Closed sebastianhelbig closed 2 months ago
The problem is exactly what you wrote. The preferred setup is to use real-time (rt) index using docker. If you insist on the legacy plain index settings, then disable the indexer.* jobs in piler's crontab in the piler container, and run them in the manticore container.
If you decided to switch to RT index, then see https://manual.manticoresearch.com/Data_creation_and_modification/Adding_data_from_external_storages/Adding_data_to_tables/Attaching_one_table_to_another#Attaching-table---general-syntax how to make it happen.
Thanks for your fast answer. Since now already quite some mails are missing in the index, may it be easier to switch to a new empty manticore instance with RT=1 and let the index be rebuild as described here: https://www.mailpiler.org/recreate-the-index-data-files/ ?
Correct. Except you don't need to run the indexer scripts any longer.
piler: 1.4.6-80dadac RT: 0
We switched from a local installation to a docker based installation some days ago and migrated all the existing data. Basically everything is working, except that newly added mails won't show up in the web ui. The counter for received mails goes up, I can see the metadata in the mysql database and I also see that the mails are added to the store.
The problem seems to be that the indexing cronjobs inside the piler container run, but they are not updating the indexes on the separate manticore container, but instead they are updating indexes in /var/piler/manticore inside the piler container.
How to fix this?
piler.conf:
config-site.php:
manticore.conf:
docker-compose.yml: