Tdengine biweekly selection of community issues | phase III

Hi, Hello everyone , To prevent too many common problems from bothering users , Easy for users to learn 、 Absorb experience , We will select the common problems encountered by users in the community recently 、 Summary , Share with you in the form of a column . This is the third issue of this column , The content shared is as follows ：

Content preview

1. Why do I write data in batches , Timestamp use NOW Function splicing will result in data loss ？
2. How to smoothly expand the capacity of a cluster ？
3. After cluster expansion ,DNode Status as Offline How to investigate ？
4. How to smoothly shrink a cluster ？
5. Novice benefits ： Quick start  TDengine  modeling .

1. Why do I write data in batches , Timestamp use NOW Function splicing will result in data loss ？

To answer this question, we first need to emphasize a concept ,TDengine As a Time series database （Time-Series Database）, The first timestamp field acts as a primary key , Construction of memory index 、 The storage of disk data is closely related to it , There must be no duplicate timestamp .

NOW function （ as well as NOW keyword ） Return the current time of the client . When performing batch writes , If the values given by the timestamp in the first column are NOW, It is indistinguishable under the time level of milliseconds by default in the database , Repeated timestamps written later will be lost or updated , The specific logic for handling repeated timestamps is provided in TDengine China Construction Library （Database） At the time of the Update Parameter determination .

2. How to smoothly expand the capacity of a cluster ？

To expand the cluster, you need to add nodes manually . The specific operation steps are as follows ：

1. After the first node is successfully deployed , Install nodes to be added in sequence TDengine Server program , edit /etc/hosts, Add the domain name resolution of all nodes in the cluster （ If deployed DNS server, There is no need to edit ）.
2. The configuration file of the first node taos.cfg Copy to the node to be added , Will be one of the fqdn The parameter is changed to the node itself fqdn (hostname).
3. Start the node to be added taosd service , Then enter taos The command line executes the following statement to add the node to be added into the cluster ：CREATE DNode ‘Node2:6030’ .
4. perform SHOW DNodes Check whether the status of the data node is Ready. by Ready On behalf of successful cluster expansion .

3. After cluster expansion ,DNode Status as Offline How to investigate ？

After the new node joins the cluster normally , The data node list will show that the node is in Ready state . If the node status is Offline, Check according to the following contents ：

1. View the node taosd Started or not 、 Whether the firewall is closed ;
2. Confirm whether the data folder of the node to be added is empty ;
3. Check all nodes /etc/hosts Whether the domain name resolution is complete 、 It works （ All nodes need to be resolved , Include arbitrator）;
4. The node firstEP、fqdn Whether the parameters are configured correctly .

4. How to smoothly shrink a cluster ？

Cluster capacity reduction is achieved through DROP DNode Delete nodes to achieve , Stop specifying directly DNode Of taosd The process will only let this node Offline, But it cannot complete the effect of volume reduction . The specific operation steps are as follows ：

1. Get into taos Command line , perform SHOW DNodes Check the status of the node to be deleted , if ready Then proceed to the next step ;
2. Delete the specified node from the current cluster DROP DNode ‘Node2:6030’ perhaps DROP DNode <DNodeId>.
3. perform SHOW DNodes View the cluster node information , Finally, confirm that the information of the node to be deleted has disappeared from the list .

Be careful ： Before deleting nodes , The cluster will DNode Data is migrated to other nodes , The completion speed depends on the bandwidth of the network 、 On disk IO, Never stop the node to be deleted before it is completed taosd service .

5. Novice benefits ： Quick start TDengine modeling .

In the process of community support , You can find many novice partners deploying TDengine I don't know how to further experience , Our suggestion is to run the sentences in the official website documents , But there are many documents , In order to facilitate community partners to get started quickly , We condense the example model of the official website document 、 I summed it up , Let users know as soon as possible TDengine Modeling methods ：[ Introduction to modeling ]

complimentary close

Thank you for reading the selected questions in this issue , I hope the above is helpful to you . You are using TDengine This kind of TIme-Series Database In the process of , If you encounter a difficult problem or have a problem with TDengine There are suggestions for improvement , Welcome to the  GitHub  Let's talk about Issue , The more detailed the content, the better ~ Last , Thank you for using TDengine！

---

Want to know more TDengine Database Specific details of , Welcome to GitHub View the relevant source code on .