This may well be obviated by future work to allow the gc to kick out other queries, but for now, just make sure we don't wait indefinitely for the exclusive lock we need when dropping a partition.
While we're waiting, all other queries that touch the "parent" table will be blocked, which may be able to quickly block all command processing threads, queries, etc., for the duration.
The report and resource event garbage collections will now time out if they have to wait longer than 5 minutes (by default) to acquire the required table lock. This prevents them from blocking other related queries indefinitely, and prevents them from participating an any permanent deadlocks.
The report and resource event garbage collections will now time out if they have to wait longer than 5 minutes (by default) to acquire the required table lock. This prevents them from blocking other related queries indefinitely, and prevents them from participating an any permanent deadlocks. See also: https://puppet.com/docs/puppetdb/latest/configure.html#PDB_GC_DAILY_PARTITION_DROP_LOCK_TIMEOUT_MS