Over the course of Solr's implementation, there have been many discussions regarding antivirus software and its interaction with Solr.
Ektron Support sometimes sees issues where antivirus software prevents Solr from crawling properly. This is because when Solr launches a crawl, it accesses and/or writes to multiple files. When performing these tasks, Solr places a lock on the files. This multiple file locking mechanism causes the antivirus application to think Solr is a virus and terminates the crawl.
This can cause you to see only two Java.exe processes running in Task Manager or a 500 error within an unknown abort cause. You may also see lock errors in the ManifoldCF logs.
lockmanager.FileLockObject; (Document cleanup stuffer thread) - Attempt to set file lock 'C:\Program Files (x86)\Ektron\Search2.0\ManifoldCF\core\.\crawldb\syncharea\737\563\lock-_Cache_JOBSTATUSES.lock' failed: Access is denied
We recommend disabling antivirus software on the Solr server due to its negative effects. If that is not an option for security reasons, it may help to apply exclusions to the following folders. However, Support has seen varying results with this approach, as every environment is different.
- SearchX.0 found at DriveLetter:\Program Files(x86)\Ektron\Search X.0
- The Java directory found at C:\Program Files\Java\jdk1.X.0_YY\
Or if Redhat OpenJDK is used
- C:\Program Files\RedHat\java-1.X.0-openjdk-Y\
After whitelisting the above, clear the crawl database and reregister.
If these exclusions do not work, we recommend disabling the antivirus software or working with your antivirus provider to investigate further.