ahuston-0/hydra
1
0
Fork 0
Code Issues Pull Requests 1 Packages Projects Releases Wiki Activity

Allow only 1 thread to send a closure to a given machine at the same time

Browse Source 
This prevents a race where multiple threads see that machine X is
missing path P, and start sending it concurrently. Nix handles this
correctly, but it's still wasteful (especially for the case where P ==
GHC).

A more refined scheme would be to have per machine, per path locks.
This commit is contained in:
Eelco Dolstra
2015-07-07 14:04:36 +02:00
parent 16696a4aee
commit 35b7c4f82b
3 changed files with 7 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
src/hydra-queue-runner/build-remote.cc
View File

@@ -70,7 +70,7 @@ static void openConnection(const string & sshName, const string & sshKey,
static void copyClosureTo(std::shared_ptr<StoreAPI> store,
FdSource & from, FdSink & to, const PathSet & paths,
TokenServer & copyClosureTokenServer, counter & bytesSent,
counter & bytesSent,
bool useSubstitutes = false)
{
PathSet closure;
@@ -99,19 +99,6 @@ static void copyClosureTo(std::shared_ptr<StoreAPI> store,
for (auto i = sorted.rbegin(); i != sorted.rend(); ++i)
if (present.find(*i) == present.end()) missing.push_back(*i);
/* Ensure that only a limited number of threads can copy closures
at the same time. However, proceed anyway after a timeout to
prevent starvation by a handful of really huge closures. */
time_t start = time(0);
int timeout = 60 * (10 + rand() % 5);
auto token(copyClosureTokenServer.get(timeout));
time_t stop = time(0);
if (token())
printMsg(lvlDebug, format("got copy closure token after %1%s") % (stop - start));
else
printMsg(lvlDebug, format("did not get copy closure token after %1%s") % (stop - start));
printMsg(lvlDebug, format("sending %1% missing paths") % missing.size());
for (auto & p : missing)
@@ -194,7 +181,8 @@ void State::buildRemote(std::shared_ptr<StoreAPI> store,
if (machine->sshName != "localhost") {
printMsg(lvlDebug, format("sending closure of ‘%1%’ to ‘%2%’") % step->drvPath % machine->sshName);
MaintainCount mc(nrStepsCopyingTo);
copyClosureTo(store, from, to, inputs, copyClosureTokenServer, bytesSent);
std::lock_guard<std::mutex> sendLock(machine->state->sendLock);
copyClosureTo(store, from, to, inputs, bytesSent);
}
autoDelete.cancel();