We are seeing issues where users update their passwords and then cannot login. We traced it down to solr (but the root cause has not been identified yet) …a primeroctrl restart usually solves the problem however. I know that solr is restarted by cron every night so I was wondering if we should add the primeroctrl script to cron as well. Thoughts?
It has to do with the couch_watcher queue worker being out of sync due to a transaction (probably a Solr commit) that keeps failing. I recommend:
-
Stop couch watcher: (as a sudo user)
sudo supervisorctl stop couch-watcher -
Manually reindex: (as primero, from the primero home):
rails sunspot:reindexand address any indexing problems -
Reset the couch watcher history file
rails couch_changes:prime_sequence_numbers -
Restart primero (as a sudo user):
sudo /srv/primero/bin/primeroctl restart
1 Like