Set the Spread Daemon Timeout Back to the Default Value
The default value for the Spread Daemon timeout is 8 seconds when the IP addresses of all cluster nodes start with the same two bytes. If all of IP addresses do not start with the same two bytes, the default becomes 25 seconds.
When you know your network or nodes may be unable to respond for a specific amount of time due to network delays or temporary pauses of a VM, you can increase the spread timeout period to longer than this time using the SET_SPREAD_OPTION function (available starting in Vertica 9.2.1-0).
You can later set the Spread Daemon timeout back to its default value using the same function setting the 'TokenTimeout' to zero.
Example:
dbadmin=> SELECT * FROM spread_state; node_name | token_timeout --------------------+--------------- v_test_db_node0001 | 16000 v_test_db_node0002 | 16000 v_test_db_node0003 | 16000 (3 rows) dbadmin=> SELECT SET_SPREAD_OPTION( 'TokenTimeout', '0'); NOTICE 9003: Spread has been notified about the change SET_SPREAD_OPTION ---------------------------------------------------- Spread option 'TokenTimeout' has been set to '0'. (1 row) dbadmin=> SELECT * FROM spread_state; node_name | token_timeout --------------------+--------------- v_test_db_node0001 | 8000 v_test_db_node0002 | 8000 v_test_db_node0003 | 8000 (3 rows)
Vertica set the default value to 8 seconds because all of the node IP addresses start with the same two bytes.
dbadmin=> SELECT node_name, dbadmin-> SPLIT_PART(node_address, '.', 1) || '.' || SPLIT_PART(node_address, '.', 2) first_2_bytes dbadmin-> FROM nodes; node_name | first_2_bytes --------------------+--------------- v_test_db_node0001 | 172.16 v_test_db_node0002 | 172.16 v_test_db_node0003 | 172.16 (3 rows)
Helpful Links:
https://www.vertica.com/docs/9.2.x/HTML/Content/Authoring/UsingVerticaOnAzure/AdjustingSpreadDaemonTimeouts.htm
Have fun!