Commit b474c80b authored by David Auer's avatar David Auer
Browse files

Optimize random slicing's placeExtent

parent d7320eeb

include/global/random_slicing/DistRandSlice.hpp

+1 −1
Original line number Diff line number Diff line
@@ -91,7 +91,7 @@ public: 
    /**

     * @see Distributor::placeExtent

     */

    virtual std::list<Disk*>*

    uint64_t

    placeExtent(int64_t position);



    /**

src/global/random_slicing/DistRandSlice.cpp

+7 −58
Original line number Diff line number Diff line
@@ -281,65 +281,14 @@ my_hash(int64_t i) { 
    return i;

}



std::list<Disk*>*

uint64_t

DistRandSlice::placeExtent(int64_t position) {



    // XXX: for tests we want to avoid id clashing with copies so

    // int64_t offset = 200;

    // if(position > INT64_MAX / offset - 1)

    //     abort();

    // position *= offset;



    std::list<Disk*>* result = new std::list<Disk*>();

    std::vector<uint64_t> selected_partitions;



    uint64_t idx = 0;



#if defined DEBUG && defined DUMP_COPIES

    uint64_t retries = 0;

#endif



    while(result->size() < (uint32_t) m_copies) {

        // XXX: at the moment, we ignore the virtualVolumeId

        uint64_t key = my_hash(position + idx);



        // lookup interval containing the key and retrieve the partition ID

    // function heavily modified for GekkoFS

    assert(m_copies == 1);

    uint64_t key = my_hash(position);

    uint64_t partition_id = 0;

    m_interval_tree->search_tree(key, partition_id);



        // check if this partition has already been selected

        bool found = false;

        for(uint i = 0; i < selected_partitions.size() && !found; ++i) {

            if(partition_id == selected_partitions[i])

                found = true;

        }



        if(found) {

#if defined DEBUG && defined DUMP_COPIES

            ++retries;

#endif

            ++idx;

            // assert(idx <= (uint64_t) offset); // XXX

            continue;

        }



        // Find the corresponding disk and return it

        // XXX: ATM, partition_id == disk_id

        uint64_t disk_id = partition_id;

        assert(m_disks->count(disk_id) != 0);

        Disk* d = (*m_disks)[disk_id];



        result->push_back(new Disk(*d));

        selected_partitions.push_back(partition_id);

        ++idx;

    }



#if defined DEBUG && defined DUMP_COPIES

//    std::cout << "  DEBUG: Extent placed with " << retries << " retries " <<

//    std::endl;

#endif



    return result;

    return partition_id;

}



std::list<Disk*>*

src/global/rpc/distributor.cpp

+1 −8
Original line number Diff line number Diff line
@@ -152,14 +152,7 @@ RandomSlicingDistributor::locate_file_metadata(const std::string& path) const { 
    auto myhash = str_hash(path);

    // TODO placeExtent hashes the hash with it's own hash, is this a problem?

    // (migh be unavoidable if multiple copies should be stored sometime)

    auto disklist = dist_impl_->placeExtent(myhash);

    assert(disklist->size() == 1);

    auto disk = disklist->front();

    auto id = disk->getId();

    delete disk;

    delete disklist;

    // TODO Avoid memory allocation for this task and make this much simpler

    return id;

    return dist_impl_->placeExtent(myhash);

}



std::vector<host_t>