File queue in Bifrost (1)

Want to make APP Same thing as WeChat , Can run small programs smoothly ？ | Experience will send you to Xinjiang 、 Huawei 、 Cherry keyboard ！>>>

In order to solve the problem of table synchronization, set parallel synchronization , Share one Binlog Parsing thread , Because of some synchronization error and other reasons Binlog Resolve the blocking situation .

stay  Bifrost v1.2.1  Version start , Bifrost Added a file queue function , When table synchronization is blocked , Auto start file queue , Does not affect Binlog analysis , It does not affect other table data synchronization .

To put it simply As long as all synchronization is fast enough , Just use the pure memory queue , As long as one synchronization fails , To make sure you make a mistake , No effect Other tables are synchronized , Current synchronization thread , A file queue that is automatically enabled . 

File queues , It's a synchronization problem , To keep pace with Binlog When analyzing the speed , Automatically activated , You don't need a human interface to click to start . At present Management system interface Points provide methods Let the user click force start !

File queue configuration

# Whether to turn on the file queue function  true|false
file_queue_usable=true

# Count whether the file queue is started or not , Unit millisecond 
file_queue_usable_count_time_diff=5000

#file_queue_usable_count_time_diff  The number of times the memory queue has been filled in time 
file_queue_usable_count=10

If the configuration file_queue_usable = false , No matter what , Does not enable file queuing

By default , The file queue is in 5000ms Inside The memory queue is continuously full 10 Time , Auto start , It doesn't need any human operation in the management interface

File queues , It is started according to the synchronization configuration of the table , in other words , If a table is configured with 10 Synchronization , The first configuration synchronization is to synchronize to Redis, Syncing to Redis Synchronization is slow in synchronization , In the case of memory queue blocking , A file queue is enabled , For buffering Synchronize to the current table to Redis