Redis queue, RDB learning

Channel publishing and message subscription

Release the news	publish queue value
Listen to the message	subscribe queue
Monitor multiple channels at once	psubscribe queue
Manage view therefore Release and monitor	pubsub queue

RDB 

Persistence ： Store the data in a device that will not disappear after power failure , Usually on the hard disk

Common persistence methods ：

1. Master-slave ： By saving and persisting from the server , Such as mongoDB Of replication sets Set up
2. journal ： Operation generates relevant logs , And recover data through logs
3. couchDB The data content is not modified , Just add , Then the file itself is a log , No loss of data

RDB Snapshot persistence

rdb How it works ：

  every other N minute perhaps N Time After writing operation , From memory dump Data formation rdb file , Compress Put it in the backup Catalog

RDB Advantages and disadvantages ： Strong recoverability to data , Fast recovery , But in some cases, data will be lost

Parameter Introduction

save 3600 1 namely 3600 Seconds later, at least 1 Keywords send changes

save 300 100 Must be 300 Seconds later 10 Keywords changed

save 60 10000 Must be 60 Seconds later, at least 10000 Keywords changed

（ These three options After comment rdb Disabled ）

 

stop-writes-on-bgsave-error stay rdb Error in file export , At the same time, prevent further writing redis Ensure the consistency of data

 

 rdbcompression    rdb The file data is too large , Compress on Export

 rdbchecksum  from rdb Import data into memory , Check rdb The integrity of

dbfilename    Derived rdb file name

 

  Derived rdb The file is in that directory

Be careful ：

stay 2 Between savepoints , Power failure will result in loss 1-N Minute data   therefore redis Added aof (append ohly file)

Please look at the next post -->Redis aof Log persistence