Question
An index request was made for a folder that has successfully been indexed for months prior to this latest error. Examining the indexing results in the Digital Hive Control Center, this error appears.
How can I resolve this issue so that the indexing completes successfully?
Answer
This error can be caused by not having enough free disk space on the Digital Hive server. When Elasticsearch, the search engine within Digital Hive, starts an indexing process, at least 20% of the total disk volume must be available. For example, if the total disk space is 100GB, then at least 20GB must be free and available to the system.
To confirm that this is the issue, open the theia.yyyy-yy-dd.log file located in the <install_location>/app/node1/tomcat/logs directory. Search for the term watermark and an error like this should be present:
blocked by: [TOO_MANY_REQUESTS/12/disk usage exceeded flood-stage watermark, index has read-only-allow-delete block]
If this error is present, free up some disk space on the Digital Hive server. Two file directories that could be consuming excessive disk space are:
- <install_location>/app/node1/tomcat/logs
- <install_location>/ContentStore/postgresql-backups
It is safe to delete all the log files previous to the current date in the logs directory, unless there is an internal need to maintain more than the current day of log files. The Postgres backups can also be deleted, but the recommendation is to maintain at least a couple of backup files in case they are needed.