We recently observed a disk full event on an-coord1003
The largest consumer of disk space was the /var/log/hive directory.
It appears that we had at least 5 months' worth of hive-metastore.log.YY-MM-DD files in this directory.
There is a log4j configuration file at: /etc/hive/conf/hive-log4j2.properties
This seems to suggest that we would be keeping 14 days' worth of log files, but perhaps this isn't being applied properly.
We should ensure that we have suitable housekeeping in place for these hive metastore logs.
