Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.
Original topic: 集群只有一个tidb节点，该节点宕机了可以直接强制缩容不？
【TiDB Usage Environment】Production Environment / Testing / PoC
【TiDB Version】5.3.3
【Reproduction Path】What operations were performed when the issue occurred
【Encountered Issue: Problem Phenomenon and Impact】The cluster has only one TiDB node, and now the machine has failed. How can this be handled? Thank you!
【Resource Configuration】Go to TiDB Dashboard - Cluster Info - Hosts and take a screenshot of this page
【Attachments: Screenshots/Logs/Monitoring】
Technically, it is possible to scale down and then scale back up later, but after scaling down, since there is no TiDB server, the business will be unable to connect.
Okay, I’ll give it a try, thank you!
TiDB nodes can be scaled down first and then scaled up.
Initially, TiKV had 3 nodes: 1.3, 1.4, and 1.5. After forcibly scaling down nodes 1.4 and 1.5, only 1.3 was left. After scaling down and then scaling up the TiDB nodes again, the TiDB nodes couldn’t start and kept logging connection information with 1.4 and 1.5, but at this point, 1.4 and 1.5 no longer exist. How can this be resolved to get the TiDB nodes up and running? Thank you!
Initially, there were 3 TiKV nodes: 1.3, 1.4, and 1.5. Nodes 1.4 and 1.5 were forcibly scaled down, leaving only 1.3. After scaling down and then scaling up the TiDB nodes again, the TiDB nodes fail to start and continuously log connection information with 1.4 and 1.5, even though 1.4 and 1.5 no longer exist. How can this be resolved to get the TiDB nodes up and running? Thank you!
Force scale-in, then scale-out again.
TiKV with 3 replicas cannot be scaled down to fewer than 3 replicas, as it may result in data loss. The specific recovery operations are as follows:
Online Unsafe Recovery Documentation | PingCAP Documentation Center
Your 3-node scale-down seems to be stuck. You should expand TiKV back first.
TiKV cannot be used this way. When there are only 3 nodes in TiKV, forcibly scaling down 2 nodes requires special recovery. Refer to the link posted above.
Did you confuse the TiDB node with the TiKV node?
Can you play like this in a production environment? There’s a high chance of data loss. If you have a backup, redeploy a new system for recovery.
I feel that the description is problematic. In the future, it would be better to display the cluster so that we can determine whether it can be scaled down.
If three TiKV nodes are forcibly scaled down to two, the cluster has already lost data. How can TiDB still connect…
Are you trying to test the robustness of the cluster? Reducing and expanding TiDB when only one TiKV is left seems inherently risky. The events that could occur at this time are unpredictable.
Your title and opening content are about TiDB, but after scaling down, you mentioned TiKV. This joke is too much.
|
username: Hacker_qeCXjAN2
|
username: 
|
username: 
|
username: 
|
username: 
|
username: 
|
username: 
|
username: 
|
username: 
|
username: 
|
username: 
|
username: 
|
username: