hydra-queue-runner: Limit memory usage

When using a binary cache store, the queue runner receives NARs from
the build machines, compresses them, and uploads them to the
cache. However, keeping multiple large NARs in memory can cause the
queue runner to run out of memory. This can happen for instance when
it's processing multiple ISO images concurrently.

The fix is to use a TokenServer to prevent the builder threads to
store more than a certain total size of NARs concurrently (at the
moment, this is hard-coded at 4 GiB). Builder threads that cause the
limit to be exceeded will block until other threads have finished.

The 4 GiB limit does not include certain other allocations, such as
for xz compression or for FSAccessor::readFile(). But since these are
unlikely to be more than the size of the NARs and hydra.nixos.org has
32 GiB RAM, it should be fine.

src/hydra-queue-runner/state.hh

@@ -9,14 +9,13 @@
 
 #include "db.hh"
 #include "counter.hh"
+#include "token-server.hh"
+
+#include "derivations.hh"
 #include "pathlocks.hh"
 #include "pool.hh"
-#include "sync.hh"
-
 #include "store-api.hh"
-#include "derivations.hh"
-
-#include "binary-cache-store.hh" // FIXME
+#include "sync.hh"
 
 
 typedef unsigned int BuildID;
@@ -354,6 +353,13 @@ private:
     std::shared_ptr<nix::Store> _localStore;
     std::shared_ptr<nix::Store> _destStore;
 
+    /* Token server to prevent threads from allocating too many big
+       strings concurrently while importing NARs from the build
+       machines. When a thread imports a NAR of size N, it will first
+       acquire N memory tokens, causing it to block until that many
+       tokens are available. */
+    nix::TokenServer memoryTokens;
+
 public:
     State();