usr/lib64/heartbeat/ipfail backup 的 ha.cf 的配置 vim /etc/ha.d/ha.cf logfile /var/log/ha-log logfacility local0 keepalive 2 deadtime 30 warntime 10 initdead 60 udpport 694 ucast auth 1 1 crc  ha.cf 的配置 master(dbserver1)的 ha.cf 的配置 vim /etc/ha.d/ha.cf logfile /var/log/ha-log logfacility local0 keepalive 2 deadtime 30 warntime 10 initdead 60 udpport 694 ucast

inlog 以及冲突检测时使用到主键以及事务的数据快照版本(gtid_set) • 比较规则。将事务所涉及的主键以及数据版本信息(gtid_set)跟write set 里面的信息进行比较: 主键不存在与write set中，不冲突。 主键存在，则比较事务版本信息，即比较gtid_set，如果是包含关系， 则不冲突。否则，冲突。 • 冲突检查 Db_name_1:table_name_3:key3:1-9578876! Key3����������� gtid_set 1-9578876��1-9578874, �����! write set: ! ��1-�������:! NODE A! NODE C! NODE B! Ta! gtid_excuted: 1-9578875! a,b,c! 1,1,1! 2,2 ���������sql)��! ��sql! mysql_parse! mysql_execute_comma nd! trans_commit_stmt! MYSQL_BIN_LOG::com mit! group_replication_trans_before_commit! waitTicket! ���������� ���������� ���������� ������paxos�

DeltaFIFO Local Storage Callbacks OnAdd OnUpdate OnDelete WorkQueue List/Watch - CRD - Pod Write Informer Worker ReadOnly • Informer：监听事件并触发回调函数的二级缓存工具包 • WorkQueue：事件合并、过滤、延时、限速 Operator CRD Host Path Volumes • 优点：读写延迟低 • 缺点：单点数据，容器漂移时 数据丢失 踩到的坑 • 现象：执行 docker 命令时，docker daemon 无响应，/var/log/messages 大量报错 libceph: osdxx 10.0.0.0:6812 socket closed (con state OPEN) 原因：libceph 触发了 linux 内核 的一个

1024M innodb_flush_log_at_trx _commit 1 0 0 代表日志只大约每秒写入日志文件并且日志文件 刷新到磁盘 ; 1 为执行完没执行一条 SQL 马上 commit; 2 代表日志写入日志文件在每次提交 后 , 但是日志文件只有大约每秒才会刷新到磁盘上 . 对速度影响比较大，同时也关系数据完整性 innodb_log_file_size 8M 512M innodb_buffer_pool_size 的 25% ，官方推荐是 innodb_buffer_pool_size 的 40-50%, 设置大 一点来避免在日志文件覆写上不必要的缓冲池刷新 行为 innodb_log_buffer_size 128K 64M 用来缓冲日志数据的缓冲区的大小 . 推荐是 8M ， 官方推荐该值小于 16M ，最好是 1M-8M 之间  设计合理的数据表结构：适当的数据冗余 设计合理的数据表结构：适当的数据冗余 执行情况，是否锁表，查看相应的 SQL SQL 语句 语句  设置 设置 my.cnf my.cnf 中的 中的 long-query-time long-query-time 和 和 log-slow-queries log-slow-queries 能 能 够 够 记录服务器那些 记录服务器那些 SQL SQL 执行速度比较慢 执行速度比较慢  另外有用的几个查询： 另外有用的几个查询：

master_host='master_db', master_user='sync', master_password='sync', master_port=3306, master_log_file='<主数据库查询得到的 File 值>', master_log_pos=<主数据库查询得到的 Positions 值>; 下面是我的脚本例子： change master to master_host='master_db' ', master_user='sync', master_password='sync', master_port=3306, master_log_file='mariadb-bin.000004', master_log_pos=789; /*开启从数据库复制*/ start slave; 最后可以通过 show slave status; 查看同步情况。 至此我们就建立了一个基于 Docker