Deployment strategy for synonyms / stopwords file

If you're using a synonym and / or stopwords file as part of your search analyzer definition my understanding is that you can update the analyzer by updating the synonym or stopwords file on every member of the cluster. Is there an API method to update the file on each cluster member, or is this a manual process? If it is manual, is there any recommended scripts to makes this task simpler?

Also, once the file is deployed I believe you just need to re-load the analyzers (since this is just a search analyzer, no re-index required). The recommended approach is to close and re-open the index, or restart each node one at a time. If you restart each node in this way, won't some have different synonym / stopwords files during the process? Does this cause any problems, or just mean you might get inconsistent results until all the nodes have restarted?

Currently there is no API to update a synonym file. This is a manual process, done on each individual node. So, it is possible that your synonym files may be inconsistent on different nodes. We currently don't check if synonym files are consistent across the nodes, so there shouldn't be problem from this side (apart from inconsistent search results). Closing and re-opening an index should be enough to load into the memory an updated synonym file.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.