Commit f0177ac9 authored by Ramon Nou's avatar Ramon Nou Committed by Ramon Nou
Browse files

Changelog change and branch 

WIP - REPL

Added NUM_REPL env variable. (0 no replicas)

NUM_REPL  num replicas  (Replicas < servers)

Remove and truncate

Metadata replication - WIP

Metadata replication - Reattempt on stat

minimal compilation issues (c++20), srand for repl

Bitset

Bit set proposal (WIP)

Read - Write with bitset (<1024 chunks)

Changed bitset to vector

Added get_fs_config reattempt

Some more resilience on create

Added Replica_Check on write (disabled)

Added helper vector-bitset functions
parent e83e86d2

CHANGELOG.md

+4 −0
Original line number Diff line number Diff line
@@ -51,6 +51,10 @@ replicas ([!166](https://storage.bsc.es/gitlab/hpc/gekkofs/-/merge_requests/141) 
  fixed([!176](https://storage.bsc.es/gitlab/hpc/gekkofs/-/merge_request/176)).



## [0.9.2] - 2024-02

- Replication without using the server. NUM_REPL (0 < NUM_REPL < num_servers) env variable defines the number of 

replicas ([!166](https://storage.bsc.es/gitlab/hpc/gekkofs/-/merge_requests/141)).

- Modified write and reads to use a bitset instead of the traditional hash per chunk in the server.

- Added reattemp support in get_fs_config to other servers, when the initial server fails.



### New

README.md

+4 −0
Original line number Diff line number Diff line
@@ -371,6 +371,10 @@ start_t_micro: [5755,9553,14132,16841] 
total_bytes: 1802366

total_iops: 4

```

The user can enable the data replication feature by setting the replication environment variable:

`LIBGKFS_NUM_REPL=<num repl>`.

The number of replicas should go from 0 to the number of servers-1. 

The replication environment variable can be set up for each client, independently.



## Acknowledgment

include/client/preload_context.hpp

+1 −2
Original line number Diff line number Diff line
@@ -109,11 +109,9 @@ private: 
    std::bitset<MAX_USER_FDS> protected_fds_;

    std::string hostname;

    int replicas_;



    std::shared_ptr<gkfs::messagepack::ClientMetrics> write_metrics_;

    std::shared_ptr<gkfs::messagepack::ClientMetrics> read_metrics_;





public:

    static PreloadContext*

    getInstance() {
@@ -241,6 +239,7 @@ public: 


    const std::shared_ptr<gkfs::messagepack::ClientMetrics>

    read_metrics();



};



} // namespace preload

include/common/rpc/rpc_util.hpp

+1 −0
Original line number Diff line number Diff line
@@ -64,6 +64,7 @@ compress_bitset(const std::vector<uint8_t>& bytes); 
std::vector<uint8_t>

decompress_bitset(const std::string& compressedString);





} // namespace gkfs::rpc



#endif // GEKKOFS_COMMON_RPC_UTILS_HPP

src/client/gkfs_functions.cpp

+10 −1
Original line number Diff line number Diff line
@@ -882,8 +882,10 @@ gkfs_do_write(gkfs::filemap::OpenFile& file, const char* buf, size_t count, 
        errno = EISDIR;

        return -1;

    }



    auto path = make_unique<string>(file.path());

    auto is_append = file.get_flag(gkfs::filemap::OpenFile_flags::append);



    auto write_size = 0;

    auto num_replicas = CTX->get_replicas();
@@ -1089,17 +1091,24 @@ gkfs_do_read(const gkfs::filemap::OpenFile& file, char* buf, size_t count, 
    std::set<int8_t> failed; // set with failed targets.

    if(CTX->get_replicas() != 0) {



<<<<<<< HEAD

        ret = gkfs::rpc::forward_read(file.path(), buf, offset, count,

                                      CTX->get_replicas(), failed);

        while(ret.first == EIO) {

            ret = gkfs::rpc::forward_read(file.path(), buf, offset, count,

=======

        ret = gkfs::rpc::forward_read(file->path(), buf, offset, count,

                                      CTX->get_replicas(), failed);

        while(ret.first == EIO) {

            ret = gkfs::rpc::forward_read(file->path(), buf, offset, count,

>>>>>>> 9ff73051 (Changelog change and branch)

                                          CTX->get_replicas(), failed);

            LOG(WARNING, "gkfs::rpc::forward_read() failed with ret '{}'",

                ret.first);

        }



    } else {

        ret = gkfs::rpc::forward_read(file.path(), buf, offset, count, 0,

        ret = gkfs::rpc::forward_read(file->path(), buf, offset, count, 0,

                                      failed);

    }