Causes | Symptoms | Countermeasures | ||
Category 1 | Category 2 | |||
Command will not be executed normally as the conditions to add or detach a node are not satisfied | Number of nodes already participating in a cluster does not match the number of nodes constituting a cluster | [Server] 180055 [Description] Status does not allow cluster addition/detachment [Format] Number of nodes already participating in a cluster, number of nodes constituting a cluster of the current cluster [Check] gs_stat /cluster/activeCount ! = /cluster/designatedCount |
Check whether a failure has occurred in a node within the cluster. If a failure has occurred, first try to see if the node can be restored in the cluster. If it cannot be restored, get ready a new node separately, and let the node join the node. After this task, check the gs_stat to see whether the number of nodes already participating in a cluster (activeCount) matches the number of nodes constituting a cluster (designatedCount) and then execute a cluster addition/detachment command again. |
|
The number of nodes trying to join the cluster is more than the number of nodes constituting the cluster | [Server] 180045 [Description] Failure to compose cluster as the number of nodes exceeds the number of nodes constituting the cluster [Format] Number of nodes constituting a cluster of the current cluster and master address in the node to be added |
Status does not allow new nodes to join as the cluster being composed has already reached the upper limit (= number of nodes constituting the cluster). If you want to further increase the number of nodes constituting a cluster, execute a gs_appendcluster command on the node to be added to the cluster. |
||
By allowing the target node to leave the cluster, the current cluster will lose data or the cluster will be dissolved. | [Server] 180030 [Description] Node cannot leave the cluster [Format] Reason why a node cannot leave a cluster (cluster cannot maintain a majority of the nodes when a target node leaves the cluster, and data loss occurs when even only one node leaves the cluster) |
If there is a risk of data being lost or the cluster being dissolved due to the execution of a gs_leavecluster command, the node will not be allowed to leave the cluster. If you want to force a node to leave the cluster, append the --force option before executing the command. However, in this case, do not append the --force option where possible as the cluster will be reset and data may be lost. | ||