nextcloud / fulltextsearch

🔍 Core of the full-text search framework for Nextcloud
GNU Affero General Public License v3.0
210 stars 51 forks source link

fulltext search reset - does not reindex all files #857

Closed ferdiga closed 5 days ago

ferdiga commented 1 month ago

I had to reinstall elastic search as it got corrupted due to other reasons. (my fault probably)

I did a fulltextsearch:reset and it seems that only files which have not been indexed before are now indexed.

I use curl -X GET "localhost:9200/_cat/indices?v=true&s=index&pretty" and the number of indexed documents increases if I create new ones, but old ones are not indexed and not available in search (I use a special non existing string to be sure) - BTW these strings are found in the synced files in finder on my macbook.

both the tables oc_fulltextsearch_index and ..._ticks have been empty.

I assume that there must be another counter which was not reset.

ferdiga commented 3 weeks ago

Tried again with fulltextsearch 29.0.1 and NC 29.0.5

curl -X GET "localhost:9200/_cat/indices?v=true&s=index&pretty" does not show the resetted/deleted index any more.

occ fulltextsearch:test Index is already running

occ fulltextsearch:stop occ fulltextsearch:live shows "waiting" - before it gets killed (by a cron job.?) which restarts indexing always shortly after "stop"

but no indexing of documents occurs.

ferdiga commented 3 weeks ago

Just want to add https://github.com/nextcloud/all-in-one/discussions/1709#discussioncomment-9127017

the culprit may be the file - but I can not find it fts-index.done

ferdiga commented 5 days ago

I reduced the "max file size to index" to 10MB - now it works - instead of crashing after a few documents

  1. fulltextsearch:stop
  2. fulltextsearch:reset - this deletes the elastic search index
  3. fulltextsearch:index - restarts the scan of all documents