diff --git a/docs/compiler.md b/docs/compiler.md
index 18660e703..6b36eeaef 100644
--- a/docs/compiler.md
+++ b/docs/compiler.md
@@ -33,6 +33,13 @@ a dialect of Go rather than a complete port of the language:
    it's up to the programmer whether assert can be performed successfully.
  * type aliases including the built-in `any` alias are supported.
  * generics are not supported, but eventually will be (at least, partially), ref. https://github.com/nspcc-dev/neo-go/issues/2376.
+ * `~` token is not supported
+ * `comparable` is not supported
+ * arrays (`[4]byte`) are not supported (https://github.com/nspcc-dev/neo-go/issues/3524)
+ * `min()` and `max()` are not supported (https://github.com/nspcc-dev/neo-go/issues/3090)
+ * `clear()` is not supported (https://github.com/nspcc-dev/neo-go/issues/3091)
+ * ranging over integers in `for` is not supported (https://github.com/nspcc-dev/neo-go/issues/3525)
+ * `for` loop variables are treated in pre-Go 1.22 way: a single instance is created for the whole loop
 
 ## VM API (interop layer)
 Compiler translates interop function calls into Neo VM syscalls or (for custom
diff --git a/docs/node-configuration.md b/docs/node-configuration.md
index b3a87ad08..3250fdd7f 100644
--- a/docs/node-configuration.md
+++ b/docs/node-configuration.md
@@ -403,3 +403,25 @@ where:
 
   Note that `Transaction` is a NeoGo extension that isn't supported by the NeoC#
   node and must be disabled on the public Neo N3 networks.
+
+## DB compatibility
+
+Real networks with large number of blocks require a substantial amount of time
+to synchronize. When operating a number of node instances with similar
+configurations you may want to save some resources by performing synchronization
+on one node and then copying the DB over to other instances. In general, this
+can be done and this is supported, but NeoGo has a lot of options that may
+affect this:
+- any differences in `ProtocolConfiguration` section make databases
+  incompatible, except for `Max*` settings, `MemPoolSize`,
+  `P2PNotaryRequestPayloadPoolSize`, `SeedList`, `TimePerBlock`.
+  Protocol configuration is expected to be the same on all nodes of the same
+  network, so don't touch it unless you know what you're doing.
+- DB types (Level/Bolt) must be the same
+- `GarbageCollectionPeriod` must be the same
+- `KeepOnlyLatestState` must be the same
+- `RemoveUntraceableBlocks` must be the same
+
+BotlDB is also known to be incompatible between machines with different
+endianness. Nothing is known for LevelDB wrt this, so it's not recommended
+to copy it this way too.