There are several steps to determine if Solr is healthy and to keep it that way. It is important to monitor certain aspects of Solr, so problems can be resolved before they become show stoppers. See below for details.
- If you are using 9.0 - 9.1 Solr, are three Java processes running on the Solr machine? Or if 9.2 is used are there four Java processes? If not, try restarting Ektron Solr Process Manager to see if all three or four processes come up. If they do not, first check for antivirus and that the proper whitelisting has been done. If there are extra Java processes, see this KB.
- The second most common issue with Ektron Solr installations is lack of allotted memory to the Java Heap. Lack of memory can cause some functions to fail, like ManifoldCF's indexing of content. So steps should be taken to verify sufficient memory. It may not be immediately obvious that an issue exists. See the following KB for information on how to verify the current Java memory heap usage and how to increase the allotted memory if necessary.
Java memory heap limit
- Is Solr crashing during the indexing of content yet the Java processes have sufficient memory? You may need to update the Tika parsers file, which allows Solr to index PDF content. This only applies if you are on the 9.0 or 9.1 version of Ektron's Solr. You can update the tika file by upgrading Solr to the 9.2+ Ektron version. It is recommended to update your Ektron sites to 9.2+ as well or crawls will not be reported at the correct times.
- Which version of Solr is being used? 6.2.0 is the latest version Ektron can use. You can resolve many Solr issues by upgrading to the latest version. You can navigate to the below page to see the current Solr version.
http://NameOfSearchServer:7611/solr/ or http://NameOfSearchServer:7609/solr/
To download 6.2.0, go to the 9.20 Ektron Downloads and select the Solr install.
- Check the Solr log files for errors. In some cases, the logs can become very large, making searching them tedious. So it is best to clear them out, recreate the issue, then review the logs.