cluster show health v5.6
Checks the health of the EDB Postgres Distributed cluster
Synopsis
Performs various checks such as if all nodes are accessible and all replication slots are working.
Please note that the current implementation of clock skew may return an inaccurate skew value if the cluster is under high load while running this command or has large number of nodes in it.
pgd cluster show health [flags]
Examples
With a node down
In this example, a 3 node cluster has two nodes up, bdr-a1 and bdr-c1, but one node, bdr-b1, down.
$ pgd cluster show health
Output
Check Status Message ----- ------ ------- ClockSkew Critical Clockskew cannot be determined for at least 1 BDR node pair Connection Critical The node bdr-b1 is not accessible Raft Warning There is at least 1 node that is not accessible Replslots Critical There is at least 1 BDR replication slot which is inactive Version Warning There is at least 1 node that is not accessible
With unsynchronized system clocks
In this example, a 3 node cluster has all nodes are up but the system clocks aren't in synchronized.
$ pgd cluster show health
Output
Check Status Message ----- ------ ------- ClockSkew Warning At least 1 BDR node pair has clockskew greater than 2 seconds Connection Ok All BDR nodes are accessible Raft Ok Raft Consensus is working correctly Replslots Ok All BDR replication slots are working correctly Version Ok All nodes are running same BDR versions
Cluster functioning correctly
In this example, a 3 node cluster has all nodes are up and the system clocks are in synchronized.
$ pgd cluster show health
Output
Check Status Message ----- ------ ------- ClockSkew Ok All BDR node pairs have clockskew within permissible limit Connection Ok All BDR nodes are accessible Raft Ok Raft Consensus is working correctly Replslots Ok All BDR replication slots are working correctly Version Ok All nodes are running same BDR versions