Cluster Operations
Operations related to the cluster are available on both the list and detail pages, and the objects of operation are the cluster and cluster nodes.
Start, Stop and Delete the Cluster
Taking shutting down the service as an example
The status of the service will dynamically change during startup, shutdown, and restart operations until the operation is completed. This dynamic change is also shown in the cluster list.
After the deletion operation, the cluster will be removed from the list.
Currently, batch operations are not supported, and the existing batch selection will default to selecting the first item for operation.
Please be careful with the above operations, which may have a significant impact on clusters already in production use. This operation is more friendly to newly built or data-free clusters.
Cluster Resizing
Add Nodes
The node configuration here refers to the current node information.
Please note that currently only expansion is supported, i.e., adding cluster nodes, and reduction of nodes is not supported.
The minimum limit is set to the original number of cluster nodes for the number of cluster nodes after expansion. The number of new nodes cannot exceed the node quota. The allocated node quota is sufficient to support your cluster use. If the quota requirement is too high, please contact the account manager to adjust the node resource quota.
The billing type is the billing type selected when the cluster was initially created. If there is a renewal or change in the billing method, please check and modify it in the console user center.
Change Specifications
The resized nodes here refer to the current data nodes, excluding dedicated master nodes.
The prerequisites for changing specifications are to ensure that the state of the cluster is normal (green) and that the index includes at least one copy; it is recommended to operate during the off-peak period of the business.
When performing node resizing operations, the health status of the cluster will be checked. If the status is normal (green), perform this node resizing and restart node operations; if the status is not normal (green), it will continue to wait and check the cluster status until it is normal (green). If the time is too long or a resizing failure occurs, please contact technical support.
Please note that currently only upgrades are supported, i.e., upgrading node CPUs and memory, and downgrades are not supported.
Modify Disk Size
The nodes where disk size is changed are the current data nodes, excluding dedicated master nodes.
Disk size modifications are made in increments of 100G, with the maximum disk size being 8000G.
Please note that shrinking disk size is not supported.
Cluster Configuration Management
The cluster overview on the details page includes the Parameter Management operation
Open parameter management, where parameters are listed, including parameter names, parameter values, application scopes, and parameter descriptions. Additionally, the function of parameter name search for quick retrieval and modification is provided.
It is worth noting that for many of the listed parameters, we have already made reasonable initializations according to the selected cluster node configuration. Unless there is a business need to modify certain parameters, modification is not recommended in other cases. Moreover, modified parameter values must comply with the conventions, otherwise, there may be adverse effects on the cluster. After modifying parameters, the cluster service needs to be restarted to take effect.
For modification of parameters that are not listed, most of them can be dynamically modified through the API. We also advocate using the API as much as possible to dynamically modify configuration parameters.
Linear Cluster Restart
Assuming that the production environment needs to restart the cluster under certain circumstances, but it requires that normal service use not be affected. The linear restart of the cluster is particularly important in this case.
The node list in the details provides the operation of single node restart.
When performing a single node restart operation, there will be a prompt after a successful restart.
It’s worth noting that, after a single node restarts successfully, the cluster needs some time to recover to a healthy state, after which you can operate on the next node
Methods to check whether the cluster has returned to a healthy state:
* Refresh the cluster information module in the cluster overview, make sure to refresh to reobtain the cluster status
Refresh once, if the status turns yellow, you can stop refreshing and wait for the cluster to automatically recover to a green healthy state. If the status is green, it means the cluster has returned to a healthy state; if the status disappears, continue refreshing at intervals until the cluster status appears. When the cluster is in a green state, the next node operation can be performed.
* Get real-time cluster status through the API `/_cluster/health` until the cluster status is restored to a healthy state.
Node Log Query
The node list on the details page includes the Node Log operation
Click to view the corresponding log.
kibana Password Reset
If you forget your kibana password, you can reset the password in the console, the kibana account currently does not support modification. If you forget the account, please contact the account manager for retrieval.