In the 3.11 release, the archivation process was reworked quite heavily, hopefully solving this problem. When the “database is locked” error happened before, old flows remained in
/var/lib/pakon.db, possibly for many days, which could lead to filling up the whole tmpfs. That hopefully shouldn’t happen anymore.
Also, as a safety measure, we added a hard limit of the
/var/lib/pakon.db database. When the database grows too much, the oldest records are deleted. The limit is pretty high, so it should be triggered only when the database is getting really insanely big. It’s probably less than ideal, but we don’t see a better way what to do in this situation. It’s probably better to lose a part of logs than to crash the router.