hydra-queue-runner: More accurate memory accounting

Dependencies

• [2] MQNJPMCU Fix bad format string
• [3] BYVRA54Q Temporarily disable machines on any exception, not just connection failures
• [4] MB3TISH2 Rate-limit the number of threads copying closures at the same time
• [5] BG6PEOB2 Make the output size limit configurable
• [6] SL3WSRAC hydra-queue-runner: Limit memory usage
• [*] 5AIYUMTB Basic remote building
• [*] UVNTWTWG Prevent download of NARs we just uploaded
• [*] MHVIT4JY Split hydra-queue-runner.cc more
• [*] HJOEIMLR Refactor
• [*] N4IROACV Move buildRemote() into State

Change contents

• replacement in src/hydra-queue-runner/build-remote.cc at line 376
  [3.8141]→[3.8141:8244](∅→∅)
  − available. FIXME: only need this for binary cache

  − destination stores. */
  [3.8141]

  [3.8244]
  + available, which is twice the NAR size (namely the

  + uncompressed and worst-case compressed NAR), plus 150

  + MB for xz compression overhead. (The xz manpage claims

  + ~94 MiB, but that's not was I'm seeing.) */
• replacement in src/hydra-queue-runner/build-remote.cc at line 381
  [3.8306]→[3.8306:8374](∅→∅)
  − auto memoryReservation(memoryTokens.get(totalNarSize));
  [3.8306]

  [3.8374]
  + size_t compressionCost = totalNarSize + 150 * 1024 * 1024;

  + result.tokens = std::make_unique<nix::TokenServer::Token>(memoryTokens.get(totalNarSize + compressionCost));
• edit in src/hydra-queue-runner/build-remote.cc at line 396
  [9.232]

  [3.8937]
  + /* Release the tokens pertaining to NAR

  + compression. After this we only have the uncompressed

  + NAR in memory. */

  + result.tokens->give_back(compressionCost);

  +
• edit in src/hydra-queue-runner/builder.cc at line 182
  [9.441]

  [10.4678]
  +

  + result.accessor = 0;

  + result.tokens = 0;
• edit in src/hydra-queue-runner/state.hh at line 53
  [12.1810]

  [9.442]
  + std::unique_ptr<nix::TokenServer::Token> tokens;
• edit in src/hydra-queue-runner/token-server.hh at line 43
  [2.112]

  [4.2573]
  + debug("acquiring %d tokens", tokens);
• replacement in src/hydra-queue-runner/token-server.hh at line 58
  [4.2807]→[4.2807:2858](∅→∅)
  − Token(Token && t) : ts(t.ts) { t.ts = 0; }
  [4.2807]

  [4.2858]
  + Token(Token && t) : ts(t.ts), tokens(t.tokens), acquired(t.acquired)

  + {

  + t.ts = 0;

  + t.acquired = false;

  + }
• edit in src/hydra-queue-runner/token-server.hh at line 68
  [4.2969]

  [4.2969]
  + give_back(tokens);

  + }

  +

  + bool operator ()() { return acquired; }

  +

  + void give_back(size_t t)

  + {

  + debug("returning %d tokens", t);

  + if (!t) return;

  + assert(acquired);

  + assert(t <= tokens);
• replacement in src/hydra-queue-runner/token-server.hh at line 81
  [4.2954]→[4.2954:3030](∅→∅)
  − assert(*inUse >= tokens);

  − *inUse -= tokens;
  [4.2954]

  [4.3105]
  + assert(*inUse >= t);

  + *inUse -= t;

  + tokens -= t;
• replacement in src/hydra-queue-runner/token-server.hh at line 85
  [4.3119]→[4.3119:3156](∅→∅)
  − ts->wakeup.notify_one();
  [4.3119]

  [4.3156]
  + // FIXME: inefficient. Should wake up waiters that can

  + // proceed now.

  + ts->wakeup.notify_all();
• edit in src/hydra-queue-runner/token-server.hh at line 90
  [4.3167]→[4.3167:3215](∅→∅)
  − bool operator ()() { return acquired; }