恢復主數(shù)據(jù)庫通常不從從庫獲取數(shù)據(jù)，但當主庫宕機且無可用備份時，可按以下步驟操作：1.檢查從庫是否具備最新數(shù)據(jù)，運行SHOW SLAVE STATUS\G確認Seconds_Behind_Master為0且Last_Error為空；2.停止從庫複製線程並重置複製信息，使用STOP SLAVE和RESET SLAVE ALL命令；3.將原從庫配置為新主庫後，更新應(yīng)用連接設(shè)置並重新配置新從庫指向該主庫，使用CHANGE MASTER TO指定正確的二進制日誌文件和位置；4.創(chuàng)建新的複制用戶並授予相應(yīng)權(quán)限；5.避免自動增長鍵衝突，需調(diào)整auto_increment_offset和auto_increment_increment參數(shù)，如兩臺服務(wù)器分別設(shè)為offset 1和offset 2，increment均為2。整個過程需謹慎操作以防止數(shù)據(jù)不一致或服務(wù)中斷。

Restoring a master database from a slave's data isn't the usual flow in replication setups, but sometimes things go sideways—like when the master is down and there's no recent backup. In those cases, pulling data from a slave becomes your best bet.

Here's how to make it work without messing things up further.

Check if the Slave Has What You Need

Before jumping into recovery, confirm that the slave has the right data. Replication lag can be a sneaky issue—if the slave was behind when the master went down, you might be restoring outdated info.

• Run SHOW SLAVE STATUS\G on the slave.
• Look at Seconds_Behind_Master . If it's 0 or close to it, you're good.
• Also check for any replication errors under Last_Error .

If everything looks clean, move forward. Otherwise, you may need to dig deeper or find another source.

Stop Replication and Promote the Slave

Once you're sure the slave is up-to-date, you'll need to turn it into the new master.

• Stop replication threads with:

   STOP SLAVE;

• Then reset the replication info:

   RESET SLAVE ALL;

• Remove the replication user on the old master (if it's still accessible) or skip this step if the master is completely gone.

At this point, the former slave acts like a standalone database. You can start setting up new slaves to point to it.

Reconfigure Applications and New Replicas

Now that the old slave is the new master, update your application settings so they connect to the right place.

Also, if you're bringing up new slaves:

• Make sure they use the correct binary log file and position from the new master.
• Use CHANGE MASTER TO accordingly and start replication.

Don't forget to create a dedicated replication user again and grant the right privileges.

Watch Out for Auto-Increment Conflicts

One common gotcha when switching roles: auto-increment values. If you're planning to bring back the original master later as a slave, you'll want to avoid duplicate key issues.

• Set different auto_increment_offset and auto_increment_increment values on each server.
• For example, if you have two servers, set one to offset 1 and the other to offset 2, both with increment 2.

This helps prevent conflicts when writing to multiple masters or when promoting a slave temporarily.

That's basically it. It's not something you do every day, but knowing how to flip replication roles can save you in a pinch. Just keep an eye on replication status, reconfigure carefully, and plan for future failover scenarios.

以上是如何從從屬數(shù)據(jù)中恢復主人？的詳細內(nèi)容。更多資訊請關(guān)注PHP中文網(wǎng)其他相關(guān)文章！