DOC: README: MOVEONLY: Move outdated db documentation apart

Author: jtimon

README.md

@@ -173,52 +173,9 @@ python3 ./run_tests.py --dbs=dummydb,postgres --tests=api_faucet.py
 python3 ./run_tests.py --dbs=dummydb,postgres --tests=api_faucet.py,process_generator.py
 ```
 
-## Starting and restating the DB ##
-
-For starting and restating the database, the create_db process in
-https://github.com/jtimon/elements-explorer/blob/master/docker/explorer/Procfile
-must be active. Conversely, to stop the db from being restarted with
-every deployment/run, one must comment that line/process of the Procfile.
-
-Currently the db needs to be restarted whenever:
-
-- The schema changes in any way.
-- Some of the data stored in blob fields changes
-- The reorg handling code shows to be faulty
-- Any daemon witnessed a 100-block reorg
-- Any daemon crashed or got into a bad state that justifies restarting
-  its persistent cache
-
-
-## greedy caching ##
-
-As explained in the previous section, processes can be activated or
-deactivated in https://github.com/jtimon/elements-explorer/blob/master/docker/explorer/Procfile .
-
-One of the processes one may want to run when the db is not being
-restarted is the greedy cacher for any selected set of chains.
-
-It will slow down initialization and in general everything until it
-runs down to height 0 for its chain from the first successfully
-getchaininfo result at least once. After that, if the reorg handling
-processes are activated, it just double checks that blocks are cached
-at least once. Greedy cachers are independent from reorg handling
-processes, and thus what they cache may be deleted.
-
-Since they are very noisy, they are only recommended when you want to
-fill the db cache but you don't want to see extra messages in your
-logs. Deactivating them should be always fine.
-
-On anything resembling production, they should be activated too, even
-if they take days to fill the db cache the first time, since any
-response coming the db without bothering the daemon is one chance less
-for a denial of service attack, which, by the way, the whole project
-is exposed to even after that unless you're working on a closed
-network where you can take care of those attacks in some other way or
-if you're just using this project as a developer as you should. In
-that latter case you just have to remember not to DoS yourself harder
-than your machine can take, but please send back any weird concurrency
-errors you get.
+## More details about handling the DBs ##
+
+See ./doc/db.org
 
 # License