Had to do several reboots of the Cacti VM tonight to do some NFS mount fixes, and noticed graphs weren’t updating and the device list was returning zero rows.  Immediately my thought was database, and this was confirmed in cacti.log

2017-09-13 22:00:00 - DBCALL ERROR: SQL Assoc Failed!, Error:145, SQL:"SELECT status, COUNT(*) as cnt FROM `host` GROUP BY status"
2017-09-13 22:00:00 - DBCALL ERROR: SQL Assoc Failed!, Error: Table './cacti/host' is marked as crashed and should be repaired

Also in var/log/mysqld.log:

170913 22:03:00 [ERROR] /usr/libexec/mysqld: Table './cacti/host' is marked as crashed and should be repaired

This blog pointed me to the easy fix:

mysqlcheck -u cactiuser -p --auto-repair --databases cacti